Johan Granlund
Health Issues Raised by
Poorly Maintained Road
Networks
Northern European road users
may be exposed to unacceptable
health and safety risks, in terms
of ride vibration and skid acci-
dents. A case study has mapped
such risks on the Beaver Road
331. The report demonstrates
methods to efficiently prevent or
reduce risks on similar roads.
THIS PROJECT IS BEING P ART- FIN ANCED BY
THE EUROPEAN UNION
EUROPEAN REGION AL DE VELOP MENT FUND
Health Issues Raised by Poorly Maintained Road Networks
March 2008
Johan Granlund
Swedish Road Administration Consulting Services
PREFACE
This is a final report fro m Task B3 of the ROADEX III project, a technical trans-national cooperation
project between The Highland Council, Forestry Commission Scotland and Comhairle Nan Eilean Siar
from Scotland; The Northern Region of The Norwegian Public Roads Administration; The Northern
Region of The Swedish Road Administration (SRA) and the Swedish Forest Agency; The Savo-
Karjala Region of The Finnish Road Administration; the Icelandic Road Administration; and the Mu-
nicipality of Sisimiut from Greenland. The lead partner in the project is The Northern Region of The
Swedish Road Administration and project consultant is Roadscanners Oy from Finland. ROADEX III
project Chairman is Per-Mats Ăhberg from The Northern Region of The Swedish Road Administration
and project manager is Ron Munro of Roadscanners Oy.
The report was prepared by Johan Granlund of SRA Consulting Services, leader of Task B3. Fredrik
Lindström, Fredrik Stensson, Ylva Magnusson, Stefan Hedlöf, Erik Cuibe, Jenny Eriksson and Ulf
Nilsson of SRA CS participated in the plans, measure ments and analysis related to the field tests at
Rd 331. Hans Johansson of SRA Central Region made very valuable contributions to the traffic safety
analysis. Mats Nilsson and Anders Larsson of Brorssons Ă
keri AB drove the tested timber logging
truck and trailer combination. Ron Munro, manager of the ROADEX III Project, checked the project
report language. Mika PyhÀhuhta of Laboratorio UleÄborg designed the report layout.
The author would like to express gratitude also to the following:
âą
Mr Torbjörn Brorsson, owner of Brorssons à keri AB. Torbjörn, you made it possible!
âą
Mr Sören Dahlquist of System Technology AB and Mr Brendan Watts of Oxford Technical So-
lutions Ltd. Thanks for your kind and professional support of the truck ride measure ments us-
ing the Dewetron Stream Machine and the OxTS RT 3050 GPS/inertial system, respectively.
âą
Mr Leif GrĂžnskov of Greenwood Engineering A/S. Thanks Leif, for your support of the refer-
ence pavement condition measurements using the accurate laser/inertial Profilograph system.
âą
Dr Boris Thorvald of Scania Co mmercial Vehicles AB. Boris, you put the finger on short step
cross slope variance, truck wheel axle roll motion, related lateral forces, their impact on road
polishing, road friction and, ultimately, increased skid risk. This really widened our focus!
âą
MSc Henrik Lindh and Dr John Aurell of Volvo 3P. Thanks both of you, for making truck dy-
na mics more understandable!
All the good results from this research task are the fruits of successful teamwork. Should the reader
find anything to the contrary, the author takes full responsibility.
Finally the author would like to thank the ROADEX III Project Partners and the Project Steering Co m-
mittee for their guidance and encouragement in their work.
Copyright © 2008, the ROADEX III Project
All rights reserved.
ROADEX III Lead Partner: The Swedish Road Administration, Northern Region, Box 809,
SE-971 25 LuleÄ. Project co-ordinator: Mr. Krister Palo.
CONTENTS
ABSTRACT ............................................................................................................................................................... 7
CHAPTER 1. THE ROADEX PROJECT IN BRIEF ...............................................................................................8
CHAPTER 2. KNOWLEDGE TO FIND GOOD SOLUTIONS..............................................................................9
CHAPTER 3. HUMAN, VEHICLE AND ROAD INTERACTION ......................................................................10
3.1
R
IDE VIBRATION AND ITS EFFECT ON HUMAN BEINGS
......................................................................................... 10
3.2
A
N OVERVIEW OF HEAVY TRUCKS DYNAMICS
..................................................................................................... 21
3.3
R
ELATING RIDE FORCES TO PAVEMENT PROPERTIES
............................................................................................ 33
CHAPTER 4. CASE STUDY ON THE BEAVER ROAD 331 ...............................................................................51
4.1
T
RUCK TEST PARTNER
-
B
RO RSSONS
Ă
KERI
AB ................................................................................................. 54
4.2
A
S
CANIA
R480
164
G
6
X
4
WAS USED AS TEST TRUCK
....................................................................................... 55
4.3
TEN ROUNDTRIPS OF
280
KM WERE RECORDED
.................................................................................................... 56
4.4
C
OMPREHENSIVE RIDE AND ROAD CONDITION MEASUREMENTS
......................................................................... 57
CHAPTER 5. EXPECTED RESULTS ARE CONFIRMED .................................................................................63
5.1
UNACCEPTABLY HIGH
W
HO LE
-B
O DY
V
IBRATION AND SHOCK
........................................................................... 63
5.2
T
HE TRUCK SUSPENSION SYSTEMS PERFORMED VERY WELL
............................................................................... 72
5.3
G
OOD FIT BETWEEN
P
ROFILOGRAPH DATA AND TRUCK RIDE
.............................................................................. 78
5.4
D
RAFTING A
âW
ARPING LIMIT
â
FOR
RBCSV...................................................................................................... 94
CHAPTER 6. SPIN-OFF RESULTS ON TRAFFIC SAFETY..............................................................................96
6.1
E
FFICIENT ANALYSIS OF INCORRECTLY BANKED CURVES
................................................................................... 97
6.2
I
DENTIFYIN G HIGH SKID RISK DUE TO WATER PONDING
.....................................................................................105
6.3
F
RIC TION ISSUES DUE TO LOW OR VARIED
M
ACRO
T
EXTURE
.............................................................................115
CHAPTER 7. ETHICAL ASPECTS ON SAFETY ISSUES ................................................................................117
7.1
V
ISION
Z
ERO FOR ROAD SAFETY
.......................................................................................................................117
7.2
T
HE
T
YLĂSAND
D
EC LARATION
.........................................................................................................................117
7.3
P
RIORITIZING VARIOUS ROAD SAFETY IMPROVEMENTS
.....................................................................................118
CHAPTER 8. SERIOUS AND USEFUL FINDINGS ...........................................................................................120
8.1
R
IDE VIBRATION SHALL BE PREVENTED AT THE SOURCE
...................................................................................120
8.2
B
UMPS ARE MOST UNHEALTHY
.........................................................................................................................121
8.3
R
OLL VIBRATION REQUIRE SPECIAL FOCUS
.......................................................................................................122
ROADEX III The Northern Periphery Research
8.4
S
OME ROADS ARE MORE HAZARDOUS
â
NOW WE KNOW WHY
...........................................................................125
8.5
L
OW AND VARYING
M
AC RO
T
EXTURE CAUSE SKID ACCIDEN TS
........................................................................127
8.6
R
ETHINK CULVERT WORKS
................................................................................................................................127
CHAPTER 9. HOW TO USE THE NEW INSIGHT ............................................................................................129
9.1
H
AULIERS MUST MONITOR DRIVERS WBV EXPOSURE
........................................................................................129
9.2
D
EVELOPING USEFUL NEW VEHIC LE TECHNOLOGY
...........................................................................................130
9.3
I
MPRO VING ROAD TRAFFIC CRASH INVESTIGATIONS
.........................................................................................131
9.4
I
MPRO VED ROAD MANAGEMENT
.......................................................................................................................132
9.5
R
OAD DESIGN POLICY IMPROVEMENTS
..............................................................................................................138
9.6
W
ORK TO BE CONTINUED
(
IN
ROADEX
IV?) ...................................................................................................139
CHAPTER 10. FURTHER READING ..................................................................................................................140
ROADEX III The Northern Periphery Research
ABSTRACT
The EU ROADEX Project 1998 - 2007 is a trans-national roads co-operation aimed at develop-
ing ways for interactive and innovative management of low traffic volume roads throughout the
cold climate regions of the Northern Periphery Area of Europe. Its goals have been to facilitate
co-operation and research into the common problems of the Northern Periphery.
The overall objective for this research task was to increase the understanding for road userâs
health risks when riding on roads in poor condition. Better knowledge will facilitate the reduction
of the risks, by means of improved pave ment management, more conscious truck, bus and am-
bulance operations, and inspire to vehicle suspension systems improvements.
The report co mmences with generic descriptions of how safety and health can be affected by
ride vibration, how truck suspension systems isolate and amplify vibration at various frequen-
cies, and how pavement properties, such as cross slope, control the important forces at work.
A case study is reported from the Beaver Road 331 in northern Sweden. A heavy timber logging
truck was instrumented to measure ride vibration and direction. Measurements were taken at a
range of points (seat, cab, frame and wheels) and the results stored together with data on
speed and interior noise. Ride vibration data fro m repeated rides over a 280 km long round trip
from forest to coast industries was then compared with reference data on pavement condition,
scanned by a laser/inertial Profilograph. Results obtained included:
âą
The daily exposure to Whole-Body Vibration, the A(8)-value, for timber truck drivers rid-
ing constantly on roads such as Rd 331 were unacceptably high, when compared to the
health and safety Action Value in Directive 2002/44/EC.
âą
The truck drivers were exposed to unacceptably health risks in the back when driving at
modest speed over the worst bumps, due to high spinal compression doses, S
ed
, as per
the ISO 2631-5 standard.
âą
A derived draft limit of 0.30 % for undesired variance in cross slope. This could be useful
in pavement manage ment to prevent roll-motion and lateral forces in road vehicles.
The case study also produced valuable spin-offs in new methods for analyzing traffic safety
risks arising from incorrectly banked curves and low drainage gradients. Hospital records from
accidents at Rd 331 (mainly skid accidents) were found to match road sections with high cross
slope variance, curves with incorrect superelevation, transition sections with low drainage gradi-
ents, and sections with high skid risk due to low/varying Macro Texture. These serious findings
call for both short and long term actions. Road agencies should use the demonstrated methods
to quickly identify hazardous sites and warn road users of them. Road repair planning and prac-
tices should be improved, and funding for road repair should be increased.
An extraordinary insight after the case study is that many new roads all over the EU Northern
Periphery area have skid risks inbuilt due to low drainage gradients at entrances and exits of
certain curves. These risks should be eliminated by modification of road design codes, road de-
sign software, road construction practices, and improved end quality control.
Page 8
ROADEX III The Northern Periphery Research
Chapter 1. THE ROADEX PROJECT IN BRIEF
The ROADEX Project is a technical co-operation between roads organisations across northern
Europe that aims to share roads related information and research between the partners. The
Project was started in 1998 as a 3 year pilot co-operation between the roads districts of Finnish
Lapland, Tro ms County of Norway, the Northern Region of Sweden and The Highland Council
of Scotland and was subsequently followed and extended with a second project, ROADEX II,
from 2002 to 2005, and now a third project, ROADEX III, from 2006 to 2007.
The partners in ROADEX III âThe
Implementation Projectâ comprised
public road administrations and forestry
organizations fro m across the European
Northern Periphery. These were The
Highland Council, Forestry Commission
Scotland & Comhairle Nan Eilean Siar
from Scotland, The Northern Region of
The Norwegian Public Roads
Administration, The Northern Region of
The Swedish Road Administration and
the Swedish Forest Agency, The Savo-
Karjala Region of The Finnish Road
Administration, the Icelandic Road
Administration and the Municipality of
Sisimiut from Greenland.
A priority of this Project was to take the collected ROADEX knowledge out into the Partner ar-
eas and deliver it first hand to practising engineers and technicians. This was done in a series of
14 seminars across the Partner areas to a total audience of 800. Reports were translated into
the 6 partner languages of Danish, Icelandic, Finnish, Greenlandic, Norwegian and Swedish as
well as English. ROADEX research continued through 5 projects: measures to improve drainage
performance, pavement deformation mitigation measures, health issues of poorly maintained
roads, road condition management policies, and a case study of the application of ROADEX
methodologies to roads in Greenland. All of the reports are available on the ROADEX website at
www.roadex.org
.
Page 9
ROADEX III The Northern Periphery Research
Chapter 2. Knowledge to find good solutions
The overall objective for the research task was to increase the understanding of the health risks
to road users when travelling on poor quality roads. Better knowledge of this will facilitate risk
reduction, by means of improved pave ment management, more conscious truck operations, in-
spire to future vehicle suspension system design i mprovements, et cetera.
Three goals were set for the research:
âą
The first goal was to assess a typical truck driverâs daily exposure to ride vibration, in re-
lation to the EU Action Value, when driving on a typical Northern Periphery rural round
trip route.
âą
The second goal was to investigate if truck drivers riding on very bumpy roads may be
exposed to so intensive mechanical shock, that there is a risk for mechanical fatigue
da mage in the hard tissue of their spine intervertebral end plates.
âą
The third goal was to validate and draft limits for an indicator of undesired variance of
pavement lane cross slope. Such variance excites roll motion which is especially prob-
lematic in high (heavy) vehicles. Roll vibration may not only be uncomfortable and un-
healthy, but it also brings transient lateral forces that may cause skid accidents on slip-
pery surfaces.
If the third goal was reached, it was hoped that the new pavement condition indicator could be
put into daily practice by road agencies in their pavement manage ment syste ms. Through this it
could be possible to identify hazardous sections and have the m repaired. Many roll-related skid
accidents could thus be prevented and truck driverâs exposure to vibration and health risk de-
creased.
Page 10
ROADEX III The Northern Periphery Research
Chapter 3. Human, vehicle and road interaction
Many readers will be able to read and appreciate the subsequent chapters, without first reading
this rather long chapter. However, since the research topic covers several disciplines, it is likely
that some readers may appreciate a condensed background that can bridge minor gaps of
knowledge in unfamiliar disciplines. This chapter gives a brief summary on ride vibration and its
effect on human beings. It also gives an introduction to heavy vehicle chassis dyna mics. Finally,
it shows how road geometry and condition of the pavement excite vital ride forces and vibration.
3.1 RIDE VIBRATION AND ITS EFFECT ON HUMAN BEINGS
3.1.1 General health risks associated with ride vibration
Back disorders are costly to society and are the main causes of sick leave in the working com-
munity. They cause great pain to those suffering, and are a significant economic burden to soci-
ety. Professional drivers are a group of workers that have been found to be at high risk for back
disorders. Many epidemiological studies have been made on the relationship between back dis-
orders and vehicle operation with vibration exposure. The results show overwhelming evidence
of a relationship that is consistent and strong, which increases with increasing exposure, and is
biologically plausible. The risk is elevated in a broad range of driving occupations, including
truck and bus drivers. Vibration exposure data indicates that current vehicles are likely to ex-
pose drivers to vibration levels in excess of the EU Action Value, as defined in directive
2002/44/EC [2]. Common control measures, such as seat suspension, are often not effective in
the low frequency range where vibration energy peaks during most highway rides. A causal link
has been found between back disorders, driving occupation and ride vibration. Numerous back
disorders are involved, including lumbago, sciatica, generalized back pain, and intervertebral
disc herniation and degeneration. Elevated risks are consistently observed after five years of
exposure, see Teschke et al (1999) [1].
Whole-Body Vibration (WBV) is the term used to describe mechanical vibration and shock
transmitted to the human body as a whole, usually through areas of the body (buttocks, soles of
the feet and the back) in contact with a vibrating surface as seen in Figure 1. Vehicles travelling
over rough surfaces expose people to periodic, rando m and transient ride vibration. Ride vibra-
tion contains many frequencies, occurs in several directions (bounce, pitch and roll) and
changes over time. Exposure to ride vibration causes various patterns of oscillatory motions and
forces within the human body; a co mplex, intelligent and active structure. WBV within the range
0.5 â 80 Hz cause resonance in various parts of the human body, such as the eye globes, head,
Page 11
ROADEX III The Northern Periphery Research
spine and stomach. Thus, WBV exposure may cause
1
stressing disco mfort or annoyance, influ-
ence human performance capability or present a health and safety risk (e.g. pathological dam-
age or physiological change). The response to ride vibration varies in a confounding way; while
bumps have a stressing alarm effect, the rocking motion when riding over long wave undula-
tions results in drowsiness.
Figure 1
Whole-Body Vibration. From the EU Guide to good practice on WBV (2006) [24]
Bovenzi & Hulshof (1999) [47] reviewed epide miologic studies conducted between 1986 and
1997 on the relationship between exposure to vibration and problems in the lumbar part of the
back. The review provided â
clear evidence for an increased risk for LBP disorders in occupa-
tions with exposure to WBV. Biodynamic and physiological experiments have shown that seated
WBV exposure can affect the spine by mechanical overloading and excessive muscular fatigue,
supporting the epidemiologic findings of a possible causal role of WBV in the development of
(low) back troubles
â. It is estimated that 4 - 7% of the working population in the EU are exposed
to potentially harmful Whole-Body Vibration.
The National Institute for Occupational Safety and Health reports that musculoskeletal injuries,
such as low back pain, vertebrogenic pain, and degenerative disk disease, account for 1 out of
5 of emergency-room-treated occupational injuries. Physical demands of many jobs make the
musculoskeletal system highly vulnerable to a variety of occupational injuries and illnesses.
WBV is one of the most important etiologic factors behind development of these disorders [27].
Hedberg (1991) [32] reported that the risk for certain types of cardiovascular disease in Sweden
is more than three times higher for commercial drivers than for the average worker. An in-
creased risk of myocardial infarction a mong professional drivers was first reported about 50
years ago, and has been reported repeatedly since then. Stress under certain driving conditions
1
A thorough guidance on evaluation of human exposure to ride vibration is given in part 1 and 5 of the international
standard ISO 2631 [
18
] [
5
].
Page 12
ROADEX III The Northern Periphery Research
is considered to explain the raised level of stress hormones found in commercial drivers, and is
believed to cause a large proportion of the health problems, see Hedberg (1993) [33]. The in-
creased incidence and mortality from ischemic heart disease among Swedish truck drivers has
remained constant over the period 1985 â 1996, as shown by Bigert et al (2004) [26]. Hedberg
& Langendoen (1989) [34] showed that amongst older commercial drivers, musculoskeletal
problems and cardiovascular diseases are the pri mary reasons for changing their occupation.
3.1.2 Truck and bus drivers vibration exposure may exceed the EU Action Value
The European health and safety directive on physical agent vibration, 2002/44/EC [2], defines a
measure A(8) for workersâ 8 hour daily exposure to Whole-Body Vibration. If the A(8) exceeds
the Action Value of 0.5 m/s
2
, the directive demands employers to take organizational and/or
technical measures to minimize the vibrations. Work tasks that bring exposures above the limit
A(8) = 1.15 m/s
2
are prohibited. In Sweden the exposure limit is sharpened into 1.1 m/s
2
. Since
2005, the directive minimum require ments have been implemented in all EU member state na-
tional laws. The directive is showed in Figure 2.
Figure 2
EU Physical Agents Directive â Vibration 2002/44/EC [2], front page
Professional drivers may be exposed to high vibration exposure and risk. The main reasons are
that vibration intensity is higher in heavy vehicles (as compared to passenger cars), and the ex-
posure time is often close to 8 hours per day. Ahlin et al (2000) [3] collated the vibration expo-
Page 13
ROADEX III The Northern Periphery Research
sure of truck drivers with road roughness, vehicle type and condition, as well as with driving be-
haviour such as speed. Among the conclusions were that many heavy vehicle drivers in Swe-
den may be exposed to vibrations above the EU Action Value A(8) = 0.5 m/s
2
. Within reason-
able ranges, the degree of road roughness was found to have much larger impact on driverâs
WBV exposure, than factors such as driving speed, vehicle type and vehicle condition.
3.1.3 Bumps are of special concern to both ride quality and health
3.1.3.1 Measuring discomfort due to bumpy rides
Human exposure to occasional shock has large impact on the perceived ride quality. It is there-
fore very important that indicators of ride quality reflect comfort disturbance caused by
shocks/bumps. Four methods to evaluate ride comfort are used all over the world today. A good
overview of these is given by Els (2005) [60]. Most i mportant is the ISO 2631-1 (1997) [18].
SpÄng (1997) [48] showed that the running Root-Mean-Square (RMS) of the weighted accelera-
tion (using integration time 1 s) is a useful definition of bumpy vibration. This definition corre-
lated very closely with annoyance perceived by a large test panel; R
2
= 92 %. The running rms
method is used for transient vibration in the ISO 2631-1 standard.
For public transport, the running rms values can be compared to the (dis-)co mfort scale in Table
1. This vibration comfort scale is used for people in public transport on roads, railways, in air
and at sea. The level of annoyance depends on passenger expectations with regard to trip dura-
tion and the passengerâs activities (e.g. reading, eating or writing) and many other factors
(acoustic noise and temperature). Therefore limits are therefore not explicit, but include over-
laps.
Table 1
Indicative comfort reactions of people in Public transportation, as per ISO 2631-1 [18]
min
max Comfort level
> 2
> 2
"Extremely uncomfortable"
1,25
2,5
"Very uncomfortable"
0,8
1,6
"Uncomfortable"
0,5
1
"Fairly uncomfortable"
0,315
0,63
"A little uncomfortable"
0
0,315
"Not uncomfortable"
a
w
rms
Hassan & McManus (2001) [51] showed that professional drivers perceive somewhat lower
comfort disturbance for a given vibration magnitude, than seen in Table 1. Two causal factors to
this finding have been identified. First, the driver can see large road obstacles and are better
prepared, (their resonance-sensitive organs are protected by increased muscle tonus) when the
resulting vibration co mes. Secondly, the driver has a steering wheel in the hands, thus being
better able to stabilize pitch and fore-aft motion of the upper body. However, there is no special
comfort scale for professional drivers yet.
Further reading is given in the section â
3.1.4 Especially vulnerable road user groups
â.
Page 14
ROADEX III The Northern Periphery Research
3.1.3.2 Measuring health risk due to bumpy rides
Sandover (1998) [4] made an review of expert opinion, stating that in general, transient vibra-
tions with multiple shocks are much more hazardous than stationary vibration. In practice, this
means that bumpy rides typically are unhealthier than ride vibration such as on a modestly
wash-boarded gravel road. There are many examples demonstrating the risk from bumps, in-
cluding spinal compression fractures when riding snowmobiles or military combat vehicles in
rough terrain.
A method to quantify Whole-Body Vibration containing multiple shocks in relation to human
health was standardized in 2004. The ISO 2631-5 [5], method uses peak vibration values to
predict compression stress in the spine, and reports equivalent daily static co mpression dose,
S
ed
. A
S
ed
value above 0.8 MPa reflects a high health risk due to transient mechanical shocks.
In contrast, a
S
ed
value below 0.5 MPa corresponds to a non-significant risk. The employer is
obliged to perform a risk assess ment for workers exposed to repeated mechanical shock, such
as from bumpy rides. In Sweden, such assess ments are, in practice, made in accordance with
ISO 2631-5. Results from such assess ments of professional drivers emphasize the importance
of a smooth road surface to keep health risks low.
Repeated driving over bumps, resulting in transient vibration yielding a
S
ed
over 0.5 MPa, may
be prohibited by the Work Environment Administration. One exa mple is the recent prohibition,
coupled with a 1 000 000 SEK (over 100 000 âŹ) fine, against risks associated with line bus traf-
fic over severe traffic calming speed humps on VikingavÀgen (the Viking Road) and Luf-
thamnsvÀgen (the Airport Road) in TÀby, Sweden. After the prohibition in the spring of 2007, the
traffic on several bus routes totally stopped until each hump was repaired or totally removed, so
that the
S
ed
was reduced to less than 0.5 MPa. See Brandt & Granlund (2008) [6].
Marjanen (2005) [61] studied transient vibration in 25 mobile machines for 30 hours. The results
showed that the ISO 2631-5 method, based on
S
ed
value, gave a worse rating of bumpy expo-
sures than the ISO 2631-1 method based on RMS-value. The latter is relevant for calculation of
the daily vibration exposure A(8), as defined in the directive 2002/44/EC [2]. An illustrating result
was an â
uncomfortable
â exposure with RMS = 0.85 m/s
2
, gave a
S
ed
âvalue of 2.92 MPa. So
while the RMS was below the exposure limit of A(8) = 1.15 m/s
2
, the exposure was high above
the 0.8 MPa limit for high health risk defined in ISO 2631-5.
When investigating methods applicable to tactical ground vehicles, Alem (2005) [62] found the
ISO 2631-5 method to be more sensitive to cross-country terrain rides than other standards.
The report mentions an anecdote on he maturia (blood in the urine) being observed in 50 % of
the company, after completing a military exercise mission.
According to the ISO 2631-5 standard, the X-Y axis natural frequencies in the human spine are
about 2.1 Hz. Therefore it is important that humans are not exposed to strong vibrations at fre-
quencies around 2.1 Hz in or across the direction of travel. This should be recognized, when
assessing risks associated with undesired variances in pavement cross slope, which in high ve-
hicles can cause transient roll motion acco mpanied by lateral (and vertical) vibration.
Further reading is given in the following section â
3.1.4 Especially vulnerable road user groups
â.
Page 15
ROADEX III The Northern Periphery Research
3.1.4 Especially vulnerable road user groups
Some people are especially vulnerable to vibration. Bumpy rides may be detri mental to:
âą
People with certain disabilities, diseases or injuries.
âą
Pregnant wo men and their unborn babies, see Armstrong et al (1988) [36] and Council
Directive 92/85/EEC.
âą
Injured
ambulance
patients.
The Academy of Pediatrics (1999) [29] states, that ambulance transports may cause decreased
vascular tone, manifested by unexpected decreases in blood pressure. Ride vibration may
cause care equipment to co me loose, cause settings to change, or produce disturbances in
monitors and equipment. Furthermore, vibration may decrease the a mbulance nurseâs ability to
perform care procedures.
Many ambulance patients report that the pain suffered during the transport as being the worst
experience in their whole life. [Personal communication with Leif Leding of the Swedish Ambu-
lance Academy].
The European trend towards fewer and more specialised hospitals is resulting in a greater per-
centage of a mbulance transports having to cover longer distances while simultaneously admin-
istering intensive care. To manage this more, and heavier, medical equipment is required to be
carried on board. As a consequence of this, modern a mbulances must have a greater load ca-
pacity than before. Large vehicles, with a similar design to trucks, are needed. An effective load
capacity of more than 1 tonne is common for Mobile Intensive Care Units (MICU). Ahlin et al
(2000) [3] showed that ride vibration is significantly worse in large MICU a mbulance vehicles,
than in small Emergency Ambulance vehicles.
One of the few efficient methods to reduce ride vibration in ambulance cars, is for the driver to
âreadâ the pavement surface condition and by risky driving avoid the worst roughness, as seen
in Figure 3.
Figure 3
Emergency Ambulance on wrong side of Road 331, avoiding edge deformations
Page 16
ROADEX III The Northern Periphery Research
3.1.5 Vibration intensities in road vehicles and mobile machinery
Most road vehicles, including modern trucks with suspended cabs, have fairly low levels of
Whole-Body Vibration (WBV), given that the pave ment is in good condition. Vehicles with less
effective suspension, such as trucks with non-suspended cabs, may cause high WBV levels.
Heavy truck vertical vibration is maximum when the truck is unloaded, while roll and lateral vi-
bration tend to peak when the truck has full payload. Of course, the WBV exposure is very de-
pendent on the quality of road surfaces, vehicle speeds and other factors such as how the vehi-
cle is operated. Therefore, it is often necessary to measure the vibrations, or the road condition,
in order to make an accurate risk assessment. An indicative exa mple of vibration levels in road
vehicles is given in Table 2. As can be seen fro m the table, the A(8) EU Action Value of 0.5 m/s
2
corresponds to a work environment being on average âfairly unco mfortableâ for the full working
day.
Table 2
Indicative example of Whole-Body Vibration magnitudes in road vehicles
Passenger
cars 0.1
to
1
m/s
2
average for route.
Up to 2 m/s
2
at bumps.
Heavy trucks
0.2 to 1.6 m/s
2
average for route.
Often over 2 m/s
2
at bumps.
Reference on co mfort, as per ISO 2631-1 [18]
< 0.315 m/s
2
is ânot unco mfortableâ.
> 0.5 m/s
2
is âfairly uncomfortableâ.
EU
Action
Value
[2]
A(8)
=
0.5
m/s
2
, âaverage over 8 hoursâ
From the table, it can be seen that truck drivers are exposed to markedly higher vibration inten-
sity than car drivers. Campbell et al (1981) [35] explained some reasons:
1. Driver location - namely, the truck driver is usually located at the extremities of the vehi-
cle, rather than near its centre of gravity.
2. Trucks are more dynamically active at low frequencies of excitation, as caused by the
use of articulation for manoeuvrability and frame flexibility for durability
3. Truck suspension systems possess substantial amounts of dry friction, thereby trans mit-
ting more road input to the vehicle.
The chassis suspensions on heavy vehicles are also designed for a much wider range of pay-
load, than on passenger cars. In addition, heavy vehicles have heavier unsuspended masses
(tyre, rim, brake and axles) than cars do. When the unsuspended mass hits a bump, it transfers
energy to the vehicle body. A heavier mass can transfer more vibration energy than a lighter
mass.
For a general co mparison purpose, examples of typical vibration levels in mobile machinery
used in civil engineering, forestry and industry works are given in Table 3. For machines that are
often operated at considerable speed, such as graders and tractors, the highest vibration levels
are usually generated in road transport mode. The table has been reproduced from the
EU
Page 17
ROADEX III The Northern Periphery Research
Guide to good practice on Whole-Body Vibration
[24], which gives useful guidelines to reduce
risks from WBV exposure.
Table 3
Examples of WBV magnitudes in common mobile machinery [24]
Page 18
ROADEX III The Northern Periphery Research
3.1.6 Ride vibrations have a negative effect on traffic safety
In the mid 1970âs, the exposure of truck drivers to vibration was an issue raised at the federal
government level in the USA, formulated as â
Do vibrations (as well as noise, toxic fumes and
other factors that contribute to truck âride qualityâ) have a negative effect on driver health and on
highway safety?
â Eventually, a five-year research programme, â
Ride Quality of Commercial Mo-
tor Vehicles and the Impact on Truck Driver Performance
â, answered this question. The findings
were summarised in the report
Truck Cab Vibrations and Highway Safety
[30]. This report was
jointly produced by leading researchers, road authorities, vehicle manufacturers, hauliers and
commercial drivers. It shows that the answer to the key question as to whether there is any cor-
relation between cab vibrations and road safety is YES; see illustration in Figure 4. Yes, there is
good reason to believe that vibrations affect driversâ health, and that vibration must be elimi-
nated at source through effective road maintenance rather than merely dampened. The report
concludes that if road network deterioration is allowed to continue, the result will be serious
health and road safety problems.
Figure 4
The primary elements in the link between truck ride vibration and safety [30].
When studying Figure 4, take note of the described driver response in terms of stress and car-
diovascular effects. This associates with the research results on increased level of stress hor-
mones and mortality from ischemic heart disease, referred in the previous section â
3.1.1 Gen-
eral health risks associated with ride vibration
â.
Road Feel
Vibration
Driver Response
Driverâs Performance
Vehicle Performance
Accident Potential
Ride Excitation
Roughness
Low Level,
Distributed
Speed Influence
Frequency
Amplitude
Vehicle has:
âą Variable Payload
âą Stiff Suspension
âą Dry Friction in the Springs
âą Flexible Frame
âą Multiple Coupled Masses
âą Rotating Nonuniform Components
Ride Environment
Multi-Modal Vibration
Dominant Frequencies
1 to 10 Hz
Dominant Amplitudes
up to 2 m/s
2
Rattling Control
Instruments and Mirrors
Possible Momentary
âUnseatingâ of Driver
Driver Response
Physiological Effects
eg. Visual
Cardiovascular
Psychological Effects
eg. Discomfort
Stress
Physical (Biodynam ic)
Response â Vibration of
Body Parts and Organs
Pathological Effects
eg. Back Pain
Driver/Vehicle
Performance
Vehicle
- Road Holding
- Component Wear
- Component Failure
Driver
- Visual Acuity
- Reaction Times
- Tracking and other
Motor Tasks
Accidents
Potentially Influenced by
-Prefatigue Vigilance
- Fatige-decreased Vigilance
-Driversâ Modulation of Steering
and Brake Controls
- Driver Affected by Ailments
- Image Clarity in Mirrors
-âUnseatedâ Driver by Severe Jolt
- Vehicle Road-Holding
Performance
Severe,
Localized
Page 19
ROADEX III The Northern Periphery Research
The most common lethal truck accident mode is rollover. Less common modes are jack-knifing
and trailer swing. Jack-knifing means the accidental folding of an articulated vehicle, similar to a
pocket knife. When the prime mover skids, the trailer can push it from behind until it spins round
and hits the trailer. Adverse road conditions, such as a slippery road surface, or an obstacle (i.e.
curbs) hitting the rear wheels, may contribute to jack-knifing. Most truck drivers are skilful
enough to correct a skid before the vehicle combination undergoes jack-knifing. Trailer swing is
easier to correct. Side forces that result from cornering, operating on a crowned road, and side
winds accelerate the jack-knife situation.
Research by Ihs et al (2002) [31] confirms a positive correlation between road roughness (ride
vibration) and traffic accident
2
frequency (crash risk) in Sweden, see Figure 5. Rough roads with
an IRI
3
-value over 3 mm/ m show more than 50 % higher crash rate than s mooth roads with an
IRI below 0.9 mm/m. The study also showed that as roughness beco mes very severe (over 10
mm/m), the crash rate increases even more than shown by the slope of the linear graphs.
The graphs at Figure 5 also show that the crash rate is much higher in the winter, than in the
summer. This is due to factors such as lower road surface friction on icy roads and darker driv-
ing conditions.
Figure 5
Rough roads have > 50 % higher crash rate. After Ihs et al [31]
2
In the study, accidents in junctions and with wild animals were excluded.
3
IRI = International Roughness Index
Page 20
ROADEX III The Northern Periphery Research
3.1.7 Summarizing health and safety risks for EU Northern Periphery road users
The Swedish National Institute of Public Health has found that in Sweden, the most common
types of preventable mortalities are lung cancer (death rate of 17.1), suicide (15.4) and cerebro-
vascular disease (11.8). Among the therapeutic treatable death causes, diabetes mellitus is
worst with a death rate of 4.5. Road traffic crashes are worse, with a death rate of 4.9 on aver-
age for the whole country. However, there are large differences in the risk of being killed in a
road traffic accident between different areas in Sweden. While the urban areas of Stockholm,
Gothenburg and Malmoe have a Standardised Mortality Ratio (SMR) of 70 for road traffic
crashes, the rural areas have a SMR of 177. This means that road users in the rural areas have
153 % higher risk in ending up in a lethal crash, as compared to road users in large cities. Of
the rural counties in Sweden, JĂ€mtland and VĂ€sternorrland have the highest SMR for road traf-
fic crashes. In these counties, road traffic crashes are taking 39 % more lives than diabetes is
[58].
This chapter has clearly showed that EU Northern Periphery road users are exposed to serious
health and safety risks. Professional drivers are exposed to a very high risk for stress-related
cardiovascular diseases, having three time higher rate for certain types of cardiovascular dis-
ease than other people. They are also exposed to high risk for musculoskeletal problems. All
road users in the Northern Periphery are at high risk for being injured in traffic crashes on rough
and poorly maintained roads. These serious findings call for responsible corrective actions.
Such actions must be very well focused, since the available funds are sparse compared to the
size of the Northern Periphery road networks.
Page 21
ROADEX III The Northern Periphery Research
3.2 AN OVERVIEW OF HEAVY TRUCKS DYNAMICS
Why do truck suspension systems isolate some vibrations very well, yet amplify others? This
section will attempt to answer this type of question. The following texts are inspired by several
sources, including handbooks such as the
Fundamentals of Vehicle Dynamics
[7]. The suspen-
sion performance figures come from a presentation on heavy trucks dynamics [8], and are re-
printed with kind permission by MSc Henrik Lindh, supervisor on vehicle dyna mics at Volvo 3P.
3.2.1 Sources of truck ride vibration
The term ride
4
vibration describes motion with frequencies from 0.5 to 25 Hz. Truck vibrations in
the ride frequency range are excited by both internal and external sources. Internal sources in-
clude engine combustion pulses, power train i mbalance, non-uniform wheel geometry and non-
uniform tyre stiffness. External sources include pavement roughness, pavement deflection vari-
ance [9] and air pressure variance (wind load, or air bursts fro m fellow vehicles or reflections
from road tunnel walls).
3.2.2 Influence of road roughness, vehicle factors and speed
In brief, truck cab vibrations are primarily determined by road condition, with vehicle properties
being secondary, as seen in Figure 6 by ForssĂ©n (1999) [10]. Note the â0.00â effect of tyre pres-
sure variance. This result is from tests within normal pressure level recommendations, while ex-
tremely low pressures when using Central Tyre Inflation (CTI) systems have a large effect.
Figure 6
Effect of road roughness and vehicle properties on truck cab vibration [10].
There is, in general, also a strong positive correlation between speed and vibration. At speed
levels below some 30 km/h, such as in parking lots and during off-road driving at construction
sites, an increase in speed by one percent will increase vibration several percent.
4
Many vehicle engineers distinguish between RIDE as < 5 Hz and SHAKE as 5 â 25 Hz vibration.
Page 22
ROADEX III The Northern Periphery Research
At highway speeds however, the effect of speed on ride vibration is rather small. To reduce vi-
bration a certain percent, it may be necessary to reduce speed twice as much. An example from
a 20 km test site is given in Figure 7. A 29 % speed reduction (fro m 70 km/h to 50 km/h) re-
sulted in only 18 % reduction in average vibration and 15 % reduction in maximum vibration.
Also at 30 km/h, the 0.5 m/s
2
EU Action Value was exceeded. The speed limit was 90 km/h [9].
Ahlin & Granlund (2002) [11] showed in a theoretical analysis that when driving at highway
speed levels, a large effect of speed change on ride vibration can only be expected when the
road roughness consists of high amplitudes at long wavelengths. If there is a high degree of
roughness with intermediate-length, the speed must be reduced to parking lot speed level, i.e.
below 20 km/h, in order to reduce vibration significantly. If only very short wave roughness is
present, the chassis vibration may in fact be reduced by increasing the driving speed. (The latter
is a very rare exception however, since most rough roads also have high amplitudes at long
wavelengths).
1,18
0,97
0,74
3,15
2,67
2,24
0
0,5
1
1,5
2
2,5
3
3,5
70
50
30
Driving speed [km/h]
R
M
S
o
f w
e
ig
h
ted
vi
b
ra
ti
o
n
ac
cel
e
rat
io
n
[
m
/s
2
]
Average level over 20 km
Max level (over 1 s)
Figure 7
Influence of speed on seat vibration in a loaded Volvo FL12 on Rd 374, n w Storfors, Sweden
Page 23
ROADEX III The Northern Periphery Research
3.2.3 Heavy trucks have several suspension systems
All highway vehicles have a suspension system designed to isolate vertical vibration from the
wheels to the vehicle body. The primary functions of a chassis suspension system are to [7]:
âą
Provide vertical compliance so the wheels can follow uneven road surfaces, while isolat-
ing the vehicle body from the roadâs roughness.
âą
Maintain the wheels in proper steer and camber attitudes to the road surface.
âą
React to tyre control forces â longitudinal (acceleration and braking) forces, lateral (cor-
nering) forces, and braking and driving torques.
âą
Resist chassis roll motion.
âą
Keep the tyres in contact with the road under minimal load variations.
Obviously, chassis suspension systems must meet many more demands - not least in a safety
perspective - than to âonlyâ isolate the cab from vibration and shock of various frequencies, di-
rections, a mplitudes and interacting histories.
A modern heavy truck has several suspension systems, as seen in Figure 8. In fact, the trucks
frame may also be considered as a suspension system, with its flexural bending modes. Inside
the cab, most truck driverâs seats are today also equipped with a suspension system.
Figure 8
Heavy trucks have several suspension systems
Page 24
ROADEX III The Northern Periphery Research
3.2.3.1 The tyre acts like a spring
Counting from the vibration main source, the road surface, the truckâs first vibration filter com-
prises its tyre.
Due to the enveloping effect of its contact patch, the tyre smears variance in pavement Macro
Texture (MaTx have waves from 0.5 to 50 mm length) and to some extent Mega Texture (MeTx
have waves from 50 to 500 mm). The tyre walls also act like springs, which - under noise gen-
eration - further absorb texture variance and interact with the vertical motions of the vehicle
body and unsprung masses. The tyre does not provide significant damping, with regard to the
lower frequencies of ride.
The vertical vibration isolation performance of a typical truck tyre is demonstrated in Figure 9.
The upper left graph shows an example of road profile spectra (tyre input), while the upper right
graph shows wheel axle acceleration (tyre output). The bottom curve shows the quotient of road
âaccelerationâ
5
and axle acceleration. This gain curve shows the tyreâs isolation performance. A
gain below 1 means that the system is isolating vibration, while a gain over 1 means that it is
amplifying. As seen, the tyre efficiently isolates vibration with frequencies higher than some 12
Hz. At the tyre eigenfrequency of about 8 - 12 Hz, vibrations are amplified. This resonance re-
sponse is known as âwheel axle hopâ, and contributes to wash boarding of poor dirt roads. Vi-
brations with low frequencies are transmitted straight through the tyre, which then follows the
road profile like a rigid body.
5
The time domain term âroad accelerationâ may be somewhat confusing to road engineers. It corresponds to the
spatial domain term âSlope Varianceâ (SV). SV was a key parameter in the 1958 â 1960 AASHO Road Test [12],
used as an index for short wave road roughness. In the giant AASHO Road Test, SV was found to be the most impor-
tant factor behind the Publicâs judgement of road serviceability. In fact, the Road Test results showed that SV (road
roughness) was many times more important for road users ratings, than rutting, cracking and patch repair altogether.
Page 25
ROADEX III The Northern Periphery Research
Figure 9
Input and output vertical acceleration of a truck tyre [8]
The basic relationship between roughness wavelength (
λ
) [unit: m], travel velocity (
Ï
) [unit: m/s]
and vertical vibration frequency (
f
) [unit: Hz] is given in Formula 1.
f
Μ
λ
=
Formula 1, Wavelength, Velocity and Frequency
When driving at 30 km/h (8.3 m/s), the tyre spring resonance of 8 - 12 Hz occurs on 0.7 - 1 m
road roughness wavelength, as per Formula 1. At 90 km/h, tyre resonance occurs on 2 - 3 m
waves.
Page 26
ROADEX III The Northern Periphery Research
3.2.3.2 The chassis suspension isolates the vehicle body from the wheels
Counting from the vibration main source, the road surface, the truckâs second vibration filter
comprises its
chassis suspension
.
The vertical vibration isolation performance of a typical truck chassis suspension is de mon-
strated in Figure 10. The upper left graph shows our example of wheel axle acceleration (chas-
sis suspension input), while the upper right graph shows frame acceleration (chassis suspen-
sion output). The bottom curve shows the quotient of the axle and fra me acceleration. This gain
curve shows the chassis suspensionâs isolation performance.
The truck chassis suspension system efficiently isolates vibration with frequencies higher than
some 5 - 6 Hz. Vibrations with some 2 - 4 Hz are a mplified. Vibrations with low frequencies are
transmitted straight through the suspension.
When driving at 30 km/h, the suspension resonance of 2 - 4 Hz occurs over road roughness
with 2 - 4 m wavelength. At 90 km/h, suspension resonance occurs over 6 - 12.5 m roughness
waves.
Figure 10
Input and output vertical acceleration of a truck chassis suspension system [8]
Ride vibration is typically small in amplitude, involving some tenfold millimetres of suspension
travel. Many truck suspension systems exhibit nonlinear properties, due to friction in struts and
bushings, or interleaf friction in leaf springs. Gillespie [7] explains that for small ride motions, the
effective stiffness may be three times greater than the no minal spring stiffness. Therefore some
Page 27
ROADEX III The Northern Periphery Research
trucks may actually be more comfortable on roads with a lot of short wave roughness, than on
roads without such roughness. However, this is not true for long waves and in cars.
3.2.3.3 Frame beaming
Counting from the vibration main source, the road surface, the truckâs third vibration filter com-
prises its
frame
. It is somewhat questionable whether or not to consider the frame as a suspen-
sion system. However, its negative effect on ride is significant at
beaming
resonance frequen-
cies.
It is custo mary to make heavy truck fra mes flexible, due to commercial demands such as low
deadweight, fatigue resistance and traction in off-road conditions.
Figure 11 shows a 34 Degree-Of-Freedom truck model. Here the frame is modelled as consist-
ing of 6 beam sections, connected with longitudinal and torsion springs and da mpers (not
shown). Examples of beaming are showed later, in Figure 13.
Note the integrated model of the suspended truck engine, constituting a powerful source of in-
ternal vibration with higher frequencies. The first harmonic of an engine running at 1500 rpm is
at 1500 / 60 = 25 Hz, while higher order engine harmonics have higher frequencies.
Figure 11
Truck model having a flexible frame [8]
Page 28
ROADEX III The Northern Periphery Research
3.2.3.4 The cab is isolated from the frame
Counting from the vibration main source, the road surface, the truckâs fourth vibration filter com-
prises its
cab suspension
.
The vertical vibration isolation performance of a typical truck cab suspension is demonstrated in
Figure 12. The upper left graph shows our example of frame acceleration (cab suspension in-
put), while the upper right graph shows cab acceleration (cab suspension output). The bottom
curve shows the quotient of the frame and cab acceleration. This gain curve shows the cab
suspensionâs isolation performance. The system efficiently isolates vibration with frequencies
higher than some 3 Hz. Vibrations with some 1 - 2 Hz are amplified. When driving at 30 km/h,
suspension resonance of 1 - 2 Hz occurs on road roughness with 4 - 8 m wavelength. At 90
km/h, resonance occurs over 12 - 25 m waves. Vibrations with very low frequencies are trans-
mitted straight through the suspension.
Figure 12
Input and output vertical acceleration of a cab suspension system [8]
Page 29
ROADEX III The Northern Periphery Research
3.2.4 Net vibration in a truck cab
As seen in the above graphs, each of the suspension syste ms in a truck isolate vibration very
efficiently. This is true for vibration with frequencies above each systemâs eigenfrequency. For
vibration close to an eigenfrequency, resonance results in amplification instead of isolation.
Resonance is seen as values above 1 in the above gain curves for each system.
For an articulated tractor-trailer truck cab, an example of net vibration content is given in Figure
13. The graph has resonance peaks at the eigenfrequencies of each suspension system.
Note that on the vertical logarithmic scale of the Figure, the cab vibration is many times more
powerful at 1 - 3 Hz frequencies than at higher frequencies. For highway speeds of 50 - 90
km/h, the 1 - 3 Hz frequency range corresponds to 5 - 25 m long road unevenness wavelengths.
Figure 13
Power Spectral Density of net vibration in a truck cab [8]
Page 30
ROADEX III The Northern Periphery Research
3.2.5 Seat suspension
A good seat improves the situation further, and the most efficient vibration isolating seats used
in road vehicles can be found among truck driver seats.
Figure 14 shows vertical vibration data taken by Ahlin et al [3] at cab floor and at the driverâs
seat in a Volvo F12 6x2 timber logging truck, when travelling on National Highway 90 in North-
ern Sweden. By comparing the graphs, it is clear that cab floor vibration at frequencies over 3
Hz are quite efficiently isolated from the drivers buttocks by the advanced air suspended truck
seat.
However, the graph for the seat pan shows highest seat vibration intensity at frequencies be-
tween 1.5 and 2.5 Hz. By comparing the graphs in Figure 14, it is clear that the expensive air
suspended seat does not isolate vibration, but rather amplifies vibration, at the dominant fre-
quencies below 3 Hz.
Cab floor
Driverâs seat
Figure 14
Power Spectral Density of vibration on the seat and on the floor in a truck cab [3]
32 kilometres of the Hw 90 test section were very rough, while 5 kilometres were smooth. The
root-mean-square for the weighted vertical acceleration on the seat was 0.96 m/s
2
on the 32 km
very rough section of Hw 90, while the corresponding figure for the 5 km smoother section was
0.38 m/s
2
. The figure 0.96 m/s
2
from the rough section is much higher than the A(8) = 0.5 m/s
2
action value for an daily eight-hour reference period, set in the health and safety directive
2002/44/EC [2]. Clearly, the ride vibration problem related to roads with similar roughness (es-
pecially long wave, as indicated fro m frequency content in Figure 14), is very serious. After the
truck ride measurements were taken by Ahlin et al [3], the 32 km long rough section of Hw 90
was reconstructed by Swedish Road Administration.
A new and promising technology for da mping in seat and cab suspension syste ms is based on
MagnetoRheology (MR). While developed for seat suspensions, MR-technology may be more
successful in large vehicle cab suspensions, where it can be used with soft springs without
comprising ride and stability. LeRoy (2006) [68] claims that MR can provide both roll isolation
and pitch stability.
Page 31
ROADEX III The Northern Periphery Research
3.2.6 Wheel axle vibration impacts on traffic safety
As a truck wheel axle is exposed to roll vibration, any arbitrary point at its outermost parts (as
well as on the attached wheels) moves in lateral direction. This results in distortion to the
tyre/road contact patch, making the laterally moving patch polishing the road into lower friction
for following vehicles. This contact patch distortion also increases the vehicleâs own need for
road friction and increases its tyre wear [Personal communication with Dr Boris Thorvald,
Scania Commercial Vehicles AB].
Less obvious is that when an axle is exposed to roll vibration, it also yaws. This is analyzed by
Ahmadian & Ahn (2003) [28]. As a wheel moves upwards and the suspension is mo mentarily
compressed, its toe angle changes. This results in a steering response, similar to the driver
slightly turning the steering wheel. If a steering effect occurs as both wheels go up parallel, the
pheno menon is called âbump steerâ. If it occurs as one wheel rises and the other falls (axle roll),
it is called âroll steerâ. Solid axles generally have zero bump steer. The occurrence of roll steer
is more or less inevitable, but the degree of severity differs between axle and truck models.
A soft chassis suspension may result in severe bump steer behaviour, especially by the front
steering axle. In Australia, this has been identified by McFarlane & Sweatman (2003) [37] as a
source of poor lane-keeping behaviour on rough road sections. Where the road width is narrow,
these lateral disturbances may require the driver to increase concentration into a stress level
significant for driver fatigue.
Roll vibration also results in significant vehicle fatigue damage, as discussed in a Ph D thesis by
Bogsjö (2007) [22]. The research reported by Bogsjö is based on a large amount of road condi-
tion data, measured with one of SRA CS Profilographs on a set of roads including Rd 331; the
road studied in this ROADEX III research task.
Page 32
ROADEX III The Northern Periphery Research
3.2.7 Final remarks on heavy truck dynamics
It is well known, even amongst non-specialists, that
stiffness
and
damping
are important sus-
pension design parameters. The potential of softer suspensions as a method to reduce cab vi-
bration has been studied by Ăijer & Edlund at Volvo 3P [14]. The results show that a very soft
suspension may reduce cab vibration by some 6 - 20 %, which is a clearly noticeable difference.
The study also showed that after resurfacing the test road, the cab average vibration was re-
duced by 67 % and its peak vibration by 85 %.
When heavy trucks are exposed to roll forces at frequencies below some 3 Hz, resonance may
cause the roll response to be larger than the input. There is a theoretical possibility that the
truck roll eigenfrequency can be reduced, and thus the entire roll resonance, by designing the
vehicle with extremely low roll stiffness (by reducing spring vertical stiffness and minimizing lat-
eral separation of left and right springs). However, as illustrated in Figure 15, the interaction be-
tween
stiffness
and
damping
also includes a third, and less publicly recognized parameter;
spring travel (
or
deflection/displacement
). Low roll stiffness brings large roll displacements and
very poor cornering performance, in terms of a major tendency to rollover in connection with fast
large lateral manoeuvres such as in the fa mous âMoose testâ (quick lane change). This is of
course unacceptable fro m a traffic safety perspective, and therefore not suitable to imple ment in
practice [Personal conversation with Dr John Aurell, Volvo 3P].
Figure 15
The dynamic triangle of stiffness, damping and spring travel [8]
When designing single syste ms therefore, the vehicle manufacturer must consider the complete
vehicle performance. Other requirements than vibration environment must also be considered,
such as commercial aspects, stability, handling and safety.
The bottom line is:
If there really would have been any good quick fixes to radically improve truck drivers
ride vibration environment without sacrificing other important issues like traffic safety, the
large and skilled engineer teams at Volvo, Scania and other truck manufacturers would
have implemented them long ago.
Page 33
ROADEX III The Northern Periphery Research
3.3 RELATING RIDE FORCES TO PAVEMENT PROPERTIES
This section gives a brief summary on pavement geometry design, required to balance ride
forces in curves (dyna mic equilibrium), and how pavement roughness - as well as bearing ca-
pacity variance - excite heavy truck ride vibration. The texts are inspired by various road design
manuals and handbooks, such as the AASHO Policy on Road Design [16].
3.3.1 Tight curves are hazardous
Tight curves are typically more hazardous than straight road sections. An example of this is
seen in the map of hospital reported accidents on the section of Rd 331 shown in Figure 16.
While the straight sections of the road showed no accidents, the wide curves show some, and
the tight curves show several accidents. In fact, more accidents happened in the sharp âRoos
curveâ than at the junction with the significantly more trafficked National Highway 87. In the fol-
lowing sections, dyna mic imbalance due to incorrectly banked curves is explained as a risk as-
sociated with tight curves. The subject of dynamic equilibrium of cornering forces was initially
not considered for the project task, but serious findings during the project work made it inevita-
ble to include it into the report.
Figure 16
The south exit of Roos curve in Ăsterforsse shows unusually many accidents
Page 34
ROADEX III The Northern Periphery Research
3.3.1.1 Steady cornering require dynamic equilibrium by correctly banked curves
This first part of this section analyses the âexcitingâ lateral force acting on a cornering vehicle.
Thereafter, the analysis continues with the âreactionâ forces needed to keep the vehicle steady in
the desired curved path. When balancing these forces, the key exciting factors are vehicle (ref-
erence) speed and the road horizontal curvature, whilst the key retaining factors are lateral fric-
tion and pavement superelevation (single sided cross slope in curves).
At highway speeds on wet road surfaces, road friction is basically a function of pave ment Macro
Texture (MaTx) only. Thus, in slippery conditions, the cornering reaction forces depend totally
on texture and the pavement superelevation. Under extremely slippery conditions, the lateral
friction may drop to almost zero. One example is when driving on black ice. In such conditions,
the only reaction force available to balance the ride is totally related to the pavement superele-
vation/cross slope.
This section ends by showing how this knowledge is used when designing superelevation in
curves on new roads, and concludes that this knowledge is not yet sufficiently used in the man-
agement of curves on the existing road network.
Failure modes in accidents related to curve design and road friction
De Solminihac et al (2007) [38] have studied accident outcomes in horizontal curves, and have
seen that light vehicles are more prone to run-off than are trucks, whereas the main failure con-
dition for trucks and SUVâs is roll over.
Strandberg (1974) [53] related the truck rollover problem to the fact that many heavy vehicle
combinations have poor rollover (overturning) stability. It is unusual that passenger cars rollover
at lateral accelerations below 10 m/s
2
. However, the rollover limit is often less than 3 - 4 m/s
2
for
trucks. A half empty tanker with a bad suspension might roll below 2 m/s
2
. While passenger cars
require high friction and extreme skid to rollover, trucks may rollover on slippery surfaces with-
out much warning to the driver. Strandberg also referred to numerous of investigations showing
that most truck drivers use larger lateral accelerations at low speeds than at high speeds. Two
of the most efficient truck design improvements to safety are utilizing maximum lateral distance
between chassis suspension springs and implementing anti-roll bars. Both actions result in
higher roll stiffness, thereby increasing roll vibration.
Persson & Strandroth (2005) [39] identified skidding as a common failure mode in lethal crashes
on Swedish roads. During wintertime, 53 % of the lethal skid accidents occurred on thin and
very slippery âblack iceâ. Wide roads with a high standard of winter operations did not feature to
any extent in skid statistics. Krafft et al (2006) [44] compared Swedish accident outcome for
cars with and without an antiskid system. They found that antiskid systems reduced the risk of
accidents involving human injury by over 13 % lower on dry road surfaces. Furthermore, on
slippery surfaces, antiskid syste ms reduced the risk by an astonishing minimum of 35 %. This
shows that the efficiency of antiskid systems as safety equipment is almost as fundamental as
of a seatbelt. This further confirms skidding as a common and very serious safety risk on icy low
volume roads in the northern parts of EU.
Page 35
ROADEX III The Northern Periphery Research
The exciting lateral force
As described by Newtonâs second law of mechanics, cornering vehicles undergo centripetal ac-
celeration acting toward the centre of the curvature. As seen in Formula 2, the associated lat-
eral
6
force
F
is a product of vehicle mass
m
[kg] and squared vehicle speed
v
[m/s], divided by
the curve radius
R
[m]. For a vehicle with given reference speed, the lateral force depends only
of the curve radius. Smaller radii (tighter curves) yield higher lateral forces. For tight curves,
even a minor increase in radius results in a large decrease of the lateral force.
R
m
F
2
*
Μ
=
Formula 2, Lateral acceleration force acting on a cornering vehicle
Figure 17
shows the factors influencing the cornering forces acting on a vehicle as described by
the âPoint mass modelâ, used in road design manuals worldwide. These are the gravitational
force
G
[N], the normal force
N
[N], the lateral force
F
[N], the side friction (demand) factor
f
s
[-],
and tangent of the angle
α
corresponding to pavement superelevation/banking/cross slope [%].
The total road grip between tyre and pave ment can be divided into a tangential part (braking
friction, longitudinal direction) and a radial part (side friction, lateral direction). The side friction is
the part of the total road grip normally utilized when cornering.
Figure 17
Vehicle cornering forces [
15
]
6
In
Figure 17
, the centripetal force is substituted by a corresponding centrifugal force in the opposite direction. Even
though people in a cornering vehicle perceive a âcentrifugal forceâ, it is fictive (not real) on the vehicle. This report
follows the practice set used in many road design manuals, by referring to the (fictive) centrifugal force, rather than to
the fundamentally correct centripetal force with opposite direction.
Page 36
ROADEX III The Northern Periphery Research
The reaction forces needed to balance the ride
If the lateral force
F
is not balanced by reaction forces, the vehicle ride will beco me unstable
and the risk of a traffic accident (run-off, skidding and rollover) will increase. There are two reac-
tion forces that may balance the lateral force
F
. One is the horizontal component of the normal
force;
N *
sin(
α
). The other is the horizontal component of the side friction developed between
the vehicle's tyres and the pavement surface friction force,
N
*
f
s
*
cos (
α
)
. This can be ex-
pressed by the equation in Formula 3.
)
cos(
*
*
)
sin(
*
α
α
s
f
N
N
F
+
=
Formula 3, Lateral equilibrium
After division by cos (
α
), the equation can be written as Formula 4.
)
)
(tan(
*
)
cos(
s
f
N
F
+
=
α
α
Formula 4, Lateral equilibrium (2)
After substitution with
N
=
m
*
g
(g being the gravitation constant) and with
F
as per Formula 2,
the equation can be further developed as Formula 5.
)
)
(tan(
*
*
)
cos(
*
*
2
s
f
g
m
R
m
+
=
α
α
Μ
Formula 5, Lateral equilibrium (3)
After elimination of
m
and recalling that cos(
α
) is close to 1 for small angles (from a mathemati-
cal point of view, pavement cross slopes are small angles), the equation is, with good approxi-
mation, finally expressed as Formula 6.
s
f
g
R
+
â
)
tan(
*
2
α
Μ
Formula 6, Lateral equilibrium (final expression)
This shows that a steady cornering is totally depending on the sum of the cross slope (banking)
and the side friction factor. The correct application of banking reduces the need for side friction;
while incorrect banking may instead increase the need for side friction.
Page 37
ROADEX III The Northern Periphery Research
Figure 18 shows the âdemandâ side, the left hand side, of Formula 6 as a linear function of lat-
eral force (Curvature
7
) for speeds of 30 to 110 km/h. A similar graph as function of curve radius
is shown in Figure 19. These graphs show that slippery surfaces in very tight curves (R < 200 to
300 m) may be a challenge at low speed levels of 30 to 50 km/h also.
In very slippery conditions, when friction approaches zero, a cornering vehicle must be retained
by another force other than friction. As seen in the right side of Formula 6, the only retaining fac-
tor beside friction is banking. Banking can be designed up to 5.5 % in Sweden. As friction gets
low, this banking can also be decisive for safe cornering in flatter curves as can be seen by the
de mand values in Figure 19. Conditions that create slippery roads when cornering at highway
speeds include black
8
ice, bleeding asphalt, surface contamination such as mud and sand, as
well as driving with threadbare slick-worn tyres on a wet road surface.
0,00
0,05
0,10
0,15
0,20
0,25
0,30
0
1
2
3
4
5
6
7
8
9
10
Horizontal Curvature, 1000/R, [m
-1
]
D
e
man
d
ed
su
m [
-]
30 km/h
50 km/h
70 km/h
90 km/h
110 km/h
Figure 18
Demanded sum of superelevation and side friction to balance the cornering force
7
Curvature is defined as 1000/
R
, thus being directly proportional to the exciting lateral force as seen in Formula 2.
Curvature also has another advantage over radius, when analyzing and reporting road alignment data. While straight
sections make the radius approach +/- infinity (which is difficult to plot in a linear scale), curvature approaches 0 and
is easy to plot. This is fundamental to plots as in Figure 71.
8
Black ice, also known as "glare ice" or "clear ice," typically refers to a thin coating of glazed ice on a surface, often a
roadway. While not truly black, it is transparent, allowing the usually-black asphalt/macadam roadway to be seen
through it, hence the term. It is unusually slick compared to other forms of ice on roadways. It often has a matte ap-
pearance rather than the expected gloss; and often is interleaved with wet pavement, which may be identical in ap-
pearance. For this reason it is especially hazardous when driving or walking because it is both hard to see and ex-
tremely slick. [Source: Wikipedia, encyclopedia on Internet]
Page 38
ROADEX III The Northern Periphery Research
0,00
0,05
0,10
0,15
0,20
0,25
0,30
0
200
400
600
800
1000
1200
1400
1600
1800
2000
Horizontal Curve Radius, R, [m]
D
e
m
a
nde
d s
u
m
[
-]
30 km/h
50 km/h
70 km/h
90 km/h
110 km/h
Figure 19
Demanded sum of superelevation and side friction to balance the cornering force
Braking tests may not always correctly reproduce what can happen in practice. Even though
adequate friction numbers are recorded in tests, cornering in under-banked curves may still end
in run-off accidents under certain winter conditions. This can occur on a snow layer. Under test
conditions a test tyre may be able to penetrate the snow during intensive braking, and find grip
in the underlying asphalt. In practice however this may not happen as the (lower) cornering
forces may be insufficient to penetrate the slippery snow layer. Under these conditions, correct
banking may be the only safeguard for safe cornering.
Page 39
ROADEX III The Northern Periphery Research
Designing superelevation in curves on new road sections
The side friction factor needed to balance the lateral force in poorly banked curves can be com-
pared with the
side friction factor used for design purpose
(set by the road agency). In order to
maintain a safe margin with respect to the inevitable temporary low friction conditions such as
due to snow, ice, water, bleeding asphalt and poorer than average tyres, the factor used in de-
sign must be substantially lower than the demand friction factor. The side friction factor used in
Sweden for superelevation design purpose is given by Formula 7. It is also shown in Figure 20.
Μ
*
6
.
3
*
0096
.
0
dim
,
*
28
.
0
â
=
e
f
s
Formula 7, Side friction design factor in Sweden [15]
where
f
s, dim
= side friction factor used for design [-]
e
= the natural logarithm [-]
v
= design speed [m/s]
Figure 20
Swedish design values for total friction, brake friction and side friction [15]
The side friction factor used in Sweden for road design corresponds to approximately 2/3 of
measured friction between good car tyres and wet asphalt pavements in good condition. Gilles-
pie (1992) [7] reports that truck tyres generally exhibit lower friction values than cars, because of
higher unit loading in the contact patch and different tread rubber compounds.
After applying the design factor, standard sheets with ideal superelevation values can be calcu-
lated for each speed li mit level. The sheet for 90 km/h is given as an exa mple in Figure 21.
Page 40
ROADEX III The Northern Periphery Research
Figure 21
Ideal ratios for superelevation at 90 km/h, as function of horizontal radius [15]
As mentioned above, superelevation design is based on analysis of the forces acting on a point-
mass model. This analysis assumes that the driver will follow a perfect curved path at the de-
sign speed. This assumption can however be far from real world conditions. For exa mple, the
path travelled in practice often includes transient curvatures (and thus lateral forces) much
higher than assumed, such as when changing lane when overtaking another vehicle, yawing to
compensate for wind bursts, or yawing to avoid road da mages such as potholes.
There is, as already stated, a margin between the side friction demand factor and the design
side friction factor, but the increased accident rate experienced at many sharp curves, such as
on the new expressway between Falun and BorlÀnge in Sweden, questions whether the current
margin is large enough. It is possible that the deviations between the design model assumptions
and the real world conditions, as described above, are too large to be covered by the existing
margin. If so, the design of reversed superelevation, as currently allowed as per note 2 in Figure
21, should be reviewed.
Page 41
ROADEX III The Northern Periphery Research
Maximum values of cross slope and banking/superelevation
For snow and ice contamination, superelevation should not exceed a slope on which vehicles
standing or driving slowly would slide toward the centre of the curve. In Norway, the maximum
allowed value for banking in existing hairpin curves is 9.5 % [41], see Figure 22. Sliding in a
section with 9.5 % banking may happen when the side friction factor is below 0.095 (9.5 %).
However, the consequences of a slip incident at low speed are likely to be milder than those of
skidding in high speed due to too low banking of a hairpin curve.
Figure 22
Guidelines for maintenance of banking in existing curves in Norway
Up until now, there has been an absence of a national guideline for superelevation on existing
roads in Sweden. For the design of new curves in new road sections, the maximum value is 5.5
% [15] (plus a construction tolerance of some 0.5 % [40]). This conservative Swedish road de-
sign code is however irrelevant for hairpin curves, which are inevitable in the mountainous roads
in Norway. Very sharp curves also exist on old roads in Sweden however, and they often have
much higher banking than 5.5 %. Values of more than 10 % are surprisingly common.
When slowly driving in a curve with higher-than-needed superelevation, the vehicle follows the
desired path only when the driver steers up the slope. Since steering against the direction of the
horizontal curve is unnatural to a driver, such curves may be perceived as âdifficultâ or uncom-
fortable.
When trucks and SUVâs with high centres of gravity and a soft suspension travel slowly on steep
cross slopes, a large share of their weight is distributed to the down-slope tyres. If this condition
becomes extre me, the vehicle may easily rollover as discussed by Strandberg (1974) [53].
Page 42
ROADEX III The Northern Periphery Research
Poor control of superelevation in existing curves
It appears that accumulated research and experience has resulted in reasonable design values
for banking/superelevation in new curves. However, much less effort has been spent on how to
manage existing curves.
The side friction factor used for design cannot be used to define a sharp limit between safe and
unsafe existing roads. It can however be used to evaluate existing horizontal curve geometry
(radius and banking/superelevation) against the very same âhighest acceptable risk levelâ as
applied when designing new road network sections, by using Formula 6.
As described above, the design of superelevation in new curves is based on analysis of the
forces acting on a point-mass model. This analysis assumes that the driver will traverse a per-
fect curved path at the design speed. For existing curves on old roads, this assumption can be
very unrealistic. Comprehensive surveys show that on low volume roads, operating speeds of-
ten exceed the design speed, see de Solminihac et al (2007) [38]. This is an undesired but hard
fact that road managers must deal with, and not ignore. Furthermore, the geometric characteris-
tics of old curves can seldo m be described by simple parameters such as a single radius value,
since the alignment can be so poor that the curvature (and thus the lateral force) varies signifi-
cantly. One such exa mple was found in Profilograph data from a Swedish National Highway,
where 20 skid accidents took place in a 200 m long section of a curve, during the winter
2006/2007. On investigation, it was found that the curvature (lateral force) was doubled within
fractions of a second just at the multi-crash length.
With the wide variance of real vehicle speeds in curves, there is always an unbalanced force
whether the curve is superelevated or not. As discussed in the AASHO Policy on Road Design
[16] unbalanced force results in tyre wall thrust, which is taken up by the friction between the
tyres and the road surface. This reaction force is developed by distorting the contact area of the
tyre. Keeping this distortion low, keeps the road surface from polishing and tyre wear low. These
are further reasons to control and correct superelevation in existing curves.
Page 43
ROADEX III The Northern Periphery Research
3.3.1.2 Road factors decisive for road grip and stability on straight roads
Hydroplaning by highway vehicles is a phenomenon characterized by a complete loss of direc-
tional control. When a tyre is moving fast enough, it rides up on a film of water and thereby
loses contact with the pavement. Although many vehicle, pavement and environmental factors
affect the risk of hydroplaning, a rule by thumb is that hydroplaning can be expected for speeds
above 70 km/h where water ponds to a depth of 4 mm or greater over a distance of 10 m or
greater. Thereby, Glennon (2004) [42] states that â
hydroplaning is a function of water depth and
length of the drainage flow path
â.
Gallaway & Rose (1971) [25] found that the pavement water depth (above the road surface tex-
ture tops) can be calculated from:
âą
rainfall intensity,
âą
cross slope,
âą
length of the drainage flow path, and
âą
texture
depth.
In addition, they defined the length of the drainage flow path as a function of:
âą
pavement
width,
âą
cross slope/superelevation, and
âą
longitudinal
gradient.
Pavement depressions (unevenness and rutting
9
) make water ponding worse, while horizontal
curvature increase the exciting lateral force and thereby the demand for lateral friction. This
raises the risk of skidding.
Both of the lists above include cross slope as a key factor for hydroplaning. Despite this, many
road agencies do not analyze cross slopes in their road network on a routine basis!
Some 535 Swedish hydroplaning accidents have been analyzed at the macro level [31]. The
results show that where cross slopes are too low, the risk of hydroplaning more than doubles;
from about 26 to 54 per Mapkm
10
. Due to uncertainty in the position of the accident sites, the
analysis was made using average values over as much as 500 m. This suppresses the influ-
ence of local damage on the road. If the analysis could have been carried out over shorter aver-
age lengths, i e 50 m, it is likely that even larger increases in risk could have been identified, as
cross slopes become too low.
The design value used in Sweden for cross slopes on new straight sections is -2.5 % on roads
with hot mix pavements, and -3 % on pavements with surface dressing as the only bound layer
9
In the EU Northern Periphery, rutting is not only caused by compaction of the pavement, but also by surface
abrasion due to the use of studded tyres in the winter. One exception is the Highlands of Scotland, where studded
tyres are not common and surface friction has instead been handled by using surface dressing pavements.
10
The unit Mapkm (million axlepairkilometer) describes traffic work.
Page 44
ROADEX III The Northern Periphery Research
[40]. These design values for new roads have been set with an allowance for future settlements.
When repaving old roads, where settle ments are likely to have stopped, a slightly smaller CS of
-2 % may be a beneficial target. A CS of -2 % is sufficient for a good water flow, and makes
driving in windy conditions easier. Occupants can also sit more upright and co mfortable than
with larger cross slope. These latter aspects are especially important in providing a sound work
environment for professional drivers.
3.3.2 Relating poor ride to pavement condition
This section starts with a short review of opinions of EU Northern Periphery road users. Then it
describes how ride vibration is affected by many factors, such as road conditions, vehicle prop-
erties and driving behaviour (including driving speed). Of these, road condition is by far the most
decisive for in-vehicle vibration. Various types of road defects cause various ride vibration prob-
lems. Examples of these are long wave unevenness, undesired cross slope variance, rough-
ness, megatexture, potholes and other local damages, particularly deflection variance in weak
pavements under heavy vehicles as shown by Granlund et al (2005) [9]. The section ends with
a description of how to measure ride quality on dirt roads.
3.3.2.1 EU NP professional road users perspective on ride conditions
Opinions of professional road users on road service levels in test areas across the EU Northern
Periphery was mapped by Saarenketo & Saari (2004) [49] in the ROADEX II Project. 330 ques-
tionnaires were issued, and with a satisfying response rate of 45 % the result was 147 answers.
The answers showed that roughness was a major proble m for the forest industry; 70 % of tim-
ber transporters stated that uneven roads were their main proble m. Also over 50 % of respon-
dents in the construction and public industries suffered from severe problems due to roughness.
Truck drivers stated that the worst sections had bumps at culverts, located at the bottom of a
valley with steep hills adjacent to the low point culvert. This situation required the m to slow their
truck down to almost zero, in order to prevent vibration da mage, and once the bump was
crossed, it did not have enough mo mentum to climb the next gradient. The drivers also reported
much higher fuel consumption on rough roads. Many problems were reported to be related to
weak pavement shoulders, poor road alignment and poor bearing capacity. A significant share
of the drivers gave poor traffic safety ratings, due to factors such as poor winter condition (main-
tenance), bad cross slope, uneven frost heave bumps, poor road alignment and lack of crash
barriers in curves on high embankments.
The truck drivers questioned also reported continual stress when driving on some long routes
(including National Highways) that the road agency believed to be in good condition for driving.
This happened when unexpected poor road conditions made the perceived maximum safe
speed drop far below the planned speed. The result was a conflict within the driver, between
making a delayed delivery and causing a major traffic safety risk. Such a conflict caused high
stress to the truck driver. Typical sources of this kind of problem are frost related roughness and
delayed snow removal. The latter allows snow to be compacted and leads to the development
of deep tracks and ruts. Wet snow freezes into ice in the lanes and these ice ruts then re main
for a long time. This type of slippery rut can be difficult to remove with a truck mounted snow
plough and typically they must be scraped off (costly) with a slow-moving heavy grader. Severe
cases may even require many repeated runs with the grader.
Page 45
ROADEX III The Northern Periphery Research
3.3.2.2 Roll vibration is excited by undesired variance of Rut Bottom Cross Slope
The Swedish Road Administration has been laser-scanning the surface condition of their road
network since the 1980âs and importing the data into the SRA Pavement Management System
(PMS). On-going evaluations by Johan Lang of SRA show that the average condition of rutting
and roughness on the Swedish highway road network is quite constant, despite gradually
increased intervals between repaving actions and large yearly increases in traffic. One reason
for this important success is the increasingly systematic use of laser-scanned condition data by
the local pavement engineers [Personal communication with Mats Wendel, SRA Head Office].
The condition parameters most commonly used in PMS so far, are Rut Depth and International
Roughness Index (IRI). These para meters do not work well on road sections with edge dam-
ages although these can cause excessive roll motion to high (heavy) vehicles. Therefore, the
SRA Central Region has in the strategic regional plan for 2004 - 2015 pointed out an urgent
need for a âRoll vibration indicatorâ as a new pave ment condition para meter [20].
At the national level, the SRAâs action plan for traffic safety 2004 - 2015 has identified that
â
Pavement edge deformations are perceived as very uncomfortable by all road user groups, es-
pecially drivers of (high) heavy trucks
â [23]. This confirms the need of a pave ment condition pa-
rameter to address this kind of distress properly.
A potentially suitable roll vibration indicator has been defined âdown underâ. As reported by
Bowler et al (2001) [45], Transit New Zealand (TNZ) has carried out excellent custo mer focused
work in the award winning
Truck Ride Improvement Initiative
. This included a two-stage re-
search process that started with a programme of qualitative research that looked at the specific
concerns of truck drivers. This was followed by a second research stage where the truck drivers
were asked to quantify their concerns. The results of the first stage were used build a list of
concerns that were prioritised by the drivers. Before the final ranking, the truck drivers were in-
formed on the relative costs for each type of improvement.
After some adjust ments for willingness to pay, the 300 truck driversâ top priorities were:
1. Build more passing lanes.
2. Repair surface undulations and settlements.
3. Straighten out too sharp corners.
4. Repair
incorrectly
banked
corners.
5. Improve road alignment and evenness at bridges; i. e. repair settlement on the
approaches to bridges.
6. Build
wider
shoulders.
7. Correct vertical alignments; take away dips and rises which block visibility.
8. Build
wider
bridges.
9. Build longer passing lanes.
Findings from the TNZ âmarket investigationâ was then used in a technological project, defining
how to detect road sections where unevenness significantly impacted on truck ride and han-
Page 46
ROADEX III The Northern Periphery Research
dling. Cenek et al (2003) [43] presented this project, where the focus included long wave undu-
lations and roll caused by roughness warping between wheel-paths. The outco me is that TNZ
now has an awareness of the need to focus on those sections of the highway network that are
of priority to truckers, and on the repair of these. The results also provided justification for ques-
tioning current road management practices and funding allocations, which are not delivering the
types of state highway improvements that professional customers require. TNZ found that cab
body roll, particularly when combined with cab body pitch, was of most concern to occupants of
trucks. This is an important finding, as existing road roughness para meters used for pavement
management purposes (i e IRI) have their emphasis on vertical vibration, not on rotation. The
threshold value for uncomfortable truck ride, related to rotational response (pitch and roll), was
found to be 4.0 to 4.5 °/s. A complex Truck Ride Index was developed from this. It is based on
existing 20 m average values of cross slope, curvature and other parameters available in TNZâs
pavement management system. Since 2001, Transit New Zealand has been given an additional
road funding of NZ$3 million annually that has been specifically allocated to the repair of critical
sections for truck ride.
In 2004, the SRA tested the truck cab roll component from TNZâs co mplex Truck Ride Index us-
ing data from road Y 953 in the SRA Central Region. The results were disappointing however.
The potential roll vibration indicator gave much higher alarms at entrances and exits of normal
left hand curves, than at critical sections with severe edge deformations and roll problems. This
was seen as a defect in the system, since it could lead to a waste of road repair funding. At a
result of this, SRA decided to define a new roll vibration indicator in-house. The new indicator is
based on road profile data, laser scanned at 16 kHz in the botto m of the truck wheel paths (left
and right) and reported in steps no longer than 1 m. This revised indicator offers 20 times better
spatial resolution than the TNZ roll indicator. From the data recovered, the Rut Botto m Cross
Slope (RBCS) is calculated. At this point a crucial filtering procedure is applied, to remove the
very long wave slope variances that relate to superelevation change at curve transitions. This is
markedly notable at left
11
hand curves. Depending on road section width and reference speed,
such desired change in cross slope takes place over some 40 - 200 m. These transitions
smoothly tilt the truck cab roll angle fro m one side to the other without producing roll-mode vi-
bration. The vital filter is calibrated with the roadâs reference speed, thereby normalizing the fil-
tering to typical heavy truck roll vibration eigenfrequencies. In the next step, undesired vari-
ances in the RBCS are calculated. This is done in two parallel runs. One run calculates the vari-
ance over âshort sectionsâ, addressing the excitation of the axle roll of the truck wheel. The
other run calculates the variance over âlong sectionsâ, addressing the excitation of the truck
chassis/cab roll. Finally, the maximum of these two variances is reported as the undesired vari-
ance of RBCS. See Granlund (2006) [21] for details.
One of the three goals in the present ROADEX III project research task, is to draft a limit for the
new âundesired Rut Bottom Cross Slope Varianceâ parameter defined at SRA.
Should the limit
be 0.25 %, 0.50 % or what? Should there be different values in curved vs. straight sections, in
long curves vs. short curves, and in wide vs. narrow sections?
11
In the UK and other countries with left hand traffic, this applies to right hand curves instead.
Page 47
ROADEX III The Northern Periphery Research
3.3.2.3 Frost related bumps and potholes are worst
The ROADEX II professional road user interviews [49] showed that uneven frost bumps and
potholes were considered to be some the worst da mage types on paved roads. These short and
high/deep local damages can cause mechanical shock that can result in damage to vehicle,
cargo and/or vehicle occupants. As can be seen in the following clause, the traditional use of
long report intervals is one of the main reasons why road agencies have not be able to focus on
local damages, despite using sophisticated laser/inertial profilometers.
3.3.2.4 Road condition data must be analyzed over relevant report intervals
An important issue when discussing road condition in relation to drive co mfort as well as health
and safety, are the properties of the used road statistics used; i e the report interval. By
tradition, rutting and roughness values have been described as mean values over long sections
such as 20 m, 100 m, 400 m, 1000 m, whole roads or even whole road networks.
Since roughness is defined as a deviation from a planar surface, it is of course less relevant to
analyze the mean value, than some kind of estimate of the worst deviations. Figure 23 shows
values from a variety of report intervals, ranging fro m 1 dm to 400 m, from an analysis of a very
rough road in the SRA Northern Region. The result shows that local bumps were 20 - 30 ti mes
worse than the average IRI value of 3.8 mm/m, giving peak IRI-values of 80 - 130 mm/ m. This is
comparable to, or actually worse than, many 10 cm high traffic calming speed bumps on urban
streets. At the worst sections of this 90 km/h road, heavy trucks almost bounce off the road at
speeds over 30 km/h. Despite this, roads with IRI lower than 4 mm/ m are not reported in the
SRA annual report as a severe problem. This is may be true for roads with low roughness vari-
ance, but definitely not for roads with severe local damages with IRI
>
80 mm/ m!
3,8
130,2
0,0
3,8
82,3
0,1
3,8
16,0
1,2
3,8
7,9
2,2
3,8
5,8
2,4
0
20
40
60
80
100
120
140
Average
Max
Min
Statistical property of IRI
M
a
gni
tude
of
I
R
I-
v
a
lue
[
m
m
/m
]
1 dm
1 m
20 m
100 m
400 m
Figure 23
Surface roughness IRI for 6.5 km of Rd 374 Vitvattnet â Storfors, 2002
Page 48
ROADEX III The Northern Periphery Research
As illustrated above in Figure 23, roughness variance is extremely high at local bumps. Thus
averaging over distances much longer than the bump itself (often about 1 m), such as with IRI
20
or IRI
100
, disguises the variance. This averaging eliminates the ability to identify those bumpy
sections that heavy vehicle operators consider intolerable. A better para meter could be the 95âth
percentile, such as is used when mapping vibration emission fro m roads and railways to nearby
dwellings. Another option could be to report a parameter related to the variance in data,
together with a mean value. By reporting the mean value, together with the âtwo sigmaâ limit
(corresponding to the 95âth percentile), a better picture is given of the worst sections.
3.3.2.5 Heavy vehicles suffer from soft spots in weak pavements
Heavy vehicles perceive not only the static surface roughness, but also a dynamic roughness
component when the pavement has âsoft spotsâ. Pavement deflection is typically less than two
millimetres under a moving heavy vehicle. This magnitude seems negligible, being co mparable
with road wearing course texture. The texture however, is smoothened by the tyreâs âenveloping
effectâ. Ride comfort is associated with vibration acceleration and vibration velocity, rather than
vibration displacement. (This makes sense; otherwise a stiff sports car would be considered
more comfortable than a soft luxury car when riding on bumpy roads). Vehicle vertical vibration
acceleration is associated with road roughness profile slope variance, rather than roughness
profile height. So, even if a pavement deflection under heavy vehicles, with few exceptions,
would not be larger than about one or two millimetres, significant vehicle vibration acceleration
could occur at soft spots where the deflection profile varies rapidly in terms of large slope vari-
ance. The importance of soft spots is confirmed by the Australian coal mining industry, where
they are recognized by handbook
Bad Vibrations
[50] as an important source of ride vibration in
transport vehicles.
Ahlin et al (2000) [3] made an unexpected observation when comparing road roughness with
ride vibration in ambulances and heavy trucks. When surface roughness drops to zero, signifi-
cant seat vibration remains in heavy trucks while vibration drops to almost zero in ambulance
cars. In the trucks, the threshold of the weighted vibration acceleration was found to be as high
as 0.2 m/s
2
rms. This value is to be compared with the Action Value of 0.5 m/s
2
rms over 8
hours, stated in directive 2002/44/EC. Clearly, other factors other than road surface roughness
can bring as much as 0.2/0.5 = 40 % of the allowed truck seat vibration. Soft spots in the pave-
ment are believed to be a causal factor behind such vibration; Forssén (2001) [10] discusses
road deflection variance as an important but hard-to-grip property. Granlund et al (2005) [9]
measured and co mpared truck wheel vibration with theoretical wheel vibration calculated from
road surface roughness (assuming a perfectly stiff road profile) on 80 km of roads in Sweden.
The hypothesis was that large differences between measured and calculated vibration indicates
possible soft spot locations along the road. The study found a correlation between soft spot in-
dications recorded in the truck and reference data on pavement bearing capacity properties,
such as subgrade stiffness module, pavement thickness, frost fatigue da mage, and overall bear-
ing capacity index. These findings give further support to the theory that significant amounts of
heavy vehicle vibration arise from soft spots in weak pave ments.
Page 49
ROADEX III The Northern Periphery Research
3.3.2.6 Measuring ride quality on very low volume dirt roads and winter roads
Many road agencies measure the condition of paved roads with laser
12
/inertial profilometers, as
with the Profilograph used in the case study of this project. The advanced and expensive
Profilograph can report a wide variety of condition parameters, including longitudinal roughness
in terms of IRI-value and other indices, such as rutting, cross slope and texture. Very low
volume roads do not however require such demanding accuracy and can be measured with
cheaper instruments. It may in fact be impossible to make relevant condition measurements
with a laser/inertial profilo meter on very poor condition dirt roads or icy winter roads.
A relatively cheap measurement method available is to ask road users about their perceived
ride quality; the so-called âno cost instrumentâ. By using a comfort scale, such as in Table 1,
âbackwardsâ, it is possible to estimate the vibration intensity and use it as a condition rating. An
obvious problem however is how to distinguish between transient shock at bumps and average
vibration by roughness.
A somewhat more expensive system can be based on vehicle seat mounted vibration sensors;
the âmedium cost instrumentâ. Two examples are the CVK Health Vib (see Figure 24) and the
Bruel & Kjaer Human Vibration Measurement Kit (see Figure 25). A more advanced state-of-
the-art example is the Dewetron Stream Machine, used in the case study and presented later in
this report. The price of such instruments ranges from a few thousand âŹ, up to some thirty thou-
sand âŹ. Then there are costs of extra sensors such as odometer, GPS, video and others. There
have been experi ments trying to measure road condition with accelerometers that have been
mounted to wheel axles. One obvious drawback is that such a system does not yield results
comparable with either the ISO 2631 comfort scale, or the EU Action Value for professional
drivers WBV exposure.
Using a cheap instrument however is no guarantee of a cheap measurement. A full
measurement process includes activities such as data collection, transfer, storage, backup,
analysis, quality control and distribution to users. There are also costs for client system
infrastructure, user education, and much more. So a low total cost may depend more on smart
purchase behaviour, than on cheap sampling with low precision (subjective comfort rating) and
limited outcome.
âCheapâ sampling can have drawbacks for road managers, particularly those who purchase
road maintenance based on road condition. If road users start to give biased ratings in order to
achieve âabove standardâ conditions, such measurements may become very expensive in a
road maintenance contract context. Alternatively, if the vast majority road users see a particular
bump and always brake for it, the vibration will be reduced and an on-board logger will not
trigger an alarm. In this case the bump will not be repaired. Such bumps however may come as
a hazardous surprise to foreign road users not familiar with the road.
12
The sensor recording the height above the pavement may be of another type other than a laser sensor; i e an ul-
trasonic sensor. However, laser sensors have proved to be able to better fulfil the high accuracy and high environ-
ment demands associated with road profiling. A current trend in the road profiling industry is to scrap cheaper sen-
sors and replace them with rugged âroad editionâ laser sensors.
Page 50
ROADEX III The Northern Periphery Research
Figure 24
The CVK Health Vib system
Figure 25
Br
Ʊ
el & Kjaer Human Vibration Measurement Kit 4447
Page 51
ROADEX III The Northern Periphery Research
Chapter 4. Case study on the Beaver Road 331
Figure 26
The southern entrance to the Beaver Road 331
Rd 331 is a 170 km long regional route in Sweden, connecting the rural forest area in eastern
JĂ€mtland County and western VĂ€sternorrland, with the heavily industrialized coast at the east of
VĂ€sternorrland County as seen on the map in Figure 27.
Rd 331âs Annual Average Day Traffic (AADT) ranges from 350 to 2000 vehicles per day.
Rd 331 is a main supply road for timber transports servicing the paper mills in the Sundsvall
area, such as SCAâs factories in Tunadal, Ăstrand and Ortviken. Thus, the share of heavy
trucks is very high, from 12 to 19 %. Most of the trucks have three axles and a trailer; the
average number of axles per recorded truck is as high as 4.8. The timber transports on Rd 331
are also expected to increase by another 150 000 m
3
/year. This is due to a redirection of timber
from the SollefteÄ area, which up until now has been transported to Utansjö Bruk (being shut
down), north of HÀrnösand.
Page 52
ROADEX III The Northern Periphery Research
Figure 27
Swedish Beaver Road 331 connects the forest area with the industrialized coast
Connecting National Highway 87 in Viksmon to the Coast, Rd 331 also is a link on the
ambulance route between the local hospital in SollefteÄ and the region hospital in Sundsvall.
Among the most sick or severely injured patients, so me are in need of specialized care not
available in SollefteÄ. These patients, often very vulnerable, are transported on this road to
Sundsvall.
The speed limit alternates between 90 and 70 km/h, with a drop to 50 km/h in so me villages.
Rd 331 is considered a normal road, while SRA have thousands of kilometres of roads in si milar
condition and use. Therefore results of this case study are therefore not unique, but relevant to
a large proportion of the road network.
In the SRA Road Data Bank (RDB), Rd 331 starts at the Coast and its distance is measured
towards JĂ€mtland. In this study, however, most of the analyses have been made in the direction
of the timber transportation. This direction is opposed to the RDB distance direction. For this
reason, most of the data graphs presented have the distance markings in the reverse direction.
Rd 331 suffers from many and severe traffic accidents. In 2005, seven people were killed in
road traffic accidents on the road network in VĂ€sternorrland County. Three of them died on Rd
331. A map over police reported serious accident black spots on Rd 331 is showed in Figure 28.
The site with highest accident rate in VĂ€sternorrland County is the Hazardous Site Stavreviken,
âHS Stavrevikenâ, at the southern exit of Rd 331. The map has been created by Hans Johans-
son, traffic safety officer at SRA Central Region.
Page 53
ROADEX III The Northern Periphery Research
Figure 28
Accident black spots at the Beaver Road 331, not normalized to AADT (Individual Risk)
HS Roos Curve
HS S Viksmon
HS AlderÄn
HS à sÀng
HS Viksjö
HS BjörknÀset
HS N Ă kroken
HS MjÀllÄsen
HS MeÄstrand
HS à kerö
HS Backe (Edsele)
HS Stavreviken
HS Helgum
Page 54
ROADEX III The Northern Periphery Research
4.1 TRUCK TEST PARTNER - BRORSSONS Ă KERI AB
Brorssons Ă
keri AB was founded in the mid 1940âs, at the end of World War II. During the eras
of water power plant construction from 1940 to 1960, the company was heavily employed trans-
porting soil, gravel and other material to the construction sites. After the end of the era, the
transport markets changed and for the last twenty years timber logging has been Brorssonsâ
core business.
The company operates 14 timber logging trucks, each with a large trailer. The trucks are loaded
by separate cranes (see Figure 29), to minimize the dead weight of the vehicle and maxi mize
payload. Each truck runs Monday to Friday in two shifts resulting in 18 hours per day. On Fri-
days, only an 8 hour shift is used. Normally each truck daily drives 4 round trips of some 2 x 140
km on the Beaver Road 331, depending on which forest the timber is to be picked up from. The
annual mileage per truck is 200 000 km. All trucks and trailers are exchanged at 3 to 4 years of
age.
Figure 29
Logging timber from forest to the coastal industries
The co mpanyâs vehicles are seldom involved in traffic accidents, other than some low speed
trailer incidents on narrow, steep and slippery forest roads. As a result, the company has a
modest insurance cost for the truck fleet. However, the drivers are very uncomfortable with
seeing foreign road users suffer from accidents at âHazardous Sitesâ. The drivers think that
many of these accidents could have been prevented. The drivers have requested road
improvements including increased width of the narrow, high and steep road banking from
Viksmon to Stavre, straightening the Roos Curve and some other sharp curves, repair of
incorrectly banked curves, repair of edge deformations and bumps at culverts, more frequent
and higher quality resurfacing, and intensified winter road maintenance such as frequent
removal of ice-ruts with a heavy grader. These requests have been raised by professional
drivers riding an astonishing total of 2 800 000 vehicle km/year on Rd 331.
Page 55
ROADEX III The Northern Periphery Research
4.2 A SCANIA R480 164 G 6X4 W AS USED AS TEST TRUCK
The test truck (licence registration number WPT 493) was three years old, and had a mileage of
609 000 km. The instrumentation was carried out on Sunday 26
th
of August, 2007.
Figure 30
A damper bush was broken, but this was not perceived under normal driving
When de mounting one of the truckâs front wheels, see Figure 30, it was found that the damper
bush was broken. None of the two experienced professional drivers, Mats Jonsson and Anders
Larsson, had noticed anything unusual during their daily driving work. The truck had been at the
Scania Workshop for service just the week before and had to be driven back for the necessary
repair work on Monday morning, before the ROADEX main test. Measurements were carried
out on this journey for reference purposes only.
The gross vehicle weight of the test vehicle combination was 60 tonnes with a full payload. The
dead weight was so me 19 tonnes. Measurements were taken under normal working conditions;
hauling timber from the forest to the coast, and then driving back unloaded. If nothing else is
stated, the data in the graphs presented are recorded with the vehicle combination fully loaded
and travelling to the coast.
The drivers were instructed to drive as they would normally do, with one exception. They were
asked to remain in the wheel tracks, and not to avoid driving over local road damages. (In
normal driving, they can avoid some bumps unless there are oncoming or overtaking vehicles).
Page 56
ROADEX III The Northern Periphery Research
4.3 TEN ROUNDTRIPS OF 280 KM WERE RECORDED
A total of ten 2 x 140 = 280 km round trips from the Ramsele forests to the coast were recorded,
in order to investigate the precision in truck ride measure ments. One trip also included the most
northwest 30 km section from Backe to Ra msele.
The round trips were carried out from Monday 27
th
to Thursday 30
th
of August 2007. Many of the
measurement trips were undertaken in rainy weather and on wet roads.
Driving a timber logging truck is extre mely busy. In the forests, the truck drivers get a short
break while the timber is loaded by a large separate crane as seen in Figure 29. At the coastal
delivery points, there are almost no natural pauses at all. The huge Svetruck log stackers with
their 8.2 m
2
grapple need only single grips to unload each of the three ti mber piles on the truck
and trailer combination. The driving shifts change at 14.00. This is done without even a few
minutes co mmon break for a cup of coffee together.
The detailed analysis focused on a series of Hazardous Sites (HS) identified from truck driver
interviews, fro m pave ment condition data and from crash statistics in the Swedish national road
traffic accident database STRADA.
Page 57
ROADEX III The Northern Periphery Research
4.4 COMPREHENSIVE RIDE AND ROAD CONDITION MEASUREMENTS
4.4.1 Truck ride measured by accelerometers and a combined GPS/inertial unit
Truck ride vibration was measured with a Dewetron Stream Machine system, owned by SRA
Consulting Services (SRA CS). This syste m carries out real time calculations of health risk, as
defined in the recent ISO 2631-5 (2004) standard [5]. The system, and the connected acceler-
ometers, satisfies the comprehensive accuracy de mands set out in the EN ISO 8041 instrumen-
tation standard [17]. The accelerometers were located at several points of the truck to record
motion in multiple directions (see Figure 31 for definitions and Figure 32 for photos of mount-
ing):
âą
Left and right front axles, z-axis, 5 kHz
âą
Left and right side of the frame, above the front axle, z-axis, 5 kHz
âą
The pan of the air-suspended drivers seat, xyz-axes, 5 kHz
Both drivers weighed around 90 kg.
Page 58
ROADEX III The Northern Periphery Research
Figure 31
Basicentric axes of the human body in seated position [18]
Figure 32
Left: Accelerometers at axle and frame, Right: Seat pad with 3-axis sensor
The vehicle speed during the tests was recorded with a GPS/inertial unit, updating speed values
at 100 Hz. This corresponds to measuring every second dm of the ride at 80 km/h. This data
was also used to calculate the distance position with a fair accuracy.
A digital video ca mera recorded front view âRight-Of-Wayâ (ROW) from the truck cab. Noise < 5
kHz was recorded with a microphone.
Page 59
ROADEX III The Northern Periphery Research
One of the main tasks of the test was to study roll vibration of the truck cab. For this purpose, an
OxTS RT 3050 100 Hz GPS/inertial unit was used, see Figure 33. This unit recorded the motion
of the cab in all 6 axes; xyz translation, as well as rotation in yaw, roll and pitch. The accuracy
and resolution was so good, that the system was able to pick up a change in elevation of 1 mm
between the left and right truck tyres road contact patch. The RT 3050 was mounted on a car-
bon-reinforced RT Strut, with very high torsion stiffness, seen in Figure 34.
Figure 33
The OxTS RT3050 GPS/Inertial unit, used for 6-axis ride measurement
Figure 34
The RT 3050 mounted on a RT Strut in the truck cab
Page 60
ROADEX III The Northern Periphery Research
4.4.2 Passenger car ride measurement
For reference purpose, vibration was also measured on the driver seat in the new Ford Mondeo
passenger car seen in Figure 35. The driverâs weight was about 90 kg. The car was driven just
below the speed limit, with a minimum of speed variance.
Figure 35
Reference driver seat vibration measurements in a new Ford Mondeo
Page 61
ROADEX III The Northern Periphery Research
4.4.3 Laser/inertial reference measurement of pavement condition
In this case study, the road alignment, the 3-D geometry of the pave ment lanes and the surface
texture of the road were scanned with one of SRA CSâs advanced laser/inertial Profilographs, as
shown in Figure 36. These Profilographs are used for routine survey of the condition of paved
public roads, airfields, test tracks et cetera. The resolution of the system is 0.1 mm (texture 0.01
mm). The accuracy expressed in terms of precision and trueness, is within fractions of a milli-
metre under normal operation conditions, as certified by third party. The Profilograph allows ac-
curate inertial compensated measurements to be gathered whilst driving at speeds up to 165
km/h, although speeds of 15 to 90 km/h are more normally used in highway surveys.
Figure 36
SRA CSâs laser/inertial Profilograph P45 [Photo: Mats Landerberg]
The Profilograph is equipped with a 2.5 m wide rut bar, as seen in Figure 37. The rut bar is
equipped with 16 kHz lasers scanning the road surfaceâs shape relative to a large scale inertial
plane. The two outermost lasers on each side are angled outwards, giving rise to a total
scanned lane width of 3.2 m. Three of the lasers sample at 64 kHz, taking accurate measure-
ments of the road surface texture. One scans the left wheel path, one the right, while the last
texture laser scans between the wheel paths.
Page 62
ROADEX III The Northern Periphery Research
Figure 37
The Profilograph with its 2.5 m wide rut bar, scanning a 3.2 m wide lane cross section
Page 63
ROADEX III The Northern Periphery Research
Chapter 5. Expected results are confirmed
5.1 UNACCEPTABLY HIGH WHOLE-BODY VIBRATION AND SHOCK
5.1.1 Daily vibration exposure exceeds the EU Action Value
It is impossible to define a long term representative daily vibration exposure to 2 decimal places
for professional truck drivers. An obvious reason, for Brorssons drivers at least, is that they pick
up ti mber at various places. Every week they drive not only Rd 331 between the coast and
Ramsele, but also other connecting local and forest roads in the Ramsele area. They drive at
different speeds on roads with varying roughness. The result is various vibration intensities.
Given this co mplexity, a set of calculations have had to be made in order to consider the various
driving routes. This analysis has been carried out using the Vibration Doses Calculator, avail-
able on the UK Health and Safety Executiveâs website
13
. Normal shifts with roundtrips from for-
est to coast, resulted in A(8) values from 0.65 m/s
2
and higher. Some of the forest roads outside
Ramsele were very rough, but since speed was low and the driving ti mes on them were si milarly
low, their contribution to the total daily exposure was lower than the main partial exposure from
the long round trips on Rd 331 between Ra msele and the coast. Figure 38 shows the resulting
A(8) of 0.76 m/s
2
for an 8 h shift example, including simulation of pauses with zero vibration.
Figure 39 shows an example of calculation details.
Daily Vibration Exposure A(8)
Scania R480G 6x4 timber logging truck
8 h morning shift: Ramsele - Backe (empty), 2 x Forest Road (e & l), Backe - Stavre - Ortviken
(loaded), Ortviken - Ramsele (e), other local and forest roads (e & l), and 100 min pauses.
0,50
0,76
0,00
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
V
ibr
a
ti
o
n [
m
/s
2
]
Measured A(8)
EU Action Value
Figure 38
The drivers daily exposure to vibration exceeds the EU Action Value
13
Internet:
www.hse.gov.uk/vibration/calculator.htm
Page 64
ROADEX III The Northern Periphery Research
Vibration intensity
Partial exposure
m/sÂČ
hours minutes
m/sÂČ
Ramsele - Backe
Empty
1,66
0
25
0,382
Backe - Ramsele
Loaded
1,32
0
28
0,316
Ramsele-Ăsterforsse
Loaded
0,54
0
43
0,160
Ăsterforsse - Viksmon
Loaded
0,66
0
7
0,080
Viksmon - Stavre
Loaded
0,66
0
56
0,226
Stavre - Tunadal
Loaded
0,44
0
9
0,061
Tunadal - Stavre
Empty
0,56
0
9
0,076
Stavre - Viksmon
Empty
0,83
0
56
0,285
Viksmon - Ăsterforsse
Empty
0,83
0
7
0,100
Ăsterforsse - Ramsele
Empty
0,58
0
40
0,168
Forest Road
Empty
0,80
0
22
0,172
Forest Road
Loaded
0,64
0
22
0,137
Misc roads, average intensity E & L
0,79
0
58
0,276
Pause, non-driving time
0,00
0
100
0,000
0,76
Exposure time
Daily exposure value, m/sÂČ A(8)
Figure 39
Vibration Dosis Calculator spreadsheet, calculating A(8) for an example route
In accordance with the ISO 2631-1 standard, the seat vibration was measured in three
directions; x (fore-aft), y (lateral) and z (vertical). The measured vibration was high in all these
three axes. The EU vibration directive states that the daily exposure value A(8) shall be
calculated from only the axis with highest vibration. Furthermore, the values for lateral (y) and
fore-aft (x) vibration shall be multiplied by 1.4, since vibration in these directions are considered
to be unhealthier
14
than vertical vibration (z). On ânormalâ roads, the vertical axis typically has
highest vibration. However, on some sections of 331, the lateral axis had the highest vibration
(after multiplication with the 1.4 factor).
All the daily exposures calculated were significantly above the EU Action Value of A(8) = 0.5
m/s
2
. This finding is very serious. The law now calls for the employer, Brorssons Ă keri AB, to
take necessary technical and/or organizational actions to minimize the driverâs exposure to
vibration. In fact, all co mpanies with similar trucking operations and conditions to Brorssons
(long and bumpy driving) are obliged by the law to make a relevant risk assessment of the
driversâ vibration exposure.
14
The factor 1.4 is only used for health risk assessment. For comfort, the factor is 1 which means no extra weighting.
Page 65
ROADEX III The Northern Periphery Research
5.1.2 Some of the worst bumps gave spinal compression stress S
ed
over 0.5 MPa
Transient vibration (mechanical shock) may cause high compression stress in the spine. This
health risk is presented in the section â3.1.3 Bumps are of special concern to both ride quality
and healthâ.
The worst bumps in the current tests were located on small roads, such as on the road to the
Sawmill in Graninge. When driven at low speeds of about 40 km/h, these bumps exposed the
truck driver to spinal co mpression stress S
ed
over 0.5 MPa. This stress level corresponds to a
health risk, as per ISO 2631-5 [5].
Also on the âmain roadâ, Rd 331, truck drivers drove over many bumps that excited significant
transient vibration. The first bump the drivers faced in the morning was only 400 m fro m Brors-
sons garage at the western exit fro m Ramsele. As seen in Figure 40, the bump was so deep,
that it had rubber marks made by retracted, non-rotating, truck bogie tyres, similar to the marks
seen in landing zones on airfield runways. This bump was due to settle ment at an old culvert
and had been present for many years.
Figure 40
Bump due to settlement at an old culvert in Ramsele, RDB section 141/336 km
Page 66
ROADEX III The Northern Periphery Research
Rd 331 had not only severe bumps due to old culverts, in some sections there were even worse
bumps at newly reconstructed culverts. One example was in Gammelmo, 7 km south of Ram-
sele. A photograph of this site can be seen in Figure 41. The section in Gammelmo is similar to
one of the most stressing driving conditions perceived by EU NP professional drivers: â
Truck
drivers stated that the worst sections have bumps at culverts, located at the bottom of a valleyâ.
Figure 41
Bumpy newly reconstructed culvert in Gammelmo, RDB section133/000 km
When driving at normal highway speed over this new culvert, the driver was exposed to âvery
uncomfortableâ transient vibration, as seen in Figure 42.
0,0
0,2
0,4
0,6
0,8
1,0
1,2
1,4
1,6
1,8
2,0
132900
133000
133100
133200
133300
133400
133500
133600
133700
Di stance (a pprox ) [m]
R
id
e
vib
ra
tio
n
[
m
/s
2
]
Seat vibration, ISO-weighted
Uncomfortable, ISO 2631-1
Figure 42
Very uncomfortable truck seat vibration when driving over the new culvert in Gammelmo
Page 67
ROADEX III The Northern Periphery Research
The cause of the transient vibration in the driverâs seat can be seen in the following 3D laser
scan from the Profilograph. Figure 43 shows a 20 m long and 57 mm deep hollow in the newly
paved asphalt over the culvert. The culvert had been reconstructed just a couple of months be-
fore the test in August 2007.
57 mm / 20 m
hollow
Figure 43
3D laser scan of settlement at reconstructed culvert in Gammelmo
An example of a series of extremely annoying bumps was found 5 km south from Edsele. At this
location there were three very bumpy culverts in a row within 200 m. When these culverts were
crossed by a heavy truck travelling at around 80 km/h, the result was repeated transient seat
vibration. The in-truck measurements recorded powerful shocks in the vertical, pitch and fore-aft
directions at all three bumps, as seen in Figure 44. The first bump is indicated by skid marks
from truck tyres, as seen on the Right-Of-Way video in the figure.
Figure 44
Three bumpy culverts within 200 m, RDB section 117/200 km
Page 68
ROADEX III The Northern Periphery Research
Two of the three bumps at RDB 117/200 km are shown below in a 113 m long 3D laser scan in
Figure 45, taken by the Profilograph. The driving direction is from the left to the right of the
graph. These bumps were up to 50 mm deep; a magnitude co mparable to the suspension com-
pression stroke of a normal road vehicle. With such a bump there is a high risk for the suspen-
sion to hit its bump stops, causing a non-linear shock. Obviously this kind of severe road obsta-
cle can be hazardous when driving at highway speeds.
Figure 45
3D laser scans of two bumpy adjacent culverts, RDB section 117/200 km
Page 69
ROADEX III The Northern Periphery Research
5.1.2.1 One of the worst bumps was found on the guest section of National Highway 87
On the short section between Viksmon and Ăsterforsse, Rd 331 is a âguest roadâ on Hw 87. It
was surprising to see that this National Highway section gave as high truck seat vibration inten-
sity as the long section Stavre - Viksmon on Rd 331, see spreadsheet in Figure 39. However,
the worst roughness on the short Hw 87 section could be reduced very efficiently. This can be
accomplished at low cost, by the repair of the bumps at a high banking over the culvert about 1
km north of Viks mon. Profilograph data in Figure 46 shows a large bump with an IRI
20
= 6.4
mm/m just above the culvert, and another bump with IRI
20
= 7.6 mm/ m at the poorly finished as-
phalt joint, just one hundred meter later.
An IRI
20
of 7.6 mm/ m is comparable to the IRI measured with a Profilograph on the 1 dm high
speed bump in front of UmeÄ Plaza Hotel. At that bump, Mrs Gunhild Högberg from
Ărnsköldsvik was severely injured by a spinal compression fracture, when riding with her
husband at low speed in their campervan.
As seen in Figure 46, the section on Hw 87 reminds of one of the most stressing driving condi-
tions perceived by EU Northern Periphery professional drivers: â
Truck drivers stated that the
worst sections have bumps at culverts, located at the bottom of a valleyâ
[49].
Hw 87, Viksmon - Ăsterforsse
Bump at culvert (reconstructed in wintertime), RDB = 77 340 m
0
1
2
3
4
5
6
7
8
77000
77100
77200
77300
77400
77500
77600
77700
77800
77900
78000
RDB-distance [m]
R
o
a
d
r
o
ughn
ess,
I
R
I [
m
m
/m
]
-1,0
-0,8
-0,6
-0,4
-0,2
0,0
0,2
0,4
0,6
0,8
1,0
Long
it
udi
na
l G
ra
d
ie
nt
[
%
]
IRI, average over 20 m updated every 1 m
Road Gradient
Figure 46
Bump at culvert in the bottom of a valley, where grade change from downhill to uphill at 0 %
The culvert beneath the high banking on Hw 87 was undermined in the winter of 2003 and an
emergency reconstruction was carried out to make the road serviceable again. The emergency
repair was done at temperatures below zero; therefore good co mpaction at optimum moisture
Page 70
ROADEX III The Northern Periphery Research
content was impossible. The present roughness is obviously not acceptable in the long-term
perspective, as can be seen in the results. This kind of winter repair must to be finished by a
second phase in summertime, after the inevitable settlements have happened. The poor winter-
paved asphalt should then be milled off, the base watered and co mpacted thoroughly, uneven-
ness s moothened and finally repaved with new asphalt. The smoothness over such recon-
structed culverts should be systematically monitored for two years, and repeated repaving ac-
tions ordered if needed.
5.1.2.2 Bumpy joints at the bridge over Fax River in Helgum
Transient vibration can also be caused by bumpy bridge joints. A 3D laser scan from Ha zardous
Site Helgum is seen in Figure 47. The joints on both sides of the bridge are tenfold times
rougher than the SRA tolerance. When timber logging trucks pass the > 4 cm bumps, snow and
mud fall off the vehicles, contaminating the road and requiring intensified ploughing.
Figure 47
3D laser scan of a bumpy joint at the bridge over the Fax River at HS Helgum
Page 71
ROADEX III The Northern Periphery Research
5.2 THE TRUCK SUSPENSION SYSTEMS PERFORMED VERY WELL
The measurements recorded during the tests permitted the performance of the vertical suspen-
sion systems of the Scania test truck to be evaluated. Figure 48 presentâs data from a 13 minute
ride at 78 km/h over 17 km of Rd 331, southbound from the junction with National Highway 87 in
Viksmon. The vertical vibration intensities recorded are plotted over frequency, using a log-log
scale. This figure does not represent response functions, so a similar figure fro m another road
section will differ somewhat, depending on the properties of the particular road profile.
10
0
10
1
10
2
10
-6
10
-5
10
-4
10
-3
10
-2
10
-1
10
0
10
1
Frequency [Hz]
Po
w
e
r/
fr
e
q
u
e
n
c
y
PSD of unweighted vertical acceleration
Axle
Frame
Cab
Seat
Figure 48
Power Spectral Density of vertical acceleration in the trucks suspension systems
Page 72
ROADEX III The Northern Periphery Research
5.2.1 The wheel axle had much 11 Hz vibration from 2 m wave roughness
The blue trace at the top of Figure 48 shows the vibration recorded at the wheel axle. There is a
wide vibration maximum at about 11 Hz, and high clear resonance peaks at about 45 Hz, 60 Hz
and wide maximums below 2.7 Hz. The lack of a clear peak at the 11 Hz maximum may be due
to superposition of several peaks. One such peak is the wheel axle parallel hop resonance; an-
other is the wheel axle tramp (roll) resonance. The 45 Hz peak may be related to the tyreâs 1
st
eccentric modal resonance. As the wheel is close to the vibration source (the road), Formula 1
can be accurately used to relate the 11 Hz frequency with maximum intensity to road roughness
with wavelengths of 2 m. The other peaks relate to 0.37 m, 0.5 m and 8 m. The intensity at the
45 Hz vibration peak is equal, or below, the intensities from 6.5 to 20 Hz. This bandwidth corre-
sponds to a road roughness ranging from 1.1 to 3.3 m. This waveband, 1.1 to 3.3 m, is obvi-
ously perceived by the truck wheel axle as the worst roughness in the road section.
5.2.2 The frame had much 1.2 Hz vibration from 18 m wave roughness
The green trace, second fro m the top, shows the vibration recorded at the truck frame. The
maxi mum is at so me 1.2 Hz, with more intense vibration than at the wheel axle. This
amplification is likely to be due to resonance in the chassis suspension system, and relates to
road profile waves with some 18 m length. The second peak is at about 2.7 Hz, and a third at 5
Hz. These peaks relate to 8 m and 4.3 m. The highest intensities are seen from 0.7 Hz up to 3
Hz. This shows that the truck fra me perceived the waveband from 7 to 31 m as the worst
unevenness in the road section.
5.2.3 The cab suspension system gave good isolation at high frequencies
The red trace, third from the top in Figure 48, shows the vibration in the truck cab
15
. Just as in
the fra me, the maximum is at 1.2 Hz, and is related to road profile waves with some 18 m
length. At frequencies above 4 Hz, the cab vibration is much lower than the frame vibration. Be-
low some 2.7 Hz, the cab suspension syste m a mplifies the fra me vibration. Similar to the fra me,
the truck cab perceived the waveband from 7 to 31 m as the worst unevenness in the road sec-
tion.
5.2.4 The seat suspension isolated high frequency vibration further
The purple trace, at the bottom, shows the vibration on the truck seat. Just as in the cab and in
the fra me, the maximum is at about 1.2 Hz. The seat suspension isolates vibration over some
16 Hz very well, as it is designed to. At frequencies between 4 and 16 Hz, vibration from the cab
seems to be amplified, getting higher on the seat pan. However, the RT 3050 truck cab
reference sensor was not mounted under the driver seat, but between the two seats. Therefore
data from the cab (input) and the seat pan (output) must be co mpared with care.
15
The cab vibration data are measured with the OxTS RT 3000 system, sampling at âonlyâ 100 Hz. This cause some
aliasing errors at high frequencies, so with respect to the sampling theorem, cab vibration data at frequencies above
50 Hz are not reliable. These data are of no practical importance for the research objectives in this case study.
Page 73
ROADEX III The Northern Periphery Research
5.2.5 Altogether, the truck suspension systems gave excellent vibration isolation
The vibration âtransmissibilityâ from the wheel axle to the driver seat is shown in Figure 49. An
amplification (gain) of â1â means âwhat comes in, gets outâ; neither isolation, nor amplification.
At frequencies over 10 Hz, the truck suspension systems have together isolated more than
99 % of the wheel axle vibration fro m reaching the driver seat. This is of course an excellent
performance. Vibration at frequencies from 3 to 10 Hz has been isolated with efficiency from 0
up to 99 % as the frequency increases. At the âslowâ frequencies below 3 Hz, amplification
makes the driver seat vibration reach up to 2.5 times the wheel axle vibration.
10
0
10
1
10
2
10
-4
10
-3
10
-2
10
-1
10
0
10
1
Frequency [Hz]
A
m
pl
if
icati
o
n
Ratio of Acceleration Seat/Axle
Figure 49
Gain of vibration from wheel axle to driver seat
5.2.5.1 The first problem is low frequency vibration, due to long wave road unevenness
As seen above, the worst ride vibrations are fro m 0.7 to 3 Hz, at 78 km/h related to road
unevenness within a waveband of 7 to 31 m.
Page 74
ROADEX III The Northern Periphery Research
When calculating the drivers exposure to WBV, seat vertical vibration at frequencies below 2 Hz
are weighted (reduced) by a factor 0.5 or smaller, as per the W
k
filter in ISO 2631-1 [18].
Despite this 50 % reduction by the frequency-weighting, very high intensities still re main at
frequencies between 0.7 and 3 Hz.
For the 17 km section south of Viksmon, the root-mean-square (RMS) âaveragedâ value for the
weighted xyz vector was 0.86 m/s
2
. This rates the 13 minute ride âuncomfortableâ on average,
as per the ISO 2631-1 comfort scale in Table 1.
5.2.5.2 The second problem is the intense lateral vibration
Calculation of the 13 minute rideâs contribution to the daily exposure A(8) is based on the single
axis having highest RMS. This was the z-axis, which had 0.59 m/s
2
. When analyzing health risk,
lateral vibration must be multiplied by a factor 1.4 [2]. After this operation, the y-axis had almost
as much vibration as the z-axis; 0.54 m/s
2
. This is remarkably high, compared to the vertical vi-
bration. âOff-roadsâ such as Rd 331 calls for a new approach to truck suspension systems. It is
obviously not enough to isolate vertical vibration; there is also a need to prevent or isolate lat-
eral vibration as well.
5.2.5.3 The third problem is the transient bumps
The worst bump in the 17 km long section southbound from Viksmon, gave a maximum tran-
sient vibration value of 2.44 m/s
2
along the XYZ vector (MTVVV). This value, calculated after
integration over 1 second, corresponds to an âextre mely unco mfortableâ ride on the ISO 2631-1
comfort scale in Table 1.
5.2.6 The broken truck suspension bush had no significant effect
The first truck test run was made without a bush on one damper in the chassis suspension sys-
tem, as seen in the photographs in Figure 30. Before the main test runs, the bush was replaced
at a Scania workshop. The truck driverâs seat vibration has been compared with and without the
bush for the first 30 km section southbound from Ramsele. The results show that the vertical (z-
axis) vibration was 3.3 % higher with the bush in place. This is within the reproducibility noise
level, so it should not be taken as a working bush makes seat vibration worse. Rather, the re-
sults show that the effect of the da mper is low. However the truck chassis suspension system
provides much of its damping by other means than the âda mperâ component.
5.2.7 The lateral vibration was 124 % higher in the truck than in the car
The average xyz vibration on the Scania truck driverâs seat was 83 % higher than the vibration
on driverâs seat in the Ford Mondeo (see photograph in Figure 35), when comparing data from
Viksmon and 17 km towards Viksjö. While the truck ride was âuncomfortableâ, the 0.47 m/s
2
car
ride was only âa little uncomfortableâ as per the ISO comfort scale in Table 1. While the worst
bump was âvery unco mfortableâ (MTVVV = 1.48 m/s
2
) in the car, the worst bump was âextre mely
uncomfortableâ (2.44 m/s
2
) in the truck. These findings confirm the indicative preferences given
in Table 2.
Page 75
ROADEX III The Northern Periphery Research
As expected, the high Scania truck was more prone to RBCSV, since it showed 124 % higher
average lateral (y-axis) vibration than the lower Ford Mondeo passenger car. The ratio between
truck and car was 34 % higher on lateral vibration (y-axis) than on vertical vibration (z-axis).
This further confirm that deformed pavement edges are a much larger problem to truck drivers
than to car drivers, (including Councillors and road agency officers in their comfortable duty
cars).
Page 76
ROADEX III The Northern Periphery Research
5.2.6 Rd 331 canât be efficiently repaired by traditional asphalt overlay
The vibration intensity of a truck wheel axle can be very high, fro m 0.7 to 3 Hz in Figure 48. At
78 km/h, these frequencies correspond to road roughness with waves ranging from 7 to 31 m.
The peak vibration in the frame, cab and seat occurs at 1.2 Hz, which corresponds to road pro-
file waves of approxi mately 18 m lengths.
These truck responses show a ride problem that is related to long wave unevenness in the road.
This finding is confirmed when analyzing the same road profiles scanned by the laser/inertial
Profilograph. A typical road profile example from the 17 km section is given in Figure 50. It re-
cords up to 60 mm deep hollows in wavelengths of over 30 metres.
Figure 50
Rd 331 Viksmon - Viksjö: Unevenness with high amplitudes at up to over 30 m long waves
The SRA Central Region plans to carry out a traditional asphalt overlay for this road section in
2008. However, its steep 7 - 31 m waves are obviously too long to be efficiently repaired by a
simple asphalt overlay. These waves are so long that the paving equipment will simply ride
along them, only raising the unevenness by the thickness of the new asphalt mat. To produce a
good solution, the road machines must be effectively controlled and forced to make the neces-
sary changes to the unevenness, as the present defects are much longer than the machines
themselves. It can be claimed that these waves can be repaired by subjective spot fillings be-
fore paving the mat. However, this ad-hoc method is unable to make the alterations necessary
without using an excessive a mount of costly asphalt. A proper and cost-effective repair of this
road requires an accurate measurement of the road 3D-geometry, and a careful (computer
aided) rehabilitation design at each 5 m section. The benefits in terms of the reduced low fre-
quency ride vibration of this type of road repair method are presented in detail by Granlund &
Lindström (2004) [13]. Another alternative is a more costly âtotal pavement reconstructionâ.
Unless one of these two methods is used, much of the low frequency vibration is likely to re main
for heavy vehicles after the road repair has been carried out.
Page 77
ROADEX III The Northern Periphery Research
5.3 GOOD FIT BETWEEN PROFILOGRAPH DATA AND TRUCK RIDE
5.3.1 Precision of repeated truck ride
There are of course variances between truck rides on a given road section. The reproducibility
for the seat vibration between two runs with truck driver A and a third run with driver B is shown
in Figure 51. The graphs are not perfectly synchronized in distance, due to slightly different lat-
eral position in curves et cetera. Despite being instructed to follow the ruts, it appears that one
of the drivers may have been more active in steering to avoid driving over bumps. This is very
human, since it is easy to revert to periods of ânormal driving behaviourâ during test driving over
a number of days. The graphs show clear differences between good and poor road sections,
even though the variance in a given section can be significant. It will be recalled that si milar
variances were seen when comparing truck ride data with the reference road profile data from
the Profilograph. However, by filtering data from repeated truck runs, variance can be reduced.
Figure 51
Reproducibility in truck driver seat vibration between three runs at HS S Viksmon
4500
5000
5500
6000
6500
7000
7500
Distance [m]
S
eat
vi
b
ra
ti
o
n
[
m
/s
2
]
Precision of 3 runs
Page 78
ROADEX III The Northern Periphery Research
5.3.2 Profilograph data is a proven reference to vertical ride vibration
Previous research has shown a very good fit between pavement roughness, as measured by a
Profilograph, and vertical ride vibration in heavy trucks. However, in road sections with very
poor bearing capacity this fit may drop significantly, as shown by Ahlin et al (2004) [19], due to
soft spots in the pavement affecting the roughness experienced through the truck tyres.
Profilograph data is frequently used in advanced studies on vehicle ride vibration. One exa mple
is seen in the Ph D thesis on heavy truck fatigue damage by Bogsjö (2007) [22], based on data
from Rd 331 and other Northern Periphery roads. Another study of Profilograph data from Rd
331 is reported in the Masters thesis on ambulance car ride quality by Nilsson (2004) [69].
5.3.3 Profilograph data emerge a good reference to truck roll angle and rate
5.3.2.1 RBCS show good fit to truck roll angle
In this project, an important issue is truck roll vibration and its relation to undesired variance of
the Rut Bottom Cross Slope (RBCS) of the pavement. The case study on Rd 331 included Pro-
filograph measurement of pave ment RBCS, as well as measurement of the dynamic roll angle
of the Scania R480 truck cab. Comparisons of the two types of data show a good fit, as seen at
Hazardous Site Backe (Edsele) in Figure 52. This confirms the value of Profilograph data as a
reference input for calculations into the dynamic roll motion of trucks.
Straigth road section in Backe (Edsele)
Pavement Rut Bottom Cross Slope vs estimated from Truck Cab Roll Angle
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
127000
127100
127200
127300
127400
127500
127600
RDB distance [m]
C
ro
s
s
S
lope
[
%
]
Rut Bottom Cross Slope from the Profilograph
Cross Slope estimated from Truck Cab Roll Angle
Figure 52
Good fit between pavement Cross Slope and Truck Cab Roll Angle
Page 79
ROADEX III The Northern Periphery Research
5.3.2.2 Rock ÂŽn Roll at HS Backe (Edsele)
Figure 53 shows a photograph of the truck, warping as it enters a section of severe long-wave
edge deformation at RDB distance 127/325 km. The truckâs large roll angle should be particu-
larly noted. Also noteworthy are the different roll angles of the truck and the trailer, as well as
the different yaw angles.
Figure 53
HS Backe (Edsele): Warping truck and trailer at RDB distance 127/325 km
A cut screen-print fro m the onboard Dewetron system is given in Figure 54. The recordings
were captured synchronized with photographs taken from the roadside as in Figure 53 above.
The forward facing video in Figure 54 actually shows the exterior photographer standing beside
the road. The white undulating top trace shows the variance of the truck cab roll angle, recorded
by the OxTS RT 3050 unit. The peak rate of the change in roll angle at the Hazardous Site at
Backe (Edsele) was over 3
°
/s. In the case of the test vehicle, a roll angle of 1
°
corresponds to
a 35 mm vertical displace ment between the left and right tyres.
Figure 54
HS Backe (Edsele): ROW video and some of the recorded truck ride data
Page 80
ROADEX III The Northern Periphery Research
Section 127/325 km was rated as â
very uncomfortable
â and â
hazardous
â by the truck drivers.
The vector of roll and pitch rates peaked at 4.48
°
/s. This touches the threshold used in New
Zealand, as presented in section â
3.3.2.2 Roll vibration is excited by undesired variance of Rut
Bottom Cross Slope
â. The sectionâs RBCSV peaked at 0.47 %.
Excessive Cross Slope in straight road sections is an ergonomic problem
The cross slope (CS) magnitude is remarkable at the Hazardous Site at Backe (Edsele). The
section has values ranging between -4 % and -7 %, see Figure 52, despite the design value for
straight roads being -2.5 %. The maximum value allowed for CS when designing extre mely
sharp curves in Sweden is +/-5.5 %. Obviously this straight section has too much CS. Excessive
CS contributes to a poor work environment for professional drivers as they have to sit in an
awkward side sloping position to counteract the adverse CS, causing the spine to be twisted,
while also being exposed to high ride vibration.
Response data are not good reference to road management
It may seem that truck response data could be useful in pave ment manage ment systems. How-
ever, this type of road condition data depends on many dynamic parameters, such as speed
and lateral position. Therefore response data from commercial trucks or other vehicles could
give poor esti mates of pavement parameters such as cross slope. This is obvious at RDB sec-
tion 127/300 km in Backe (Edsele), where the truck roll response differs by almost 2 percent,
compared to the pavement cross slope, as seen in Figure 52. This difference is many times lar-
ger than the tolerance limit applied on quality certified road condition data used in normal pave-
ment management systems.
Page 81
ROADEX III The Northern Periphery Research
5.3.2.3 Rock Žn Roll at HS à kerö
The HS Ă
kerö pavement edge da mage is seen on photo in Figure 55. Take note of the ex-
ploded truck tyre to the right! Traditional optical photographs do not reflect unevenness very
well. A better way of visualizing unevenness is to use a 3D laser scan. This highlights all of the
unevenness features. A Profilograph scan of the HS Ă
kerö damage can be seen in Figure 56.
The deformation at this site was found to be 69 mm deep.
A cut screen-print fro m the onboard Dewetron Strea m Machine system is reproduced for HS
Ă
kerö in Figure 57. The white undulating top trace shows the variance of the truck cab roll an-
gle. The RT 3050 had registered a peak rate for the roll angle of 6.8
°
/s. The lateral acceleration
measured in the truck cab was 2.0 m/s
2
, and it was 75 % higher (3.5 m/s
2
) at the driverâs seat
pan. The latter confirms that current seat suspension syste ms are unable to isolate lateral vibra-
tion as they have been designed primarily to isolate vertical vibration. These seats appear to
actually a mplify the problem, in fact, with lateral jolts at low frequencies. The HS Ă
kerö section
was rated as âvery uncomfortableâ and âhazardousâ by the truck drivers.
The cross slope magnitude is also remarkable at the straight road section at Hazardous Site
Ă
kerö. Recalling that the design recommendation for cross slope was -2.5 %, this section has
values ranging between -3 % and -6 %. The peak CS exceeds the maximum banking of +/-5.5
% allowed when designing the sharpest curves in Sweden.
Figure 55
Pavement edge deformation at HS à kerö. Take note of the exploded truck tyre!
Page 82
ROADEX III The Northern Periphery Research
Figure 56
Profilograph 3D plot of the HS à kerö 69 mm deep edge deformation
Figure 57
HS à kerö: ROW video and some of the recorded truck ride data
Page 83
ROADEX III The Northern Periphery Research
As stated above, an important issue for this research was the relationship between undesired
variance of the Rut Bottom Cross Slope (RBCS) of the pavement and the truck roll vibration. In
the previous section, a good fit was found between RBCS and truck cab roll angle. The variance
of cab roll angle is a measure of the cabâs roll vibration. Further analysis confirms a good fit be-
tween variance of the roll angle and variance of the RBCS (RBCSV), as can be seen in data
from HS Ă
kerö in Figure 58.
Variance of Truck Cab Roll Angle vs Variance of Cross Slope
HS à kerö edge damage at 125 275 m
0,0
0,2
0,4
0,6
0,8
1,0
1,2
125000
125200
125400
125600
125800
126000
126200
126400
RDB distance [m]
Va
ri
a
n
c
e
Variance of truck cab roll angle
Pavement RBCSV
Figure 58
Good fit between Variance of RBCS and Variance of Truck Cab Roll Angle
The RBCSV parameter has been designed to identify those sections with cross slope variance
that cause roll vibration in the suspended masses (body, cab and payload) of heavy trucks, as
well as in the wheel axle [21]. As a result of this multi-purpose requirement, one should not look
for a perfect match between RBCSV and the roll vibration of the cab. There can be significant
variances between reproduced truck rides, as seen in Figure 51. With this in mind, the match
seen in Figure 58 seems good for the intended purpose of the RBCSV parameter.
Page 84
ROADEX III The Northern Periphery Research
5.3.2.4 Warping RBCS at Hazardous Site N Ă kroken
The Hazardous Site north of Ă
kroken shows an unusually high accident number, as seen in
Figure 28. This site also show extre mely high RBCSV; 1.04 %. This is ten times the ânoise levelâ
of 0.1 %, as seen in Figure 59. This clear alarm is caused by a warping change in cross slope
from -4 % to -7.5 % and then back to -4 %. The net change of 2.5 % cross slope corresponds to
a change of 5 cm in elevation between left and right wheel track, as they are spaced 200 cm;
0.025 * 200 = 5 cm.
HS N Ă kroken
-8
-6
-4
-2
0
2
4
53500
53600
53700
53800
53900
54000
54100
54200
54300
54400
54500
RDB Distance [m], lane in reverse direction
Cr
o
ss S
lo
p
e [
%
]
0,0
0,2
0,4
0,6
0,8
1,0
1,2
N
e
t va
ri
a
n
ce
[
%
]
Rut Bottom Cross Slope
Effective Unde sired RBCS Variance
Figure 59
High RBCSV indicate severe pavement edge deformation at HS N Ă kroken
Page 85
ROADEX III The Northern Periphery Research
5.3.2.5 Warping RBCS at Hazardous Site MeÄstrand
In contrast the Hazardous Site at MeĂ„strand âonlyâ shows a slightly increased accident number.
However, the truck drivers report this to be one of the most dangerous sites. The modest con-
sistency between driver opinion and accident black spot map, may be explained by the fact that
this site has low traffic intensity. This calls for an âIndividual Riskâ mapping, where accident
number is divided by traffic intensity AADT as described by Ogden & Daly [64]. However, such
an analysis requires further resources not available within this project which has its focus on
health issues rather than traffic safety.
HS MeÄstrand shows a high degree of warping RBCSV; 0.95 %. This is almost ten ti mes the
ânoise levelâ of 0.1 %, as seen in Figure 60. This clear âalarmâ is caused by a warping change in
cross slope fro m -2.9 % to -5.2 % and then back to -3.0 %. The net change of 2.3 % cross slope
corresponds to a change of about 4.6 c m in elevation between left and right wheel track, as they
are spaced 2 m. A photo of a truck yawing to avoid this pavement edge is showed in Figure 61.
Take note of how the painted road marking line reflects the lateral component of the pavementâs
deformation. Also take note of the glare of the asphalt repair in the outer wheel track. Friction
aspects on this kind of single track patch repair will be further discussed later in the report.
HS MeÄstra nd
-6
-4
-2
0
2
4
6
112000
112100
112200
112300
112400
112500
112600
112700
112800
112900
113000
RDB Distance [m], lane in reverse direction
C
ro
ss S
lo
p
e
[%
]
0,0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
1,0
N
e
t var
ia
n
ce
[%
]
Rut Bottom Cross Slope
Pavement RBCSV
Figure 60
High RBCSV indicate severe pavement edge deformation at HS MeÄstrand
Page 86
ROADEX III The Northern Periphery Research
Figure 61
Severe pavement edge deformation at HS MeÄstrand
Page 87
ROADEX III The Northern Periphery Research
5.3.2.6 Warping RBCS at Hazardous Site AlderÄn
HS AlderÄn shows an unusually high accident number, as seen in Figure 28. This site also
shows several peaks with RBCSV up to 0.55 %. This is over five times the ânoise levelâ of 0.1
%, as seen in Figure 62.
-10
-8
-6
-4
-2
0
2
4
6
127500
128000
128500
129000
129500
130000
RDB Distance [m], lane in reverse direction
C
ros
s
S
lope
[
%
]
0,0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
1,0
N
e
t va
ri
an
c
e
[
%
]
Rut Bottom Cross Slope
Effective Undesired RBCS Variance
Figure 62
High RBCSV indicate warping pavement edge deformation at HS AlderÄn
HS AlderÄn also has other problematic features. It is a sharp right hand curve at the foot of a
long and steep hill. In the curve, the cross slope is worse than -8 %. This is a very large slope,
especially when appearing just after a long grade. Furthermore, the cross slope transition
lengths are too short. On top of all of this, the pavement Mega Texture (MeTx) is unacceptably
high in the curvesÂŽ outer wheel path.
MeTx is longer than Macro Texture, but shorter than roughness. These short waves range from
5 up to 50 cm. High MeTx causes distortion in the tyre/road contact patch, thus being a source
of friction problems. MeTx is also a significant source of annoying interior and exterior noise. At
HS AlderĂ„n, the MeTx peaked at 0.9 mm, being 4.5 times the normal ânoise levelâ of 0.2 mm.
Page 88
ROADEX III The Northern Periphery Research
5.3.2.7 Warping RBCS at Hazardous Site à sÀng
HS Ă
sÀng shows an unusually high accident number, as seen in Figure 28. A photo of a recent
truck skid accident can be seen in Figure 63. The Police reported the friction to be zero, due to
polishing. The truck driver said that the truck responded neither to steering, nor to braking, at
the exit of the left hand curve.
Figure 63
Truck skid accident in à sÀng, 2007-02-22 [Photo: Torbjörn Elverheim, ST]
This site also shows several peaks with RBCSV up to 0.49 %. This is over five times the ânoise
levelâ of 0.1 %, as seen in Figure 64.
There are three transient RBCSV peaks in Figure 64, at distances of 11/450, 11/435 and 11/403
km. Such peaks relate to short wave CS variance, which excite wheel axle tra mp. This rolling
motion of the axle results in lateral tyre displacements polishing the road, as discussed in sec-
tion â
3.2.6 Wheel axle vibration impacts on traffic safety
â. The Police observation on the road
being polished into low friction give support to the theory of RBCSV causing low friction due to
polishing.
As mentioned, Figure 64 shows three transient RBCSV peaks between 11/450 and 11/403 km.
These transients increase the RBCSV by some 0.1 %, and correspond to about 1 % change in
RBCS as seen in the graphs. In other terms, the transients correspond to 0.01 x 200 = 2 cm
Page 89
ROADEX III The Northern Periphery Research
warping in profile elevation in one of the rut bottoms. Such warps are characteristic for steps at
the start and end of poorly made 2 cm thick pavement repair patches. Macro Texture results
(not shown here) also indicate likely starts and stops of patch work at the sections distance
11/450, 11/435 and 11/403 km. Furthermore, the MaTx was re markably low, down to 0.2 mm at
the patch indications. A benchmark mini mum value is 0.6 mm MaTx for acceptable wet friction
when braking at highway speeds. As seen on the photograph in Figure 63, the ice rut bottoms
showed some asphalt, so the low MaTx may possibly have contributed to the observed low fric-
tion.
Junction w ith Rd 703 bound for Ljustorp at section 11 497 m
-8
-6
-4
-2
0
2
4
6
8
11375
11425
11475
11525
11575
11625
11675
11725
RDB Distance [m], lane in reverse direction
C
ro
s
s
S
lope
[
%
]
0,0
0,2
0,4
0,6
0,8
1,0
1,2
1,4
E
ff
e
ct
iv
e
RB
CS
va
ri
an
c
e
[
%
]
R ut B ottom Cross Slope
Undesire d Rut Bottom Cross Slope Va riance
Figure 64
High RBCSV indicate warping pavement edge deformation at HS à sÀng
The photograph in Figure 63 was taken at about section 11/400 km, and the skidding may have
occurred just at the peak RBCSV in section 11/435 km. There is also another indication of the
accident being related to the pavement condition. This is discussed in the section on insufficient
Drainage Gradient.
Page 90
ROADEX III The Northern Periphery Research
5.3.2.8 Warping RBCS at Hazardous Site MjÀllÄsen
HS MjÀllÄsen shows an unusually high accident number, as seen in Figure 28. It has a long
curve with varying curvature. In this curve, there are sections with excessive CS; up to 8.23 %
negative CS. An ideal CS for dyna mic cornering balance with respect to current curvature is -4
% in this section.
This site also shows several RBCSV peaks up to 0.85 %. This is more than eight ti mes the
ânoise levelâ of 0.1 %. This clear âalarmâ is caused by a warping change in cross slope from -3.3
% to -0.8 % and then back to -3.9 %. The net change of up to 3 % cross slope corresponds to a
change of 6 cm in elevation between left and right wheel track, as they are spaced 2 m. The
RBCS trace down to -0.8 % indicates that the pavement centre, rather than the edge, has
collapsed. If this is the case, this road section could have serious bearing capacity proble ms.
Such problems should be considered when planning repair of the road section. If only a si mple
surface repair is done, the road will most likely deteriorate in very short ti me.
-10
-8
-6
-4
-2
0
2
4
6
196 00
19 700
19800
1990 0
20 000
20100
20200
RD B D istance [m], lane in reverse direction
C
ros
s
S
lop
e
[
%
]
0,0
0,2
0,4
0,6
0,8
1,0
1,2
E
ff
e
c
ti
v
e R
B
C
S
var
ia
n
c
e
[
%
]
Rut Bottom Cross Slope
Undesired R ut B ottom Cross Slope Varia nc e
Figure 65
High RBCSV indicate warping pavement deformation at HS MjÀllÄsen
Page 91
ROADEX III The Northern Periphery Research
5.3.2.9 RBCS is a more accurate pavement parameter than Lane Regression CS
There are many ways to define/ measure the Cross Slope (CS) of a pavement lane. Since 1997,
the Swedish Road Administrationâs Pavement Management System (PMS) has measured CS
by the âregression methodâ. This uses the 17 laser measurement spots on the Profilograph dis-
tributed over a width of 3.2 m. In the regression method, data fro m the whole 3.2 m wide cross
section are used to calculate the CS. In sections with severe edge deformation, the regression
method may report significant smaller slopes than perceived between the left and right wheels
of a truck. One example is found at HS Ă
kerö, in the section showing severe edge damage at
RDB distance 125/275 km. As can be seen in Figure 66, the lane regression CS differs one unit
from the 6 % truck cab roll angle; a relative difference of 1/6 = 17 %. In 2006, the SRA defined
Rut Bottom Cross Slope (RBCS) as a parameter focusing on pavement slopes as perceived as
a priority by drivers of heavy trucks [21]. Figure 66 shows a good fit between lane regression CS
and the new RBCS parameter, except at the pavement damage in section 125/275 km, where
the RBCS matches the truck roll angle much better. The difference between these two meas-
ures of cross slope can be large at sections with severe edge deformations in the shape of a
basin, where a âwallâ of displaced material is raised outside the outermost wheel as seen in
Figure 55 and Figure 56. This kind of damage has a great effect on truck ride and RBCS must
therefore be considered to be the most accurate road-user oriented para meter of the two.
Figure 66
RBCS is a better estimate of truck roll angle, than Lane Regression CS is
Page 92
ROADEX III The Northern Periphery Research
5.3.3 Road edge deformation may excite as much lateral vibration as a curve
Lateral acceleration is commonly recognized as a key para meter for vehicle driving stability, and
thus for traffic safety. This is especially relevant on slippery surfaces, where the lateral friction
forces are small. When a vehicle changes its roll angle quickly, the roll motion is accompanied
by a lateral acceleration. An example from HS Ă
kerö is given in Figure 67. This shows a left
hand curve (curvature -1.6) at distance 126/200 km, reflected by the change of sign in cross
slope as it becomes superelevation through the curve. The graph for âRunning Root-Mean-
Square of Truck Cab Lateral Accelerationâ shows a semi-static level of 0.78 m/s
2
through the
curve. This can be co mpared to the value of 0.66 m/s
2
for lateral RMS acceleration recorded on
the section of straight road with severe edge damage at HS Ă
kerö, section 125/275 km. In this
latter section, the peak lateral acceleration was -1.37 m/s
2
, whilst the peak lateral acceleration in
the curve at 126/200 km was only -0.94 m/s
2
.
The HS Ă
kerö example clearly shows that severely deformed pavement edges are a serious
safety hazard, as they may result in lateral acceleration forces comparable to the lateral forces
experienced when travelling a horizontal curve.
The grey trace in Figure 67 shows that the pavement RBCSV parameter was registering ap-
proximately 0.1 % through the curve, where the cab lateral acceleration was fairly constant with
low vibration. However as intended, the parameter quickly gives a clear alarm of 1.18 % (being
over 6 times larger than the 0.1 to 0.2 % noise level) when it enters the HS Ă
kerö section of
pavement edge damage. This example also shows that the RBCSV parameter does not give
âfalse alarmâ due to normal superelevation transitions at left hand curves, where the truck cab
roll angle s moothly tilts from side to side.
Truck Cab Lateral Accelaration vs Pavement Cross Slope
HS à kerö edge damage at 125 275 m
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
125000
125200
125400
125600
125800
126000
126200
126400
RDB distance [m]
C
ro
ss S
lo
p
e [
%
]
-1,5
-1,0
-0,5
0,0
0,5
1,0
1,5
L
a
te
ra
l accel
er
at
io
n
[
m
/s
2
]
Rut Bottom Cross Slope
Cab lateral acc
Running RMS of Cab lat acc
RBCSV
Figure 67
Edge damages may excite as much lateral acceleration as a horizontal curve do
Page 93
ROADEX III The Northern Periphery Research
5.4 DRAFTING A âWARPING LIMITâ FOR RBCSV
One goal in this project was to draft limit values for maxi mum warping between the road profile
in left and right wheel track; the âundesired varianceâ of the pavementâs Rut Bottom Cross Slope
(RBCS). For this task, it is important to understand typical distributions of warping on road
sections in ânormalâ and in bad condition.
The 26.5 km section from Ăstergraninge down to Viksjö is a âquite normal old roadâ. Co mpared
to the Hazardous Sites
N Ă kroken
and
N Viksjö
(see the black spot map in Figure 28), it shows
a modest accident record. The distribution of Rut Botto m Cross Slope Variance (RBCSV) values
on this section is shown in Figure 68.
"Quite normal old road", RDB 58 300 - 31 760
51%
74%
87%
93%
97%
99% 99% 100%100%100%100%100%100%100%100%100%100%100%100%100%100%
0
2000
4000
6000
8000
10000
12000
14000
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
1,10
Undesired RBCS variance [%]
Fr
e
que
nc
y
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Sh
a
re
Frequency
Cumulated share
Figure 68
RBCSV distribution at a ânormalâ old road, Ăstergraninge - Viksjö
The 17 km from the junction with Hw 87 in Viks mon, and down southeast to Ăstergraninge, is a
rough section with severe pave ment edge deformations, resulting in intense lateral vibration in
the truck cab. An unusually large number of traffic accidents have taken place on this section,
including the HS BjörknÀset, as seen on the black spot map in Figure 28. The distribution of Rut
Botto m Cross Slope Variance (RBCSV) values on this section is shown in Figure 69.
Page 94
ROADEX III The Northern Periphery Research
j
"Rough road", RDB 74 989 - 58 300
2%
12%
29%
47%
61%
73%
81%
86%
90%
94%
95% 97%
97% 98% 98% 99%
99% 99% 100%100%100%
0
500
1000
1500
2000
2500
3000
0,10
0,20
0,30
0,40
0,50
0,60
0,70
0,80
0,90
1,00
1,10
Undesired RBCS variance [%]
Fr
e
que
nc
y
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Sh
a
re
Frequency
Cumulated share
Figure 69
RBCSV distribution at the rough road section S Viksmon â Ăstergraninge
Sample results of collation of Profilograph data, truck ride data and truck driver perception of
comfort and safety are given in earlier report sections. These results show that a RBCSV of 0.4
% is too high as a limit value. On the other hand, 0.1 % RBCSV corresponds to the âbackground
noiseâ on old roads, and is obviously too low to be a limit value. A reasonable limit could be
somewhere between 0.2 % and 0.3 % RBCSV. The graph in Figure 68 shows that 3/100 of the
old road length exceeds 0.30 % RBCSV, while 13/100 of the road length exceeds 0.20 %
RBCSV. Since it is i mportant to focus road repair to a limited fraction of the road network, a rea-
sonable draft limit value could therefore be 0.30 % RBCSV.
On the rough road, 0.30 % RBCSV is exceeded on 39/100 of the length, as seen in Figure 69.
Again, 0.30 % RBCSV is exceeded on 3/100 of the length of the old road Ăstergraninge - Vik-
sjö, which include so me Hazardous Sites. This shows that 0.30 % RBCSV can be a good draft
limit value.
A statistical analysis of the data from the section fro m Ra msele to Ărtrik shows that 0.3 %
RBCSV gave approxi mately 2.0
°
/s roll rate in the test truck at the normal operation speeds on
this road section.
Page 95
ROADEX III The Northern Periphery Research
Chapter 6. Spin-off results on traffic safety
A map of accident black spots on Rd 331 is shown in Figure 28 and data recording intense truck
vibration and high road roughness values have been presented above for a number of the Haz-
ardous Sites involved. These findings support the theory that a causal relationship between
pavement da mages and traffic accident risk.
When taking ride measurements in the high truck, the SRA CS team clearly perceived
unexpected high lateral forces in many curves. This indicated that the curves may be incorrectly
banked. This suspicion was further enhanced by complaints from truck drivers. As a
consequence of this it was decided to make some analysis of the dyna mic equilibrium of
cornering forces due to road alignment in the curves. For many curves (and straight sections as
well), the analysis resulted in some alarming results as seen below.
A refined analysis method was de monstrated to quickly show if a curve is correctly banked or
not. The results confirm that many curves on the Beaver Road 331 are incorrectly banked and
thereby hazardous.
Data clearly show an overrepresentation of incorrect banking in left
16
hand curves. The causal
reason has been analyzed and explained.
Several left hand curves include sections where cross slope is 0 (zero) %. These have also
been investigated with respect to Drainage Gradient (DG), the resultant vector to cross slope
and longitudinal grade. The results show that, on roads with modest grades, the vast majority of
left hand curves have spots with unacceptably low DG at their entrance and/or exit, resulting in
a high skid risk due to water ponding. Analysis on new sections on other roads confirms that this
is a ubiquitous problem in road design.
Finally, analysis shows that the spot repair of pavements, in one wheel track only, may cause
hazardous split friction when braking hard at high speed (emergency baking) in wet weather
conditions.
16
Sweden has right hand traffic; in the UK the problems are focused into right hand curves instead.
Page 96
ROADEX III The Northern Periphery Research
6.1 EFFICIENT ANALYSIS OF INCORRECTLY BANKED CURVES
6.1.1 Ideal ratios of cross slope and horizontal curve radius
As seen in section â
3.3.1 Tight curves are hazardous
â, road agencies worldwide have defined
ideal ratios for cross slope versus horizontal curve radius, in order to create a dynamic balance
to the lateral cornering force. An example of ideal ratios at the reference speed 90 km/h used in
Sweden is given in Figure 21.
It is not practical to relate cross slope (CS) to radius (
R
) when analyzing data from real roads,
including straight sections of roads where
R
approaches infinity and beco mes difficult to plot.
Curvature, defined as 1000/
R
, is a more practical parameter for real roads since it approaches 0
(easy to plot) on straight sections. Furthermore, curvature is directly proportional to the lateral
cornering force. Therefore, the ideal CS to
R
ratios in Figure 21 has been plotted as CS to Cur-
vature in Figure 70. The green boxes in the figure correspond to a high standard of road align-
ment, whereas the orange boxes correspond to a moderate to low standard. These boxes in-
clude +/- 0.5 % tolerance limits, as implied by (complex) tolerance demands stated in the Swed-
ish road construction code [40]. The sign convention in Sweden is illustrated by a two lane road
cross section to the left in the figure. Sweden has right hand traffic, so the focus is on the right
hand lane. In straight road sections, the correct cross slope is -2.5 % (-3 % for roads with cold
non-mixed pave ments) and curvature is 0. In right hand curves, the absolute value of cross
slope (banking/superelevation) should be increased where the curvature is high (radius low); the
most extreme design value is -5.5 %.
With +/- 0.5 % tolerance, the most extreme box goes from -6 % to -5 % CS. Significant left hand
curves (> 0.3 % negative curvature) call for the cross slope to be tilted to the other side, thereby
changing the CS sign. Corresponding sharp right hand curves, the most extre me box for left
hand curves goes fro m + 5 to + 6 %. The CS transition between -2.5 % and +2.5 % should at
roads with 90 km/h reference speed be carried out at negative curvature smaller than 0.3,
corresponding to a radius wider than -3200 m.
Page 97
ROADEX III The Northern Periphery Research
(
-
)
(
-
)
(
+
)
(
+
)
(
-
)
(
-
)
Figure 70
Ideal Ratios between Cross Slope and Curvature at 90 km/h reference speed. After [15]
Page 98
ROADEX III The Northern Periphery Research
6.1.2 Reference patterns of fair alignment on real roads
In Figure 71, 12300 values from a reconstructed section of Hw 90 are plotted. Each point repre-
sents the average value for 1 m road section, in total 12.3 km. This plot gives a reference to pat-
terns created by fair ratios between cross slope and curvature on real roads.
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
7
8
-4
-3
-2
-1
0
1
2
3
4
Curvature = 1000 / Radius [m]
Cr
o
s
s sl
o
p
e
[%
]
Figure 71
New Hw 90: Reference ratios between CS and Curvature, 90 km/h reference speed
It is easy to identify a handful of reference road alignment âfa miliesâ in the plot, as seen in
Figure 72. Straight sections are marked â1â, low cross slope (CS) in wide right curves â2â, high
CS in sharp right curves â3â, low superelevation in wide left curves â4â, high superelevation in
sharp right curves â5â and cross slope transitions to/from left curves â6â.
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
7
8
-4
-3
-2
-1
0
1
2
3
4
Curvature = 1000 / Radius [m]
Cr
o
s
s sl
o
p
e
[%
]
Figure 72
Identifying reference road alignment families
1
2
3
4
5
6
Page 99
ROADEX III The Northern Periphery Research
6.1.3 Incorrectly banked curves on Rd 331
If the road alignment data does coincide with the fair road alignment reference fa milies, the road
section is likely uncomfortable and possibly hazardous at the reference speed.
Data from 12.3 kilometres of the old Beaver Road 331 (Ramsele - Edsele) are compared with
data from new Hw 90 in Figure 73. In this plot, several uncomfortable and hazardous families of
road alignment data can be identified. Note that the Curvature axis has been widened, to make
it possible to plot data from the sharp curves on Rd 331.
Straight sections with excessive Cross Slope (CS) are marked â7â in Figure 73. These sections
include CS greater than the permissible banked sharp curve allowed in Sweden. These cause
an uncomfortable ride. They are also a health risk, since they force the driverâs spine into an
awkward twisted position making it much more susceptible to Whole-Body Vibration. These sec-
tions are also hazardous when overtaking another vehicle, as the large difference in CS be-
tween the two lanes causes large lateral vibrations if the overtaking is done with a quick lateral
manoeuvre. On the section NordankÀl - Backe, a straight section had CS down to - 8.5 % (not
shown here).
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
7
8
-5
-4
-3
-2
-1
0
1
2
3
4
5
Curvature = 1000 / Radius [m]
Cro
ss sl
o
p
e
[%
]
Figure 73
Comparison of data from
old Rd 331
and new
Hw 90
Wide right hand curves with too little negative CS are marked â8â in Figure 73. These curves
contribute to skid accidents as they do not generate sufficient lateral support to reach a dynamic
balance when cornering in slippery conditions.
Sharp right hand curves with high negative CS are marked â9â. These curves contribute to slip
accidents in vehicles driving at lower speeds than the reference speed.
7
8
9
10
11
12
13
Page 100
ROADEX III The Northern Periphery Research
Left hand curves where the banking is tilted to the wrong side are marked â10â. These are of
course extremely hazardous, as it is difficult to avoid skidding in slippery conditions. For heavy
vehicles, the risk for rollover accidents is obvious.
Sections with close to 0 (zero) % CS are marked â11â. With few exceptions, these are entrances
or exits of left hand curves. Unless these sections happen to be in a longitudinal grade, they will
also have a Drainage Gradient close to zero. This will cause water to pond, and the road will
often develop local (surprising) ice spots in cold weather. These bring unacceptably high skid
risk, so these sections should be checked for Drainage Gradient. This type of analysis is done in
the next section of this report.
Left hand curves with too little CS are marked â12â. These curves contribute to skid accidents,
as they do not generate enough lateral support to reach a dynamic balance when cornering in
slippery conditions. It is noticeable that almost none of the left hand curves have excessive
positive CS.
The family marked â13â can be described as âpoorly synchronized CS transitionsâ. This family
includes sections where CS transitions take place at a curve radius sharper than - 3200 m
(Curvature -0.3). In practice, these are sections where the curve has started, but the
superelevation is applied later in the curve. Or even worse; in curves where positive
superelevation suddenly becomes negative CS before the curve is finished. This kind of road
feature can come as a dangerous surprise to road users, unfamiliar with local hazards. The data
families â10â to â12â may also include poorly synchronized transitions.
Figure 73 also includes families of unacceptably sharp left and right curves on Rd 331. When
these curves are evaluated against the acceptable risk level stated by the SRA Road Design
Manual [15], many are even too sharp at the lower reference speeds of 70 and 50 km/h. As per
the Tylösand declaration (see section â
7.2 The Tylösand Declaration
â), SRA must as soon as
possible make sure that all of these curves gets warning signs, and should start planning for
straighten the m out.
It is interesting to see that Figure 73 is non-symmetric. It shows that hazards are more common
in left hand curves, than in right hand curves.
Page 101
ROADEX III The Northern Periphery Research
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
7
8
-10
-9
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
Curvature = 1000 / Radius [m]
Cr
o
s
s sl
o
p
e
[
%
]
Figure 74
Poor road alignment cause dynamic imbalance at the Hazardous Site Roos Curve
The Roos Curve in Ăsterforsse is extremely hazardous, as can be seen in the hospital record
plot in Figure 16. Today, the speed limit is 70 km/h through the curve. When analyzing ratios
between CS and Curvature for Roos Curve in Figure 74, the data can be classified into several
of the above defined hazardous families of road alignment. This non-uniform horizontal curve
includes a minimum âradiusâ of about - 120 m, as Curvatures reach - 8.32. These extreme Cur-
vatures cause very high lateral forces. Considering the curveâs CS, and a lateral friction factor of
less than 0.1 for slippery conditions, the maximum safe speed is definitely lower than 50 km/h
as per the graphs in Figure 18. The posted speed limit is therefore more than 40 % higher than
the safe maxi mum speed. Obviously, it is extremely important to maintain high road surface fric-
tion in this curve. It is reco mmended that this curve should be straightened out, or at least
should have the banking very carefully redesigned, as soon as possible. (Each point in Figure
74 corresponds to an average value over 1 m; the plot includes 500 m).
There are two Hazardous Sites (HS) at Viksjö; one just north of the village and the other just
south of Viksjö, as seen in the hospital record plot in Figure 75.
The HS south of Viksjö shows three fatal heavy truck accidents at exactly the same location. In
the flat village of Viksjö the speed limit is 50 km/h. At the south exit, the limit is raised to 70
km/h. The road makes a short and sharp left hand curve, as it begins a long and steep downhill
grade. Then it makes a wide right hand curve, followed by a short but very sharp left hand
curve. This third curve also ends the grade, and the road goes over a bridge at the botto m of the
valley. In the grade, truck drivers have lost control of their vehicles. At the exit of the third curve,
each of the trucks have missed the bridge and made a large hole in the - obviously undersized -
crash barrier. All of these lethal rides ended with a 20 m long and 12 m deep jump into the rift.
Page 102
ROADEX III The Northern Periphery Research
Figure 75
Hospital record plot from HS N Viksjö and HS S Viksjö [Hans Johansson, SRA]
Figure 76
HS S Viksjö: Truck crash in Oct 2005 [Photo: High Coast Rescue Dept]
When analyzing ratios between CS and Curvature for HS S Viksjö in Figure 77, the data can be
classified into several of the above defined hazardous families of road alignment. The non-
uniform horizontal curves include a minimum âradiusâ of about -150 m, as Curvatures reach -
6.77. These extre me Curvatures cause very high lateral forces. Considering the curveâs CS, and
Page 103
ROADEX III The Northern Periphery Research
a lateral friction factor below 0.1 for slippery conditions, the maximum safe speed is about 50
km/h as per the graphs in Figure 18. The posted speed limit is 40 % higher.
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
7
8
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
Curvature = 1000 / Radius [m]
C
ros
s
s
lope
[
%
]
Figure 77
Poor road alignment cause dynamic imbalance at the Hazardous Site S Viksjö
Page 104
ROADEX III The Northern Periphery Research
6.2 IDENTIFYING HIGH SKID RISK DUE TO WATER PONDING
Where the âDrainage Gradientâ of a non-permeable road surface is lower than 0.5 %, water will
not run off and water pools can be formed in wet weather. Water ponding, such as seen in
Figure 78, increase the skid accident risk.
Figure 78
Water ponding at a CS transition section [Photo from the UK road network survey]
Road design manuals worldwide recognize the risk for water ponding and demand a minimum
Drainage Gradient of 0.5 %.
From a mathe matical point of view, Drainage Gradient (DG) is the resultant of the Cross Slope
(CS) and longitudinal Gradient (G) of road surface, as illustrated in Figure 79 and defined by
Formula 8.
Page 105
ROADEX III The Northern Periphery Research
Figure 79
Drainage Gradient is resultant of Cross Slope and longitudinal Gradient
2
2
G
CS
DG
+
=
Formula 8, Calculation of Drainage Gradient
Straight roads are generally designed with 1 to 3 % negative CS. Curves are designed with a
superelevation up to +/-5.5 % (Sweden), +/-7 % (UK), +/-9.5 % (Norway, maintenance of
existing roads).
With more than +/-0.5 % CS, the DG should never drop below the minimum li mit of 0.5 %. Nei-
ther straight sections, nor curves, have less than +/-0.5 % CS. So in what type of road sections
could DG become insufficient? The Swedish road design manual [15] does not include guidance
on this important question. The UK road design manual [54] gives a clue on the topic:
â-
Care must be taken to ensure that a minimum longitudinal gradient of at least 0.5 % is
maintained wherever superelevation is to be applied or reversed
â.
So, critical sections are the transitions where superelevation starts or stops between straight
sections and curves. As shown later in this section, the critical sections are further limited to
left
17
hand curves, where CS change direction and sign as they pass through 0 (zero) %.
An important question is: â-
How can unacceptably low DG be avoided at entrances and exits of
left hand curves in flat terrain
?â Again, the Swedish road design manual does not give guidance,
while the UK manual does:
â-
In flatter areas, the vertical alignment should be manipulated by the introduction of vertical
curvature simply to achieve adequate surface water drainage
â.
The solution presented in the UK manual, is to construct local vertical curves so there are at
least 0.5 % longitudinal Gradient in the sections where CS is close to 0 % as it changes sign. To
create a 0.5 % slope over 50 m length, an elevation of 0.25 m is required. This is a reasonable
17
In the UK the opposite is true;
to right hand curves.
This is due to the left hand driving in the UK.
Page 106
ROADEX III The Northern Periphery Research
design option when building a new road, where profiling is made by local material such as
gravel and rock fill or by cut section. It is important to use smooth gradient changes, so the ride
doesnât get bumpy at highway speeds.
There is also another solution. Where there is a grade before the left hand curve starts, the CS
can be moved into this grade. Thereby the CS can make a transition from -2 % to +2 % long
before the left hand curve starts. One drawback with this method is that a longer road section
will have somewhat longer drainage path, thereby also a slightly larger water film depth.
A third option is to minimize the length of road where the CS is close to 0 %. This can be done
by varying the âtilt rateâ. The transition from i.e. -2 % to -0.5 % can have a low tilt rate, from -0.5
to +0.5 % a higher tilt rate, and from +0.5 % to +5.5 % the tilt rate is slow again. Road sections
designed with this method should be checked to ensure that it does not excite significant roll
vibration.
A fourth option to reduce water ponding and thereby the skid risk, is to construct a permeable
pavement in the sections where DG is low.
6.2.1 Low Drainage Gradient gave an unacceptable skid risk at HS N Viksjö
The northern HS at Viksjö shows an unusually high number of accidents leading to hospital
care, as seen in Figure 75. The curve is sharp and induces high lateral cornering forces; the
curvature reaches -5.12 (radius tighter than 200 m). Despite this fact, the maxi mum allowed su-
perelevation of 5.5 % has not been utilized, as the curve only has about 4 % in the northbound
direction. Furthermore, it has a warping Rut Bottom Cross Slope Variance of 0.54 %, which is
significantly above the above proposed âwarping limitâ of 0.30 % RBCSV. In addition to these
features, at the southern entrance of the curve, the Drainage Gradient is below the 0.5 % mini-
mum limit value on a long section as shown in Figure 80.
Page 107
ROADEX III The Northern Periphery Research
Rd 331, Viksjö - Ăstergraninge
Curve at the north of Vik sjö, RDB c a 3 2000
Skid risk: If the Drainage Gradient doesnÂŽt ex ceed 0.5 %, water pools will be forme d
0,0
0,5
1,0
1,5
2,0
2,5
3,0
3,5
4,0
4,5
5,0
5,5
6,0
31700
3180 0
3 1900
32000
3210 0
32 200
3 2300
32400
3250 0
32 600
32700
RDB Distance [m], lane in reve rse direction
D
ra
ina
ge
G
ra
d
ie
nt
[
%
].
Dra inage Gradient
Unac ceptably low Dra inage Gradient
Figure 80
Extreme skid risk due to low Drainage Gradient at entrance of the left curve at HS N Viksjö
Page 108
ROADEX III The Northern Periphery Research
6.2.2 Low Drainage Gradient gave unacceptable skid risk at HS S Viksjö
Figure 76 shows a photograph of a truck accident in front of the bridge at southern Hazardous
Site in Viksjö. As previously presented, this site has poorly banked curve and a low Drainage
Gradient at the exit of the sharp left hand curve (in front of the bridge), as seen in Figure 81.
Bridge in Viksjö at 30 050 and 30 021 m
0
2
4
6
8
10
12
29600
29800
30000
30200
30400
30600
30800
31000
RDB Distance [m], lane in reverse direction
G
ra
d
ie
n
t [
%
]
Drainage Gradient
Unacceptably low Drainage Gradient
Unacceptably high Gradient
Figure 81
High skid risk due to low Drainage Gradient before the bridge at HS S Viksjö
Page 109
ROADEX III The Northern Periphery Research
6.2.3 Extreme skid risk at HS BjörknÀset
Hazardous Site BjörknÀset shows an unusually high accident number, as seen in Figure 28. At
this site, the Drainage Gradient is low for hundreds of metres, see Figure 82. This causes water
ponding and the formation of ice during the winter, bringing extremely high skid risks. This haz-
ardous geometry may be explained by the road section being very weak so the pavement has
collapsed totally.
HS BjörknÀset
0
1
2
3
4
5
6
7
61000
61200
61400
61600
61800
62000
62200
62400
62600
62800
63000
RDB Distance [m], lane in reverse direction
D
ra
ina
g
e
G
ra
d
ie
nt
[
%
].
Drainage Gradient
Unacceptably low Drainage Gradient
Figure 82
Skid risk due to low Drainage Gradient over hundreds of metres at HS BjörknÀset
Page 110
ROADEX III The Northern Periphery Research
6.2.4 Extreme skid risk at HS Helgum
Hazardous Site Helgum shows an unusually high accident number, as seen in Figure 28. At this
site, the Drainage Gradient is unacceptably low for hundreds of metres, see Figure 83. This
causes water ponding and the formation of slippery ice during the winter, bringing extre mely
high skid risks. Another problem is excessive Cross Slope (CS). In the junction with Rd 950 at
RDB distance 86/422 km, Rd 331 makes a curve with CS up to + 6.3 %. In this 70 km/h section,
a CS of + 2.5 % is sufficient with respect to current curvature, as per the Swedish road design
code. As seen in accident records, many vehicles turning in the junction with Rd 950 are skid-
ding.
0
1
2
3
4
5
6
7
8
85500
85600
85700
85800
85900
86000
86100
86200
86300
86400
86500
RDB Distance [m], lane in reverse direction
D
ra
ina
ge
G
ra
d
ie
nt
[
%
].
Drainage Gradient
Unacceptably low Drainage Gradient
Figure 83
Extreme skid risk due to low Drainage Gradient on one hundred meter at HS Helgum
Page 111
ROADEX III The Northern Periphery Research
6.2.4 Skid risk at HS à sÀng
Hazardous Site Ă
sÀng shows an unusually high accident number, as seen in Figure 28. As
shown in a previous section, the pavement on this site is significantly deformed. Just before the
section of the crash photograph in Figure 63 (taken at RDB section about 11/400 km), the
Drainage Gradient is very low as seen in Figure 84.
Junction with Rd 703 bound for Ljustorp at section 11 497 m
0,0
0,5
1,0
1,5
2,0
2,5
3,0
3,5
4,0
4,5
5,0
11300
11350
11400
11450
11500
11550
11600
11650
11700
RDB Distance [m], lane in reverse direction
G
ra
d
ie
n
t [
%
]
Unacceptably low Drainage Gradient
Drainage Gradient
Figure 84
Unacceptably high skid risk due to too low Drainage Gradient at HS à sÀng
Page 112
ROADEX III The Northern Periphery Research
6.2.5 Many areas with unacceptable skid risk from Ramsele to à kerö
South of Ramsele, there is a 18.5 km long road section that, despite low traffic volume, shows
an increased accident number, as seen in Figure 28. This section has a lot of skid risk areas
with insufficient Drainage Gradients, as seen in Figure 81.
Skid risk: If the Drainage Gradient doesnÂŽt exceed 0.5 %, water pools will be formed
0
1
2
3
4
5
6
7
8
9
125000
127000
129000
131000
133000
135000
137000
139000
141000
143000
RDB Distance [m], lane in reverse direction
D
ra
ina
ge
G
ra
d
e
[
%
].
Drainage Gradient
Unacceptably low Drainage Gradient
Figure 85
Many sections between Ramsele and à kerö show unacceptably low Drainage Gradient
Page 113
ROADEX III The Northern Periphery Research
6.2.6 Also new roads have skid risk areas due to too low Drainage Gradients
For reference purposes, Drainage Gradient (DG) was calculated for a 12.3 km new section on
Hw 90 north of SollefteÄ. This section had been reconstructed totally, after the study by Ahlin et
al (2000) [3]. However, the resulting DG plot was surprising and gave some very valuable
knowledge. As seen in Figure 86, the new road section has 12 skid risk areas; one per km. The
black Curvature trace clearly shows that all skid risk areas are located at the entrances or exits
of left hand curves (having negative curvature). No skid risk areas can be seen at right hand
curves or on straight sections. Tests on data from highways and expressways in various parts of
Sweden demonstrate that this new knowledge on skid risk hot spots has a generic application.
0,0
1,0
2,0
3,0
4,0
5,0
6,0
0
2000
4000
6000
8000
10000
12000
Distance [m]
G
ra
d
ie
n
t [
%
]
-3,0
-2,0
-1,0
0,0
1,0
2,0
3,0
C
u
rv
at
u
re
=
1000 /
R
a
d
iu
s
[
m
]
Drainage Gradient
Unacceptably low Drainage Gradient
Curvature
Figure 86
The new section on Hw 90 has unacceptable skid risk at left hand curves
Page 114
ROADEX III The Northern Periphery Research
6.3 FRICTION ISSUES DUE TO LOW OR VARIED MACROTEXTURE
6.3.1 Low wet friction at the extremely Hazardous Site Stavreviken
The spot with most skid accidents in the VĂ€sternorrland County is HS Stavreviken. Numerous
skid accidents take place there every year. In one week there were three accidents. In
southbound direction, the road makes a long and steep downhill grade, finished by a hairpin
curve over a railway. The skidding incidents take place at the end of the grade, at about RDB
section 5/380 km, just before the hairpin curve begins. Most skidding vehicles crash within a
zone of 10 metres length. The SRA Central Region is planning to solve the troubles with HS
Stavreviken by building a 2.1 km new road and railway bridge section at a cost of about 3 MâŹ
[71].
Mahone (1975) [55] showed that the friction in hard emergency braking at highway speeds on
wet road surfaces is mainly determined by the surface Macro Texture (MaTx). At HS Stavre-
viken, all vehicles must brake hard to keep speed low in the grade. Many vehicles brake with
significant tyre slip, thereby polishing the road surface. In Figure 87, Macro Texture values from
left and right wheel track are reported from HS Stavreviken. As seen by the graphs, the values
seldom exceed the benchmark minimum level of 0.6 mm.
A low cost action to increase the Macro Texture and reduce the skid risk in wet conditions could
be a double surface dressing. There are also extremely skid resistant special surfacings avail-
able. One such surfacing, based on steel slag, is currently being tested in Dalarna County within
the SRA Central Region [66]. A pair of speed-activated âYour speedâ displays could be benefi-
cial at the top of the grade. If initial speeds were lowered, the need for braking in the grade
would be much less.
RDB 5216 m at junction w Rd 684 (on which Rd 331 then is guestroad) in Stavreviken
0 m at junction with E4
0,0
0,2
0,4
0,6
0,8
1,0
1,2
1,4
1,6
1,8
2,0
5200
5300
5400
5500
5600
5700
5800
RDB Distance [m], lane in reverse direction
M
a
cr
o
T
ext
u
re M
P
D
[
m
m
]
MPD Right
MPD Left
Minimum Macrotexture
Figure 87
Insufficient Macro Texture at HS Stavreviken
Page 115
ROADEX III The Northern Periphery Research
6.3.2 Hazardous Split Friction due to patch repair in one wheel track only
Split Friction (SF) is an extremely hazardous condition, known to cause instability phenomena
such as jack-knife and trailer swing when braking hard with heavy vehicle combinations [33]. SF
happens when the friction is much lower in one wheel track than in the other. SF may be difficult
to recognize when cruising or braking normally. However, it is detrimental when (emergency-)
braking hard. When doing so, the vehicle rotates over the wheel track offering high friction.
ABS-brakes reduce the instability problem, but at the price of a longer braking distance. An ex-
treme example of SF is ice in one track and bare asphalt in the other.
SF may occur after a patch repair in one wheel track only. Such a repair can result in large
differences in colour, as well as in MaTx, between the wheel tracks. This can create very high
SF condition, especially in mornings after a night with temperatures slightly below 0 °C. When
this happens the road surface can become covered with thin ice. As the sun rise, its radiation is
absorbed by the black bitumen-rich patches so the ice on these thaws quicker than the greyish
old asphalt in the non-patched track. When braking hard on such split friction surface, the result
may be a skid into the ditch or over to the opposite side of the road.
A photograph fro m HS MeÄstrand is shown in Figure 61; take note of the glare slick patch repair
in the right wheel track. Profilograph results for the siteâs MaTx are showed in Figure 88. The
right side vertical scale is for MaTx, while the left is for the turquoise Split Friction risk indication
trace. SF risk indication is defined as difference in MaTx in the wheel tracks, divided by the low-
est MaTx of the two tracks. While there are several sections with low MaTx in the right wheel
track, there are fewer sections with high SF risk. The most hazardous section is at 125/770 km.
On investigation, Split Friction due to asphalt patch repair has been identified as a likely causal
factor behind five skid accidents within two weeks after patch repair at a curve on Hw 61 in
VĂ€rmland, Sweden.
Figure 88
Macrotexture values indicating low and split friction (due to patch repair in only one track)
Page 116
ROADEX III The Northern Periphery Research
Chapter 7. Ethical aspects on safety issues
7.1 VISION ZERO FOR ROAD SAFETY
âVision Zeroâ
is the basis for all modern road safety work in Sweden. The approach was ratified
by the Swedish Parliament in 1997, and has resulted in changes to road safety policy and the
work undertaken.
Since
Vision Zero
was established in Sweden, fewer people have been killed on roads. Now the
ideas behind
Vision Zero
have also had an international breakthrough.
Vision Zero
is an i mage of a future in which no one will be killed or serious injured. It is both an
attitude to life and a strategy for making a safe road transport system. Road safety in the spirit
of
Vision Zero
means that roads, streets and vehicles must be much more adapted to human
capacity and tolerance.
The responsibility for safety is shared between those who design, and those who use the road
transport syste m.
7.2 THE TYLĂSAND DECLARATION
The Tylösand Declaration lays down the principal rights of citizens for road traffic safety. These
rights serve to protect the m from the loss of life and health caused by road traffic. They rest on
the general assumption that no road user wishes to harm either himself or herself or any other
fellow human being, whatever the circumstances under which they are using the roads.
The Declaration was signed at the annual conference in Tylösand 2007, by Jörg Beckmann
(Executive Manager of European Transport Safety Council), Ă
sa Torstensson (Swedenâs
Minister on Infrastructure), Ingemar Skogö (Director-General of Swedish Road Administration),
together with other decision makers and experts within Europe as well as fro m other continents.
Page 117
ROADEX III The Northern Periphery Research
The Declaration includes five articles:
1. Everyone has the right to use roads and streets without threats to life or health.
2. Everyone has the right to safe and sustainable mobility: safety and sustainability in road
transport should complement each other.
3. Everyone has the right to use the road transport system without unintentionally imposing
any threats to life or health on others.
4. Everyone has the right to information about safety problems and the level of safety of any
component, product, action or service within the road transport system.
5. Everyone has the right to expect systematic and continuous improvement in safety: any
stakeholder within the road transport system has the obligation to undertake corrective
actions following the detection of any safety hazard that can be reduced or removed.
7.3 PRIORITIZING VARIOUS ROAD SAFETY IMPROVEMENTS
It is well known, that one, or several, of the following factors are involved in the vast majority of
traffic accidents:
1. Drugs, including alcohol, narcotics et c.
2. High
speed.
3. Not using a seat belt.
4. Suicide.
Thus, it is rational that most road safety improvement actions should be focused on reducing the
above human factors, mainly no 1 - 3.
However, acting rationally does not always mean the same as acting ethically.
The above listed factors have much in common: most road users are aware of risks associated
with factors listed; they have personal control over each of the factors; and, finally, road users
18
have decided to expose themselves to the risks. Taken all together, this means that road users
should be able to take large responsibility for these risks.
Below is a list of other factors involved in traffic accidents. These also have much in common:
many road users are unaware of them and/or their association with risk; it is difficult for road us-
ers to exercise control over the associated risks; road users have generally not decided to ex-
pose themselves to the risks. Taken all together, this makes it difficult for the road user to take
responsibility for this second list of risks. However, a fourth co mmon feature is that the road
18
An exception is the second party, suffering from actions by the causal individual. One example is an
âinnocentâ driver crashing due to a drunk driver over-speeding at the wrong side of the road.
Page 118
ROADEX III The Northern Periphery Research
agency is able to exercise control over all of the factors. The control option makes it obligatory
for road agencies to take responsibility for the risks listed below:
1. Longitudinal or lateral roughness, causing ride vibration related phenomena such as
driver fatigue, bump steer and loss of friction due to weight transfer.
2. Bumps, without a warning sign.
3. Poorly banked curves, not giving relevant lateral support for cornering vehicles.
4. Pavement local areas with too low drainage gradient, where slippery water puddles
ponds and in the winter freeze to ice spots.
5. Split friction between left and right wheel tracks during (hard) emergency braking from
high speed and on a wet (thin ice) surface, caused by different texture after spot repair in
only one wheel path.
Considering the above discussion on the individualâs responsibilities versus the road agencyâs
responsibilities, it may actually appear more ethical to spend road agency funding on road
repair, rather than on rational campaigns aimed at reducing drunk driving, over speeding and
reminding to buckle up. It seems i mportant to discuss this balance further, both in- and outside
road agencies.
Page 119
ROADEX III The Northern Periphery Research
Chapter 8. Serious and useful findings
8.1 RIDE VIBRATION SHALL BE PREVENTED AT THE SOURCE
The Rd 331 case study has shown that many professional truck drivers are likely to be exposed
to a daily vibration exceeding the EU Action Value A(8) = 0.5 m/s
2
. Timber hauliers like
Brorssons Ă
keri AB are now obliged, by law, to carry out a risk assessment and i mplement
organizational and/or technical actions to minimize the driverÂŽs vibration exposure. These
actions will bring significant costs to hauliers and their customers in the forest industry.
Why is truck ride vibration a problem at all? Shouldnât vehicle manufacturers be able to isolate
all ride vibration?
Road managers are likely to ask these kinds of questions, and more, after
reading the daily vibration exposure results from this research project.
The answer is that if feasible technological solutions were at hand, they would already be a suc-
cess on the market. The vehicle industry, unlike the average road agency, spends very large
resources on product development and their engineers work hard to continually develop new
solutions and improvements to overcome perceived problems. However these organisations
work within many constraints, such as commercial aspects, handling and stability. The net effect
of their i mprovements to vibration in vehicles is therefore typically small, when compared to the
potential improve ments by road repair. A good example of this can be seen in the case study,
where the truck driver seat vibration did not change after a missing chassis suspension damper
bush had been replaced (see section
5.2.6 The broken truck suspension bush had no significant
effect
).
The SRA has thousands of kilo metres of roads in a condition si milar to Beaver Road 331, used
for the case study. Thousands of kilometres of roads are likely to be in a similar condition also
across the ROADEX partner areas. The truck response recorded on Rd 331 includes very high
roll vibrations at frequencies below 5 Hz. The
Handbook of Vehicle - Road Interaction
[52],
states that roll motions at frequencies under 5 Hz are not common when driving heavy trucks on
roads with ânormalâ roughness and at normal speeds. This implies that roads in this kind of con-
dition should be considered as non-compatible with normal heavy vehicles.
Article 5.1 in directive 2002/44/EC states: â
Taking account of technical progress and of the
availability of measures to control the risk at source, the risks arising from exposure to
mechanical vibration shall be eliminated at their source or reduced to a minimum
â.
A similar conclusion was made in the five-year US research program â
Ride Quality of Commer-
cial Motor Vehicles and the Impact on Truck Driver Performance
â, performed by leading re-
searchers, road authorities, vehicle manufacturers, hauliers and commercial drivers;
Vibration
Page 120
ROADEX III The Northern Periphery Research
must be eliminated at source through effective road maintenance rather than merely dampened
.
See section â
3.1.6 Ride vibrations have a negative effect on traffic safety
â.
8.1.1 Long wave unevenness require improved road repair methods
The case study showed that low frequency, below 3 Hz, truck ride vibration is a serious
problem. Low frequency vibration is difficult to reduce, both in todayâs vehicle fleet and with all
currently demonstrated truck suspension solutions.
Because of this road managers should pay particular attention to the prevention, and repair, of
those forms of road damage that cause low frequency ride vibration. Vibration at 0.5 - 3 Hz
frequencies relates to road unevenness with 5 - 40 m long wavelength. Asphalt pavers cannot
repair such long wave uneveness efficiently, since the paver only ârides alongâ in waves longer
than the paver itself. Repair of long wave unevenness requires a more advanced approach than
seen in many current road maintenance practices. Milling machines and asphalt pavers must
therefore be âforcedâ by machine control systems to follow a carefully engineered repair design.
This type of solution relies on two prerequisites:
1. A carefully engineered (computer aided) design of the geo metric asphalt repair works.
2. Asphalt pavers being operated with a suitable machine control system, such as used
when repaving airfield runways.
Current standard road repair practice cannot repair long wave unevenness efficiently.
8.1.2 Excessive Cross Slope is an ergonomic problem
Based on the testing carried out in the current trials it appears likely that many professional
truck drivers in the Northern Periphery of Europe are being exposed to higher health risks than
truck drivers in central Europe. Not only is the ride vibration level high, but many truck drivers in
the Northern Periphery may also be sitting in an awkward side sloping position due to excessive
pavement cross slope (CS) on straight road sections. In the case study on Rd 331, many
straight road sections had a CS that exceeded the maximum superelevation/banking allowed in
the tightest horizontal curves. This excessive side slope causes the spine to be twisted, which is
not only uncomfortable, but also makes the back more susceptible to Whole-Body Vibration
exposure. Excessive CS has not been reported as a systematic problem in central Europe.
8.2 BUMPS ARE MOST UNHEALTHY
Transient vibration (shock) is much more detrimental to health than stationary vibration. Many
bumps give shocks that can be compared to those recorded on city bus driversâ seats when
they driving at 30 - 50 km/h over traffic calming speed bumps with 1 dm height. The worst
bumps in the current tests were located on s mall roads, such as on the road to the Sawmill in
Graninge. These bumps, when driven at low speeds of about 40 km/h, exposed the truck driver
to spinal compression stress S
ed
of over 0.5 MPa. This stress level corresponds to health risk.
Also on the âmain roadâ Rd 331, truck drivers drove over many bumps that excited significant
transient vibration. These bumps were due to settlement at old culverts, poorly reconstructed
culverts and settle ments at bridge joints.
Page 121
ROADEX III The Northern Periphery Research
8.3 ROLL VIBRATION REQUIRE SPECIAL FOCUS
Many of the Hazardous Sites in the case study were found to have local severe pavement edge
da mages, characterized by high Rut Bottom Cross Slope Variance (RBCSV).
Repair of these pavement damages will minimize lateral vibration in trucks. (Truck suspension
systems cannot isolate such vibration). This kind of road repair will bring better health and
safety to professional truck and bus drivers. It will also improve safety for fellow road users, due
to the reduced risk of collision with skidding trucks.
Road damages resulting in high variance of Rut Bottom Cross Slope (RBCSV) has been
identified as a critical factor behind truck rollover accidents by the group analysing lethal
crashes in the Central Region of Norway. Repeated lethal crashes (2005, 2006) have occurred
in an âeggshapedâ curve at SmalĂ„sen on road E6, where a culvert bump reduced the Cross
Slope in a section where the curve is tightened (and the lateral force increased). Deformed
pavements have also been found to be a cause in accidents, as the deformations affect the driv-
ing stability of the vehicles. Deformations have been found very hazardous to cars with low pro-
file tyres, motorcycle traffic and at slippery road conditions [72] [Personal communication with
Mr BĂ„rd Ăien, head of the crash investigation group in Central Norway].
8.3.1 Lateral vibration
Lateral acceleration is commonly recognized as a key para meter for vehicle driving stability, and
thus for traffic safety. This is especially relevant on slippery surfaces, where the lateral friction
forces are small. When a vehicle changes its roll angle quickly, the roll motion is accompanied
by lateral acceleration. Results fro m the case study show that severely deformed pavement
edges are a serious safety hazard, as they may result in lateral acceleration forces co mparable
to the lateral forces experienced when travelling around a horizontal curve.
The high lateral vibration seen in the case study raises a question whether it is sufficient to have
truck suspension systems that deal with vertical vibration alone. There seems also to be a
similar need to prevent and/or isolate lateral vibration. This is a significant challenge to truck and
seat designers, since the conflicts with traffic safety are obvious. Any additional isolation
systems to the present provision will however increase the deadweight of the vehicle and thus
reduce the payload that can be carried, thereby increasing the number of trucks necessary to
meet a given transport need. The development of new efficient solutions will cost money, and
new components also bring new costs. The net effects of this is that new systems for isolation of
lateral vibration are likely to be accompanied by increased transport costs, and increased
number of trucks on the roads.
The conclusion is that the cross slope variations on badly deformed EU Northern Periphery
roads makes them incompatible surfaces on which to drive normal heavy vehicles. Such road
sections should be repaired as soon as possible.
Page 122
ROADEX III The Northern Periphery Research
8.3.1.1 Tramp-related polishing creates extremely slick areas
Road roughness with short (0.7 - 3 m) wavelengths causes truck tyre resonance. Roughness
with these short wavelengths often has a low coherence between the left and right wheel track.
This means, for example, that the left track may have a short bump whilst the right track may be
flat, or even have a depression. This can result in the wheel axle starting to roll, accompanied
by lateral forces and an oscillating motion between road and tyre. Such âtramp-motionâ gives a
polishing effect which, after accumulated truck passes, can make the road extremely slippery.
Short wavelength road roughness is therefore a road safety issue that should be controlled and
kept below safe levels. This kind of road roughness can be repaired by a simple asphalt overlay.
8.3.2 The change in climate calls for increased repair of RBCS damages
Change in climate is likely to make freezing and thawing more frequent in the Northern
Scandinavia. Data from year 1961 - 1990 (left) can be compared with a computer modelled
scenario for 2071 - 2100 (right) in Figure 89. The result shows that the number of days with
temperature shifts of around 0 °C will increase.
Figure 89
Freezing and thawing Dec - Jan; increased number of times temperature passes zero [46].
Slippery âblack iceâ occurs more frequently at temperature shifts of around 0 °C, than at very
cold te mperatures. Thus, extremely slippery conditions will become more and more common on
the rough roads in the Northern Scandinavia.
The combination of slippery surfaces and pave ment edge da mages results in lateral forces and
can be very dangerous. Thus, the need for repairing Rut Botto m Cross Slope Variances
(RBCSV) will increase as climate change continues.
A relevant exa mple is the strong increase in road crashes in the High Coast Ă
dalen (SollefteÄ,
Kra mfors and HÀrnösand) in SRA Central Region, as seen in Figure 90. In the period from 1
January to 10 March, the number of crashes has increased fro m 18 to 23 during the 10 year
period 1998 - 2007. Given the mean value of 23 in the past ten years and considering the
natural variance of this statistic (the standard deviation over the past ten years was 3.7
crashes), there was over 95 % probability of less than 31 crashes in 2008. However, the 2008
outcome was 42 crashes. This is an increase by 109 %, as compared to the ten years before.
The Rescue Leader had reported poorly maintained roads as a causal factor behind more than
Page 123
ROADEX III The Northern Periphery Research
50 % of the crashes in 2008, while 25 % also involved deep ice ruts. The High Coast Rescue
Department also see a clear relation to the extreme and unsteady climate. Most crashes were
single car accidents. Most of these took place on straight road sections in daytime, and many of
the crashed cars were driven by wo men. [Personal communication with Peter Carlstedt, Head of
High Coast Rescue Department] [73].
Road traffic crashes in the High Coast
SollefteÄ, Kramfors and HÀrnösand
From 1'st of January to 10'th of March
42
23
23
14
17
27
20
22
18
18
19
0
5
10
15
20
25
30
35
40
45
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
Year
N
u
m
b
er
o
f cr
ash
e
s
Num ber of road traffic crashes
Moving average o ver 2 years
Figure 90
A strong trend of increased road traffic crashes
Page 124
ROADEX III The Northern Periphery Research
8.4 SOME ROADS ARE MORE HAZARDOUS â NOW WE KNOW WHY
The objectives for this project focused on health issues but the work also gave valuable spin
offs in safety issues. From the overall accident records, it is obvious that there is still a long way
to go before âVision Zeroâ can be reached. The findings in this study however show that there is
great potential to reduce accidents by make existing road surfaces safer.
Antiskid systems in cars were recently recognised to be as important pieces of safety equipment
as seatbelts [44]. This confirms skidding to be a common and very serious safety risk. Seatbelts
are accepted as being a very successful safety aid, yet society still continues its efforts to pre-
vent traffic accidents. In the same way, society should not rely on antiskid systems as the sole
solution to skid problems. Greater efforts should still be made to make road surfaces more skid
resistant. After all, a vehicleâs braking distance is much shorter on a skid resistant surface, than
in a similar vehicle with antiskid system braking on a slippery surface.
Technological advances in the future vehicle fleet are likely to make large improve ments to
traffic accident outcomes. However these solutions are also likely to require even better friction
than current road surfaces are offering, in order to make use of the full safety improvement
potential. Obviously, it is time to start making the road network more skid resistant!
8.4.1 Incorrect banking cause dynamic imbalance in curves
Many of the Hazardous Sites in the case study were found to have incorrectly banked curves,
causing dynamic imbalance when cornering.
The case study demonstrated that plots of Cross Slope (CS) versus Curvature can be useful
tools when analyzing dynamic balance in curves, using similar safety margins to those
employed in designing new roads. Groups of data, âfamiliesâ, with safe and co mfortable road
alignments in such plots were identified in Figure 72. Data outside these fa milies are from sec-
tions with incorrect Cross Slope / Superelevation.
8.4.1.1 Evaluating / redesigning old roads is different from designing new roads
When designing new roads, specified CS values are used; i.e. 2.5 %, 4 % and 5.5 %. When
evaluating or repaving old roads however using such fixed values has a poor cost-benefit return
as it can be very costly to modify existing CS. For modest Curvatures (absolute values ranging
from 1 to 3 for a reference speed of 70 km/h), the magnitude of CS between 2.5 % and 5.5 %
are of low to moderate importance for safety as well as for comfort. However, it is very important
that the CS does not vary to the extent that high vehicles start to roll, as measured by the new
RBCSV parameter.
The red dots in Figure 91 are inside the tolerance box for the design of new 70 km/h roads,
whereas the green stars are outside the box. The green stars however represent a road section
offering a safer and more comfortable ride, than a section represented by the red dots.
Page 125
ROADEX III The Northern Periphery Research
g
g
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
7
8
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
Curvature = 1000 / Radius [m]
C
ros
s
s
lope
[
%
]
Figure 91
What is worse on an old road: CS outside tolerance boxes, or high RBCS variance?
8.4.2 Large hydroplaning risk at left hand curves, also at new roads
Where the road surface Drainage Gradient (DG) is lower than 0.5 %, water will not run off the
road and water pools will be formed in wet weather. Water ponding, such as seen in Figure 78,
increases the risk of skidding accidents.
Many of the Hazardous Sites in the case study were found to have DG lower than the 0.5 %
lower limit used in road design manuals worldwide. The case study demonstrated that
entrances and exits of left hand curves are hot spots for a low DG. The reason of this was
explained.
A further finding was that even new roads had been designed with very low DG at many left
hand curves, thereby creating an unacceptable skid risk.
Low DG was found to correlate so mewhat with pavement deformation in terms of high RBCSV
and high IRI (the details have not been shown in this report). This indicates that water on the
road, and by i mplication within the pave ment itself, can hasten permanent deformation. Keeping
DG sufficient is therefore a prequisite for keeping the service life ti me costs to manageable
levels. Insufficient DG can bring unneccesary road agency costs through deformation.
Page 126
ROADEX III The Northern Periphery Research
8.5 LOW AND VARYING MACRO TEXTURE CAUSE SKID ACCIDENTS
The location with most skid accidents in the VĂ€sternorrland County road network is HS Stavre-
viken on Rd 331. Numerous skid accidents take place there every year; in some cases several
accidents take place over a few days! Macro Texture values from HS Stavreviken are generally
below the minimum level of 0.6 mm. A low cost action to reduce the skid risk at the site could be
to make the road surface more skid resistant. A surface dressing can often give a very good ef-
fect in terms of increased friction factor. The texture of the existing aggregate may also be reju-
venated by ultra-high pressure water cutting, as investigated by Pidwerbesky & Waters (2008)
[63]. However, the most efficient action is to prevent polishing, by careful selection of polish-
resistant aggregates for those sections with high polishing energy, i.e. tight corners, downhill
end of grades, roundabouts, and junctions et cetera. Using steel slag as asphalt aggregate
could also be one cost effective solution [66] [67].
Split Friction (SF) is an extre mely hazardous condition, when the friction is much lower in one
wheel track than in the other. SF may be difficult to recognize when cruising or braking normally.
However, it can be detrimental when braking hard in an emergency. When doing so, the vehicle
tends to rotate over the wheel track offering high friction. SF may occur after a patch repair in
one wheel track only. The case study demonstrated the use of a new Split Friction risk
indication parameter, based on Macro Texture data from the laser/inertial Profilograph.
8.5.1 Double surface dressings have better Mega Texture
It is well known that double surface dressings give off less noise than single surface dressings
[65]. However, there is a myth that the noise difference is due to the lower Macro Texture on
double surface dressings. The true causal factor is that the double surface dressings have less
Mega Texture (MeTx) than single surface dressings. Lower MeTx levels result in reduced inte-
rior and exterior noise, due to reduced tyre vibration. The reduced tyre vibrations are also likely
to reduce Hand-Arm Vibration to vehicle operators and improve friction. These benefits of dou-
ble surface dressings are not generally recognized yet, but they are likely to beco me more ap-
parent with the increasing use of pavement texture analyses.
8.6 RETHINK CULVERT WORKS
The case study showed that culverts can be critical locations for bumps, giving poor ride quality
and health damage to the spine. There are three problems. First, current culvert repair practices
can be poor resulting in centimetre-deep initial unevenness immediately after repair. Secondly,
poorly compacted backfill may add settlement of several cm within a couple of years. Third,
culverts have been found to collapse at a fraction of the design age. This brings unacceptable
costs to taxpayers.
8.6.1 Poor culvert construction practice
A 6 cm deep hollow appeared immediately after constructing a new culvert on Rd 331 in
Ga mmelmo. The cause of this initial bump is likely to be found in poor construction work, rather
than in deficiencies in the road culvert installation code. It is most unlikely that such a large
settlement could occur in an established road embankment unless the construction work
(material selection, compaction et cetera) was poor.
Page 127
ROADEX III The Northern Periphery Research
8.6.2 Poor culvert-related road surface maintenance management
There is a need for an improved maintenance regime for culverts. Many culverts on the Beaver
Road 331 collapsed during the winter, making the repair even more difficult. âEverybodyâ knows
that there will be settlement when a culvert is constructed or repaired, due to difficulties in the
compaction of the thick backfill, differences in material properties, etc. The case study shows
however that the resulting bumps due to such settle ments are totally unacceptable for road
users, in terms of comfort, health and traffic safety. There a clear need for robust management
of bumps at culverts. Road users can probably be expected to cope with a culvert bump for a
few months. However, it is a modest demand that culverts should be inspected for road
roughness, in the first and the second year after reconstruction, so that repair of any local
roughness can be carried out in a timely fashion where necessary. It is reco mmended that a
culvert repair should always be followed by systematic roughness inspection and additional road
repair where necessary.
8.5.3 Water-piping in permeable culvert foundation beds
Culverts manufactured with Portland Cement Concrete (PCC) are usually manufactured for a
design life of up to 100 years. It is therefore surprising that so many apparently sound concrete
culverts need to be reconstructed within only 5 - 20 years after their installation. As seen on Rd
331, a co mmon failure mode is a full collapse, caused by water piping in the soil below the cul-
vert. This shows a need for a revision of the culvert design code. Can water really be expected
to flow within a culvert, when the culvert itself is founded on a permeable gravel bed at the bot-
tom of the culvert ditch in low-permeable soil (so water can pipe its way beneath the culvert), as
seen in Figure 92. Should not the foundation need to be made low-permeable? These are ques-
tions that culvert experts should consider.
Figure 92
Section of a culvert on its permeable foundation bed [40]
Culvert
(Permeable) foundation bed
Page 128
ROADEX III The Northern Periphery Research
Chapter 9. How to use the new insight
This chapter sets out recommendations for stakeholders across in Northern Periphery, from
vehicle manufacturers to road agencies, based on the project results. It is hoped that these will
be accepted in a constructive fashion, improving the present situation for the benefit of all.
9.1 HAULIERS MUST MONITOR DRIVERS WBV EXPOSURE
The daily exposure value A(8) for Whole-Body Vibration (WBV) must be determined for truck
drivers. The value should be representative of year round operations and measurements should
include driving during winter. In good winter conditions, the road can be smoother than in the
summer. Conversely poor winter condition, through poor road maintenance, can result in
significantly higher ride vibration. The A(8) value will therefore be depending on the road main-
tenance standard, both in summer (pavement condition) and in winter (snow ploughing). A fur-
ther factor to be considered is vibration when driving on road sections under reconstruction.
Discussions should be held with the owners of local roads in poor condition, such as the access
road to the Graninge Sawmill, regarding the condition of the road. If the worst bumps on the
Graninge access road are not repaired, it may be necessary to close the road to heavy trucks.
Where alternative routes are possible, a longer route with lower roughness / vibration could be
an option. An example in the case study is the smoother route on Hw 87 - Hw 86. This route
could be used instead of the rougher, but shorter, Rd 331. An i mportant question that arises
here is â
Who is prepared to pay for the additional costs for the longer route
â? How can
competing truck hauliers be equally treated?
An efficient tool to reduce WBV on roads with excessive shortwave roughness is tyre pressure
control or Central Tyre Inflation (CTI) syste ms. (CTI cannot isolate long waves). Using CTI it is
possible for the driver to change the pressures in the tyres of the vehicle while driving, and this
has proven to reduce WBV by 8 % on four test roads in central Sweden.
In the case study, a chassis suspension da mper bush was out of order, as seen in Figure 30,
without anyone being aware of the problem. This lack of detection indicates a need for i mproved
vibration control in vehicles in truck fleets.
Hauliers in the Northern Periphery are recommended to use special winter tyres. Brorssons
Ă
keri AB use the new Michelin XFN+ winter tyre on the steer axle, offering 10 % better side
friction against the road surface and 5 % shorter braking distance.
Page 129
ROADEX III The Northern Periphery Research
9.2 DEVELOPING USEFUL NEW VEHICLE TECHNOLOGY
On-board vibration loggers could be useful for hauliers, especially as they are obliged by law to
assess the risks to their drivers from vibration exposure. One proble m with seat pan mounted
automatic logging syste ms is that the driverâs ingress and egress may cause data artefacts
(false âshocksâ, resulting in unacceptably high S
ed
values), as discussed by Mansfield & Newell
(2004) [59]. A more robust solution could be to have the vibration sensor at the cab floor, and
calibrate it to predict seat pan vibration. This may require a calibration that takes account of the
driverâs weight, and should be further investigated. The vibration logger could also be managed
by a condition stating that vehicle speed must exceed a specified minimum, e.g. 5 km/h, before
vibration data is stored.
As seen in the case study, many EU Northern Periphery roads give high lateral vibration.
Currently heavy vehicles are not so good in isolating lateral vibration. Increased efforts on
preventing and isolating lateral vibration could therefore be beneficial. The potential use of
MagnetoRheology (MR) technology in trucks should be further investigated. However it is
important not to imple ment solutions that increase bounce, when decreasing roll.
There is a need for a declaration of a vehicleâs vibration emission value to be made, so that
drivers can be in a position to request the most appropriate truck to be purchased. A general-
ized test is defined in the EN 1032 standard [57].
Road profile data from laser/inertial profilo meters can be used to develop even better trucks in
the future. At the date of writing, Profilograph data from Rd 331 is being used in a road simulator
hydraulic test rig at Volvo 3P. Figure 93 shows a photograph from a simulation of a very rough
road, causing bounce vibration as seen by the photographed truck vertical âtracesâ. The simula-
tion shown was of obviously made on a road without pavement edge damage, otherwise roll
traces would have been seen as well. Severely damaged road sections, such as at HS Ă
kerö on
Rd 331, may require road simulators to have larger hydraulic ranges than the current versions.
Figure 93
Truck ride vibration tests in a hydraulic road simulator [Photo: Volvo 3P]
Page 130
ROADEX III The Northern Periphery Research
9.3 IMPROVING ROAD TRAFFIC CRASH INVESTIGATIONS
A wise man said: â
When investigating road traffic crashes, it is important to define a clear
objective. Investigators trying to identify who to blame, tend to search after deviant behaviour.
Investigators trying to understand normal road users need for better technology and
infrastructure, tend to search after repeated patterns
â. Which of these strategies has the best
potential to support improve ments into a safer future road transport syste m?
As seen in the case study results, each of the Hazardous Sites at Rd 331 show re markable
properties in the laser/inertial Profilograph road condition measurement. With exception of
crashes with obvious causes, such as suicides, crash investigators should study pavement
profilometer data on a routine basis.
Key safety parameters should include Cross Slope (CS) by magnitude and undesired variance
(RBCSV), dyna mic imbalance due to suboptimal combinations of CS and Curvature in
incorrectly banked curves, high Curvature (lateral force), insufficient Drainage Gradient (hot
spots at left hand curves and at deformed pavement sections), excessive Mega Texture (MeTx),
insufficient Macro Texture (MaTx) and heterogeneous MaTx causing Split Friction. It is
recommended that all of these parameters should be analyzed in road safety ratings, such as in
the Euro Road Assessment Programme.
The work of improving road network safety could gain much from being benchmarked with
aviation safety work. Before using an airfield for international air traffic, the facility administrator
must demonstrate that the runway surface provides the mini mum required level of safety, as
defined by the International Civil Aviation Organization. Today, aircraft seldom crash due to
deficiencies in the runway condition.
In Norway, all lanes of all paved highways are profiled annually in both directions. In Sweden,
only one lane in one direction is monitored at least every fifth year. One consequence of this
lesser measurement strategy is that when a serious crash occurs, it is often necessary to carry
out extra profilometer measurements to obtain sufficient accurate data for the investigation. It is
not possible to take laser/inertial measurements on icy roads or at te mperatures below 0 °C.
Page 131
ROADEX III The Northern Periphery Research
9.4 IMPROVED ROAD MANAGEMENT
It should be a top priority within a road agency to recognise the importance of good road
condition to comfortable, stress-free, healthy and safe road use.
A key factor in the reduction of professional driversâ daily exposure to vibration A(8) is the
effective reduction of road roughness. This calls for a good focus in the selection of road repair
sections, the planning of repair methods, the performing repair work, and in the end control.
The profilometer data held in a road agencyâs Pavement Manage ment System (PMS) is a
powerful tool for the management of paved roads, and its use should be encouraged and
developed. Special training courses in the application of data would be beneficial, as would
interregional, and international, benchmarking between road agencies local offices.
As seen in the case study, much of the truck ride problem relates to long wave unevenness and
pavement edge deformations. The new Rut Botto m Cross Slope Variance (RBCSV) parameter
should be implemented immediately in PMS. Effective repair of long wave unevenness and
RBCSV is a new target for the road maintenance sector, requiring new methods. Detailed
drawings should be made per 5 m section, showing target values for asphalt overlay thickness
(including depth of grinding, if milling machines are to be used) and redesigned Cross Slope. A
computer aided method for this is already in practice for high volume roads. This method should
also be used on low and medium volume roads. Asphalt machines should be equipped with
machine control systems, so they can perform the designed repair work effectively. Such
systems are in use on airfield runways and high volume roads. The time has come to i mplement
similar solutions on low and medium volume roads as well.
Why does such a large proportion of the road length in Sweden have severe pavement edge
deformations as a result of weak road shoulders? Perhaps the design of pavement
edge/shoulders should be reviewed. Is the quality of the road materials too poor? Are the road
structure layers too thin? Could the reason be insufficient shoulder width and/or too steep em-
bankment? These questions need to be answered. The deformed pavement edge at HS
MeÄstrand on Rd 331 in Figure 61 does not compare favourably with the stable Danish road
edge design shown in Figure 94.
Figure 94
A Danish stabile pavement edge with paved shoulder and a wide grass verge
Page 132
ROADEX III The Northern Periphery Research
The SRA Northern Region has been using laser/inertial profilometry in the control of roughness
of new pave ments for almost 10 years. Their experience is that the technology can result in
much smoother new pavements without raising the price of paving. The outcome is lower ride
vibration, longer pave ment service life, and thereby lower road lifetime costs.
Good road maintenance practice could be encouraged by giving an award to the âs moothest
resurfacing project of the yearâ. This is already an accepted practice in Norway, but not in
Sweden. Such an award would gain extra attention and status, if sponsored by stakeholders
using the road, such as a truck haulage association.
Good practice amongst contractors could also be encouraged by openly reporting profilometer
measurement results after surfacing operations, possibly accompanied by a comment fro m the
project manager on any gap between target and outco me.
In the Northern Periphery the daily vibration exposure A(8) can be affected by winter
maintenance operations. Poor snow ploughing response times can result in the formation of ice
roughness on the surface of the road, which can cause intense vibration in the vehicle. Such
vibrations can also cause interior noise. Professional drivers are exposed to many stress
factors, including vibration and noise. Stress can also occur in the internal conflict in a driver
when he, or she, has to decide whether to reduce vehicle speed to match poor road conditions,
and thereby delay a delivery, or continue to try to meet the schedule, possibly as an accident
risk. As presented in this report, researchers suspect the prevalence of increased stress
hormones in the blood to be the cause of the strongly increased prevalence of myocardial
infarction among the drivers. Therefore it is important to reduce as many stressors as is
possible. Some types of stress can be reduced by good information on the route conditions, but
the prevention of poor road conditions should not be underesti mated.
A special concern is transient vibration/shock at bumps. A special progra m should focus on re-
pair and prevention of bump hot spots, such as at culverts (see section
8.6 Rethink culvert
works
), bridge joints, frost related deformations and potholes.
Roughness after culvert repair should be carefully monitored for the following three years.
Culverts requiring emergency repairs during the winter should be revisited in the first summer.
A significant share of the road network is repaired each summer. Road roughness can be
extreme during road repairs, and this can contribute to high vibration exposure A(8) to truck
drivers. It is therefore important to try to restrict road repair section length and maximum
roughness levels during reconstruction works.
It could be cost-efficient to make a special contract for bridge joint roughness repair. This could
permit the successful contractor to assemble a team of specialists to repair the settle ments at
low cost in the spring, ahead of the main paving season.
Page 133
ROADEX III The Northern Periphery Research
Transient wheel vibration can be caused by poor joints in patch repairs. The case study shows
several examples of some 2 cm high joints, causing wheel axle bounce and tra mp-related pol-
ishing resulting in very low friction. It is recommended that no patching should be carried out
without 2 cm deep edge grinding, to enable a proper joint with the adjacent road surface. Only
very local patches such at potholes should be exceptions from such a practice. Preparatory
edge grinding should be carried out at the repair of bumps at culverts and of pavement edge
deformations. An example of small size grinding machines now available on the market is
shown in Figure 95.
Figure 95
A small asphalt grinder [Photo: Tobias Edberg, SRA Production]
Double surface dressings should be considered in preference to single surface dressings as
they have less Mega Texture (MeTx).
Maximum limits for MeTx should be implemented as soon as possible.
The current standards for laser/inertial road condition profilometry of newly laid pavements
should be revised. The Swedish profilometry standard âVVMB 116â does not require reporting of
key safety parameters such as Curvature (plots of Cross Slope vs. Curvature), Longitudinal
Gradient (to be combined with CS, when calculating Drainage Gradient), Rut Bottom Cross
Slope Variance, Mega Texture and Macro Texture.
Pavement condition data should be stored in 1 m steps, rather than the present 20 m steps, or
longer, used by the national road administrations in the Northern Periphery. Should it be
required, this new style of data can be readily re-calculated as a ârunning 20 mâ value for
comparison with old 20 m data. Such results are still â20 m valuesâ, directly comparable with
existing data, limits and preferences. However, ârunning 20 mâ values with a 1 m update step
length are much better in reflecting local bumps, than traditional 20 m values with a 20 m update
step length are.
Road workers can be exposed to unacceptable Whole-Body Vibration (WBV). Typical examples
of these are drivers of snow ploughing trucks and operators of asphalt paving machines, as in-
Page 134
ROADEX III The Northern Periphery Research
dicated in Table 2. Road agencies should start to measure WBV for these workers. In Sweden,
drivers of contracted snow-ploughing trucks have filed complaints in respect of newly milled
rumble strips in the centre of many roads. These are alleged to cause work related health prob-
lems due to their excessive ride vibration and noise.
The Tylösand Declaration states that road administrators are obliged to identify Hazardous Sites
(HS) quickly, warn road users, and make appropriate repairs.
The proper identification of Hazardous Sites on low volume roads is an important issue. The
case study demonstrates that there is a strong correlation between accident black spots and
poor road condition. However, many of the worst pavement damages were in the section Backe
- Ramsele which does not show any major black spots. The reason for this is the very low traffic
volume on the section. It has an AADT of less than 350 vehicles per day. Obviously there is an
urgent need to use the âIndividual Riskâ approach for low volume road networks. In this ap-
proach, it is not enough to analyze the number of accidents; the numbers must also be normal-
ized to (divided by) the AADT figure. The Individual Risk approach is promoted by road user or-
ganizations, such as the Royal Automobile Club of Victoria, as it increases the likelihood of
identifying very Hazardous Sites on low volume roads [64].
One drawback with the Individual Risk approach is that the low numbers involved make the
assessment more susceptible to randomness. A way to increase the accident numbers, and
thereby reduce the influence of risk, is to include data from registers of insurance companies.
These registers hold more data than databases such as STRADA in Sweden where only Police
and Hospital reported crashes are registered.
Another method to identify Hazardous Sites is to analyze road condition data from laser/inertial
profilometers, see the previous section for examples of key safety parameters. The collation of
accident data from STRADA versus Profilograph data shows a valuable potential that should be
further explored. However, in the case study many examples were found of poorly positioned
crash records in STRADA, including lethal crashes such as at HS S Viksjö. A separate project
should seek to improve the quality of such crash record positions. It is also important to develop
and refine generalized relationships between road condition and accident risk, such as in Figure
5.
A further option is to ask road users to identify their perceptions of hazardous sites. Focus
groups, incorporating regular road users such as Brorssons Ă
keri AB on Rd 331, may be a good
tool for this. In Finland, the Internet-based âStreet Channelâ / KatuKanava [56] is used to map
road user opinions. Improved custo mer focus by roads management may also bring better road
user satisfaction.
There is a need for better standardization in setting up warning signs on roads. At present there
are no objective limits for erecting a bump warning sign. Such a limit should be defined with re-
spect to road condition data from profilometers as well as a subjective decision on âneedâ. Simi-
larly, there should be objective limits for when to warn for high Curvature (lateral force), rutting,
incorrectly banked curves, and skid risk due to too low, or varying, Macro Texture. There should
also be li mits on the length of road section that can have a Drainage Gradient below 0.5 m, be-
Page 135
ROADEX III The Northern Periphery Research
fore a warning should be given. It is impossible to sign for all of such flat areas, as they currently
exist at almost every left hand curve (right hand curve in the UK).
It is not enough however to identify Hazardous Sites and put up additional warning signs. The
road network needs to be more skid resistant and less unhealthy, by focusing actions at the
identified hot spots. This requires a long-term program, with significant funding. Such a program
could benefit of a benchmarking with Transit New Zealandâs Truck Ride Improvement Initiative,
running with a 3M$ annual budget since 2001.
Cross Slope (CS) is identified as a key factor for skidding and its modification requires
substantial amounts of road material, and consequently substantial funding. For these reasons it
is recommended that CS should also be analyzed on a road network level. The new RBCSV
parameter is suitable for such a purpose as it is easy to interpret. A higher RBCS value shows a
higher need for road repair and greater funding.
A slightly more difficult analysis is that of dynamic imbalance due to the ratio of CS to Curvature,
the âincorrectly banked curvesâ in the report. Further research should seek methods to quantify
the problems with incorrect banking at the road network level. This is important, as the recon-
struction of slopes in existing curves requires significant quantities of road materials and thus
significantly larger funding than traditional overlays of ruts and short wave roughness. Plots, as
Figure 73, should be employed in the programming, planning and detailed design of the repair,
as well as in the quality control of finished work.
An easy-to-use parameter is the Drainage Gradient (DG). This is a key safety parameter as
de monstrated in the case study and should exceed 0.5 % to avoid water ponding proble ms.
Road agencies with DG in their databases will find it easy to identify those sections that have
DG below the safety limit 0.5 %.
Suitable software can identify road sections with insufficient DG from existing databases and
during a typical search it can also be possible to identify other flat sections requiring
rehabilitation. Exa mples of the latter type can be found at the HS BjörknÀset and the HS
Helgum. Repair of such weak sections should be designed by support of bearing capacity
testing with a falling weight deflectometer.
Road agencies should require their road designers to report the designed DG, especially at
entrances and exits of left hand curves (UK: right hand curves). Consultants and Contractors
should face high penalties if their work results in Drainage Gradients that are too low.
The limit of 0.5 % DG is tight, and requires high measurement accuracy. DG is calculated from
Cross Slope and Longitudinal Gradient. It is possible to report both of these parameters from
most laser/inertial profilometers on the market. However, most profilo meters do not measure the
roadâs Longitudinal Gradient. Rather they only measure the gradient of the profilometer vehicle
body. The grade of the vehicle and the road grade often differ significantly, especially when the
profilometer vehicle accelerates or brakes. When this occurs the difference can be very large.
The vehicle grade can also change in response to changes in wind load, the level of fuel in the
fuel tank, and other changes in load. The SRA CS Profilograph used in the case study has an
Page 136
ROADEX III The Northern Periphery Research
accurate system for road grade measurement, taking into account the vehicles own pitch angle
in relation to the road. Without such a system, profilometer reported grades/Drainage Gradients
might not be sufficiently accurate to be useful in analyzing the risk for skid accidents. This
should be considered when purchasing road condition measurements.
The allocation of existing road maintenance funding should be reviewed. The repair of
pavement deformation with high RBCSV in the Northern Periphery costs more than overlays of
rough, but much more planar, surfaces in the southern areas. Repair of incorrectly banked
curves and insufficient Drainage Gradient require even more funding, as significant quantities of
road materials are needed. Such road repairs are one-time investments, since slopes, once
created, do not normally change significantly over time.
There should be an extra focus on maintaining high road surface friction in sharp and incorrectly
banked curves and at the downhill part of long and steep grades. This can be done by
increasing the use of high friction surfacings and intensified winter maintenance.
The failure mode at HS S Viksjö shows that the barrier may be undersized. In three lethal
crashes, heavy trucks have made a big hole in the standard crash barrier. There is therefore an
acute need for crash barriers to have the capacity to retain heavy vehicle combinations with up
to 60 tonne gross vehicle weight (GVW). There are plans to increase the max GVW to 80 - 88
tonnes in Sweden. Such plans should be reconsidered on those road networks that do not have
crash barriers with the relevant heavy truck capacity.
Measurements of Split Friction risk potential should be carried out on road sections with new
patch repairs, such as the repair seen in Figure 61. The friction numbers should be measured in
both the left and right wheel track, focusing on the difference between the m.
Profilo meter results should be systematically used in traffic safety inspections and analysis. Left
hand curves (Right hand curves in the UK) are hot spots for hazardous road alignment errors.,
Only one lane in one direction is currently monitored in Sweden in accordance with the road
surface profilometry strategy of the SRA. This lane is scanned at least every fifth year. One of
the net effects of this strategy is that the PMS does not include relevant data for 50 % of the left
hand curves on the network (important geometrical parameters may be totally different in the
opposed directions). Temporary changes should be made in the measurement strategy, ai ming
to result in having relevant geometrical data fro m every lane within three years from now.
A consequence of a limited road condition measure ment strategy is that, in the event of a
serious having to be investigated, it may be necessary to carry out additional profilometer
measurements to obtain sufficiently accurate data for the investigation. Such non-scheduled
measurements are more expensive than syste matically planned measurements.
When planning actions to improve safety at road sections suffering from many skid accidents on
wet pave ment or thin ice, such as HS Stavreviken, resurfacing with high friction double surface
dressings should be tried before planning to build expensive new road sections. Speed
monitoring displays could reduce skid accidents at such sites.
Page 137
ROADEX III The Northern Periphery Research
As recommended by the group analysing lethal crashes in Norwayâs Central Region, the time
tolerance for removing wet snow (which is risky as snow, and later, after forming ice ruts) should
be reduced. Furthermore, the contractor performing daily road maintenance should be paid for
making extraordinary friction improvement actions when weather conditions become extreme
[72].
9.5 ROAD DESIGN POLICY IMPROVEMENTS
Left hand curves should be identified as hot spot sections for insufficient Drainage Gradient
(DG) in road design manuals (right hand curves in the UK).
The minimum limit on DG should be raised. The present design limit of 0.5 % should be
redefined as a rare exception, as it is too close to the normal deviations in road construction
works (about 0.5 % is allowed Cross Slope deviation in the Swedish ATB VĂG road
construction manual). In practice, this means that a âproperly designedâ section of road may end
up with an insufficient DG if the construction on site is only just inside the tolerance limits.
Considering the need for a reasonable production tolerance, the normal design limit needs to be
raised, up to about 1 %.
The RBCSV concept should be implemented in manuals for redesign of existing roads.
Low-permeable material should be considered for culvert foundations.
Page 138
ROADEX III The Northern Periphery Research
9.6 WORK TO BE CONTINUED (IN ROADEX IV?)
This project suggests a draft limit value of 0.30 % for undesired Variance of the pavementâs Rut
Botto m Cross Slope (RBCS). Further work should draft differentiated RBCSV limit values,
depending on road/lane width, curvature and length of curve. It should be noted that the RBCSV
is normalized to a user defined reference speed. Therefore the same limit value can be used for
50 km/h roads as for 90 km/h roads.
How common are the road da mages found on Rd 331 on roads in the Northern Periphery
partner area?
The comfort scale in ISO 2631 is relevant for people in public transportation and there are
indications that professional drivers may have a somewhat higher comfort tolerance. Should a
special comfort scale be developed for professional drivers?
A scale and a limit value should be drafted for short road roughness that causes tramp-related
polishing and thus extremely low friction. Mega Texture could be used as one parameter, but
0.5 - 2.5 m roughness could also be addressed.
The relation between speed and vertical truck seat vibration is fairly well known. But the relation
between speed and roll/lateral vibration should be further explored. (However, there are no
indications on a larger speed dependence on roll, as co mpared to vertical bounce).
The correlation of road surface texture and truck interior noise should be mapped.
Whole-Body Vibration, Hand-Arm Vibration as well as interior noise should also be measured
when driving heavy vehicles on winter roads with ice ruts accompanied by high Mega texture on
the ice edges.
The effect of changing winter road maintenance standards on ride vibration should be
investigated.
The impact of general road roughness should be mapped against the ability to perform efficient
snow ploughing in winter.
Why does such a large proportion of the road length have severe pavement edge deformations?
The design of pavement edge/shoulders could be reviewed. Is the quality of the road materials
too poor? Are the road structure layers too thin? Could the reason be insufficient shoulder width
and/or too steep e mbankment?
A method to quantify the problems with incorrect banking at the road network level could be de-
veloped. Such a method could be drafted from an analysis of plots such as Figure 73.
Page 139
ROADEX III The Northern Periphery Research
Chapter 10. Further reading
[1] Teschke, K., Nicol, A-M., Davies, H. & Ju, S. (1999).
Whole Body Vibration and Back
Disorders Among Motor Vehicle Drivers and Heavy Equipment Operators: A Review of
the Scientific Evidence
. Report to Workers' Co mpensation Board of British Columbia.
[2]
Directive 2002/44/EC on the minimum health and safety requirements regarding the
exposure of workers to the risks arising from physical agents (vibration).
The European
Parlia ment and the Council.
[3] Ahlin, K., Granlund, J. & Lundström, R. (2000).
Whole-Body Vibration When Riding on
Rough Roads â A shocking Study
. Swedish National Road Administration, Vol. 31E.
Internet 2008-02-24:
http://www.vv.se/filer/skakstudie.pdf
.
[4] Sandover, J. (1998).
High acceleration events: An introduction and review of expert
opinion
. Journal of Sound and Vibration 215(4), p 927 â 945.
[5]
ISO 2631-5, Mechanical vibration and shock - Evaluation of human exposure to Whole-
Body Vibration â Part 5: Method for evaluation of vibration containing multiple shocks.
(2004). International standard.
[6] Brandt, A. & Granlund, J. (2008).
Bus Driversâ Exposure To Mechanical Shocks Due To
Speed Bumps
. Society for Experimental Mechanics, IMAC-XXVI Conference on
Structural Dynamics, Orlando, Florida, USA.
[7] Gillespie, T. D. (1992).
Fundamentals of Vehicle Dynamics.
[8] Lindh, H. (2002).
Ride Quality in Trucks â A Dynamic Challenge
. Scandinavian Vibration
Society, Symposium on comfort in vehicles and model verification, RiksgrÀnsen,
Sweden.
[9] Granlund, J., Lenngren, C.A., Lindström, F. & MÄrtensson, B. (2005).
Measuring
Pavement Deflection Variance at Highway Speeds
. 7'th Int Conf on Bearing Capacity of
Roads, Railways and Airfields, Trondheim, Norway.
[10] Forsén, A. (1999).
Heavy Vehicle Ride and Endurance. Modelling and model validation
.
Royal Institute of Technology, Stockholm, Sweden. PhD thesis, TRITA-FKT 99:33, ISSN
1103-470X, 92 p
[11] Ahlin, K. and Granlund, J. (2003).
Relating road roughness and vehicle speeds to
human Whole-Body Vibration and exposure limits
. The International Journal of
Pavement
Engineering, Vol. 3, No. 4, pp.207-216.
Page 140
ROADEX III The Northern Periphery Research
[12]
The AASHO Road Test
. Special Reports 61A-61E. HRB, National Research Council.
Washington, D.C., 1961
[13] Granlund, J. & Lindström, F. (2004).
Reducing Whole-Body Vibration by geometric repair
of pavements.
J. Low Frequency Noise, Vibration and Active Control, Vol. 23, No 2, pp.
103-114.
[14] Ăijer, F. & Edlund, S. (2004).
Identification of Transient Road Obstacle Distributions and
Their Impact on Vehicle Durability and Driver Comfort
. Supplement to Vehicle System
Dyna mics, Vol 41, pp 744-753.
[15]
VĂ€gars och gators utformning. (Design manual for roads and streets).
(2004). Swedish
Road Administration, publ 2004:80
[16]
A Policy on Geometric Design of Highways and Streets.
(1984, 1990, 1994 and 2001).
American Association of State Highway Officials (AASHO), Washington D.C.
[17]
EN ISO 8041, Human response to vibration â Measuring instrumentation.
(2005).
European and International standard.
[18]
ISO 2631-1, Mechanical vibration and shock â Evaluation of human exposure to Whole-
Body Vibration â Part 1: General requirements.
(1997). International standard.
[19]
Ahlin, K., Granlund, J. & Lindströ m, F. (2004).
Comparing road profiles with vehicle
perceived roughness
. Int. J. Vehicle Design, Vol.
36
, Nos. 2/3, pp.270-286.
[20]
Strategisk plan för drift, underhĂ„ll och bĂ€righet 2004 â 2015 (Strategic plan for road
operations, maintenance and bearing capacity)
. Swedish Road Administration, Central
Region.
[21] Granlund, J. (2006).
Nytt mÄtt pÄ tvÀrfall. (How to Measure Rut Bottom Cross Slope and
its Variance)
. SRA Consulting Services, report for Swedish Road Administration, Head
Office.
Internet 2008-02-24:
http://www.vv.se/fudinfoexternwebb/pages/PublikationVisa.aspx?PublikationId=241
[22] Bogsjö, K. (2007).
Road Profile Statistics Relevant for Vehicle Fatigue.
PhD thesis, Lund
University, ISRN LUTFMS-7032-2007, ISBN 978-91-628-7291-5.
[23]
Den goda resan: Förslag till nationell plan för vĂ€gtransportsystemet 2004 â 2015.
Underlagsrapport TrafiksÀkerhet. Handlingsplan för trafiksÀkerhet.
(Proposed national
long term plan for the road transport system â Traffic safety action plan)
. Swedish Road
Administration, VV publ 2003:101
Page 141
ROADEX III The Northern Periphery Research
[24]
Bereton, P.F., Donati, P.M., Fischer, S., Griffin, M.J., Howarth, H.V.C., Kaulbars, U. &
Pitts, P.M. (2006).
EU Guide to good practice on Whole-Body Vibration
.
[25]
Gallaway, B.M. & Rose, J.G. (1971).
The effect of rainfall intensity, pavement cross
slope, surface texture and drainage length on pavement water depths
. Texas
Transportation Institute, Research Report No. 138-5.
[26]
Bigert, C. Klerdal, K., Hammar, N., Hallqvist, J. & Gustavsson, P. (2004).
Time trends in
the incidence of myocardial infarction among professional drivers in Stockholm 1977-96
.
Occup Environ Med; Vol 61, p 987 - 991.
[27]
Perspectives in Disease Prevention and Health Promotion Leading Work-Related
Diseases and Injuries
. US National Institute for Occupational Safety and Health.
[28]
Ahmadian, M. & Ahn, Y.K. (2003).
On-Vehicle Evaluation of Heavy Truck Suspension
Kinematics
. SAE TB 51.
[29]
Guidelines for Air and Ground Transport of Neonatal and Pediatric Patients
. (1999).
Task Force on Interhospital Transport, American Acade my of Pediatrics.
[30]
Campbell, K.L., Erwin, R.D., Gillespie, T.D., Segel, L. & Schneider, L.W. (1982).
Truck
Cab Vibrations and Highway Safety
. Highway Safety Research Institute, University of
Michigan. FHWA report RD-82/093
[31]
Ihs, A., Velin, H. & Wikström, M. (2002).
VÀgytans inverkan pÄ trafiksÀkerheten. (The
influence of road surface condition on traffic safety)
. VĂ€g- och
TransportforskningsInstitutet, VTI medd 909
[32]
Hedberg, G.E., et al. (1991).
Mortality in circulatory diseases, especially ischemic heart
disease, among Swedish professional drivers
. J Human Ergol., Vol 20, p 1-5.
[33]
Hedberg, G.E., et al. (1993).
Risk indicators of ischemic heart disease among male
professional drivers in Sweden.
Scand. J Work Environ Health, Vol 19, p 326-333.
[34]
Hedberg, G.E. & Langendoen, S.M. (1989).
Factors Influencing the turnover of Swedish
Professional Drivers
. Scand. J Soc Med, Vol 17, p 231-237.
[35]
Campbell, K., Gillespie, T., Segel, L. & Schneider, L. (1981)
State of knowledge review:
Relationship of truck ride vibration to highway safety.
Highway Safety Research Institute,
University of Michigan. FHW A report RD-81/083
[36]
Armstrong, B., Cherry, N.M., Cote, R., Lavoie, J., McDonald, A.D., McDonald, J.C.,
Nolin, A.D. & Robert D. (1988).
Fetal death and work in pregnancy
. Br J Ind Med, Vol
45(3), p 148-157.
Page 142
ROADEX III The Northern Periphery Research
[37]
McFarlane, S. & Sweatman, P.F. (2003).
Investigation into the Specification of Heavy
Trucks and Consequent Effect on Truck Dynamics and Drivers: Final Report.
Report
prepared for FORS by Roaduser International Pty Ltd
[38]
De Solminihac, H. E., Echaveguren, T. & Vargas, S. (2007).
Friction Reliability Criteria
Applied to Horizontal Curve Design of Low-Volume Roads
. Transportation Research
Board of the National Academies, Washington D.C. Transportation Research Record,
No 1989, Vol 1, pp 138-147
[39]
Persson, J. & Strandroth, J. (2005).
Halkolyckor med dödlig utgÄng 2000 - 2004 (Fatal
skid accidents)
. Swedish Road Administration, VV publ 2005:83
[40]
AllmĂ€n Teknisk Beskrivning för VĂGkonstruktioner - ATB VĂG.
Swedish Road
Administration, publ 2005:112
[41]
Standard for drift og vedlikehold, HĂ„ndbok 111 (Standard for operations and
maintenance).
(2003). Norwegian Road Administration.
[42]
Glennon, J.C. (2004).
Hydroplaning - The Trouble With Highway Cross Slope
.
Internet
2008-02-24:
http://www.johncglennon.com/papers.cfm?PaperID=8
[43]
Cenek, P., Jamieson, N. & Owen, M. (2003).
Transit New ZealandÂŽs Truck Ride
Improvement Initiative
. REAAA/ARRB International Conference.
[44]
Krafft, M., Kullgren, A., Lie, A. & Tingvall, C. (2006).
The Effectiveness of Electronic
Stability Control (ESC) in Reducing Real Life Crashes and Injuries
. Traffic Injury
Prevention, Vol 7, No 1, pp 34 â 43.
[45]
Bowler, R., Dunne, M. & McCormick, M. (2001).
Survey of Commercial Truck Drivers:
Valuing Their Priorities for Improving New Zealandâs State Highways.
Australasian
Transport Research Forum, Hobart.
[46]
Jonforsen, H. (2007).
Climate change and effects on airports
. Innovative Pavements
Europe, Stockholm
[47]
Bovenzi, M. & Hulshof, C.T.J. (1999).
An updated review of epidemiologic studies on
the relationship between exposure to Whole-Body Vibration and low back pain (1986-
1997).
Int Arch Occup Environ Health (1999) 72:351-365
[48]
SpÄng, K. (1997).
Assessment of Whole-Body Vibration containing single event shocks.
Noise Control Eng. J.,
45
(1), 1997, pp 19-25.
[49]
Saarenketo, T. & Saari, J. (2004).
User perspective to ROADEX II test areaâs road
network service level
. The ROADEX II project.
Internet 2008-02-24:
www.roadex.org
Page 143
ROADEX III The Northern Periphery Research
[50]
Foster, G., Long, A. & McPhee, B. (2001).
Bad Vibrations. A handbook on Whole-Body
Vibration exposure in mining
. The Joint Coal Board Health & Safety Trust
[51]
Hassan, R. & McManus, K. (2001).
Heavy Vehicle Ride and Driver Comfort
. SAE World
Congress, session on Human Factors in Automotive Design. Detroit, Michigan, USA
[52]
Cebon, D. (1999).
Handbook of Vehicle-Road Interaction.
University of Ca mbridge,
England.
[53] Strandberg,
L.
(1974).
The dynamics of heavy vehicle combinations
. Statens VĂ€g- och
Trafikinstitut, Stockholm. Internrapport 172
[54]
UK Design Manual for Roads and Bridges
. Volume 6, Road Geometry.
[55] Mahone,
D.C.
(1975).
An Evaluation of the Effects of Thread Depth, Pavement Texture,
and Water Film Thickness on Skid Number : Speed Gradients
. Virginia Highway and
Transportation Research Council, Charlottesville, USA.
[56] Saarenketo,
T.
(2005).
Monitoring, Communication and Information Systems & Tools for
Focusing Actions - Ideas and Innovations.
The ROADEX II project.
Internet 2008-02-24:
www.roadex.org
[57]
Mechanical vibration - Testing of mobile machinery in order to determine the vibration
emission value.
European Standard, EN 1032 (2003).
[58] Westerling,
R.
(2008).
à tgÀrdbar dödlighet som en indikator i den folkhÀlsopolitiska
uppföljningen. (Avoidable mortality as an indicator in the evaluation of public health
politics
). Statens folkhÀlsoinstitut, Rapport 2008:03
Internet 2008-02-24:
www.fhi.se/shop/material_pdf/Atgardbar_dod_webb_0802.pdf
[59]
Mansfield, N.J. & Newell, G.S. (2004).
Exploratory study of Whole-Body Vibration
âartefactsâ experienced in a wheel loader, mini-excavator, car and office workerâs chair.
39âth United Kingdom Group Meeting on Human Responses to Vibration.
[60] Els,
P.S.
(2005).
The applicability of ride comfort standards to off-road vehicles
. Jour
Terramechanics, Vol
42
, p 47-64.
[61] Marjanen, Y. (2005).
Using ISO 2631-5 as an additional Whole-Body Vibration
evaluation method with ISO 2631-1 to include also transient shocks to the analysis
. 12âth
International Congress on Sound and Vibration, Lisbon, Portugal.
[62]
Alem, N. (2005).
Application of the New ISO 2631-5 to Health Hazard Assessment of
Repeated Shocks in U.S. Army Vehicles.
Industrial Health, Vol
43
, p 403-412.
Page 144
ROADEX III The Northern Periphery Research
[63]
Waters, J.C. & Pidwerbesky, B.D. (2008).
Watercutting â investigating the lifecycle of
watercutter rejuvenation of aggregates
. Land Transport New Zealand Research Report
336
[64]
Ogden, K.W. & Daly, P,S. (2007).
Some Roads are Safer than Others, and Hereâs Why
.
World Road Association, 23âd World Road Congress, Paris.
[65] Hunter,
R.N.
(2000).
Asphalts in road construction
. Thomas Telford Publishing.
[66] Jacobsson,
T.
(2008).
StÄlslagg i asfaltbelÀggning. (Steel slag in asphalt mixtures
).
Swedish National Road and Transport Institute, VTI notat nr 5.
Internet 2008-03-03:
http://www.vti.se/EPiBrowser/Publikationer/N5-2008.pdf
[67] Jacobsson,
T.
(2007).
Polering av asfaltbelÀggning. (Polishing of the coarse aggregate
).
Swedish National Road and Transport Institute, VTI notat nr 25.
Internet 2008-03-03:
http://www.vti.se/EPiBrowser/Publikationer/N25-2007.pdf
[68]
LeRoy, D. (2006).
Smooth operators
. SAE Off-Highway Engineering.
[69] Nilsson,
U.
(2004).
Ride Quality in Ambulances â Modelling and Model Validation
. Royal
Institute of Technology, Stockholm, Sweden. MSc thesis, TRITA-AVE 2004:20, ISSN
1651-7660.
[70]
The Swedish Work Environment Act, chapter 3 paragraph 14
. (1977, 2005).
Internet 2008-03-03:
http://www.av.se/inenglish/lawandjustice/workact/chapter03.aspx
[71]
Förstudie VÀg- och jÀrnvÀgskorsning i Stavreviken, VÀg 331, 681, 684 och 683.
(Feasibilty study of road- and railway crossing in Stavreviken).
(2003). VĂ€gverket Region
Mitt.
Internet 2008-03-03:
http://www.vv.se/filer/5304/stavre_forslh_jan03_liten.pdf
[72]
Dybteanalyse av dödsulykker i vegtraffiken i Region midt i 2006
.
(In-depth analysis of
lethal crashes in the road traffic within the Central Region during 2006).
(2007). Statens
Vegvesen, Norge.
Internet 2008-03:
http://www.vegvesen.no/cs/Satellite?c=Page&cid=1168004567453&pagename=vegvesen%2FPage%2FSV
VsubSideInnholdMal
[73]
Fördubbling av trafikolyckor i à dalen (Doubled number of road traffic crashes in à dalen).
(2008).
Allehanda.
Internet 2008-03-16:
http://allehanda.se/avdelning/kramfors/22811
Page 145
ROADEX III The Northern Periphery Research
Comfort, comfort my people
, says your God.
A voice of one calling in the desert;
-Prepare the way for the Lord, make straight paths for him. Every valley shall be filled in, every
mountain and hill made low. The crooked roads shall become straight, the rough ways smooth.
The path of the righteous is level; O upright One, you make the way of the righteous smooth.
And all mankind will see God's salvation.
[Isaiah 26:7, Isaiah 40:1,3-5, Luke 3:5]
ROADEX III PUBLICATIONS
Developing Drainage Guidelines for Maintenance Contracts
Tyre Pressure Control on Timber Haulage Vehicles
Understanding Low-Volume Pavement Response to Heavy Traffic Loading
Health Issues Raised by Poorly Maintained Road Networks
Road condition management policies for low volume roads â tests and development of proposals
Policies for Forest Roads â Some Proposals
Road Construction in Greenland - The Greenlandic Case