Anda di halaman 1dari 29

Journal of Transportation Safety & Security

ISSN: 1943-9962 (Print) 1943-9970 (Online) Journal homepage: http://www.tandfonline.com/loi/utss20

Vertical curve design insights of road tunnels


versus highways

Shy Bassan

To cite this article: Shy Bassan (2016): Vertical curve design insights of road tunnels versus
highways, Journal of Transportation Safety & Security, DOI: 10.1080/19439962.2016.1206049

To link to this article: http://dx.doi.org/10.1080/19439962.2016.1206049

Accepted author version posted online: 13


Jul 2016.
Published online: 13 Jul 2016.

Submit your article to this journal

Article views: 16

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at


http://www.tandfonline.com/action/journalInformation?journalCode=utss20

Download by: [Ryerson University Library] Date: 08 October 2016, At: 04:37
JOURNAL OF TRANSPORTATION SAFETY & SECURITY
http://dx.doi.org/10.1080/19439962.2016.1206049

Vertical curve design insights of road tunnels versus


highways
Shy Bassan
Amy Metom Engineers & Consultants, Ltd., Tel Aviv, Israel

ABSTRACT KEYWORDS
The design of road tunnels is an essential infrastructure crest; sag; road tunnel; sight
component in the highway system. The study implements distance; trucks; vertical
reasonable criteria for obtaining vertical curve length based on curve
sight distance requirements of road tunnels versus open
roadways. The infrastructure impact of sight distance tunnel
characteristics on the design of crest and sag vertical curve radii
is based on safety and driving comfort criteria. Both results of
crest and sag vertical curves show signicant reductions of
vertical radii in road tunnels compared to open roadways, that
is, percentage reduction range of 35% to 71% depending on the
design speed, curve type, and tunnel pavement characteristics.
The driving comfort criterion generally overrides in the lower
range of design speeds whereas the safety criterion overrides in
the upper range of design speeds: 100 to 120 km/h for moist
and dry tunnel sag vertical curve, 90 to 120 km/h for dry tunnel
crest vertical curve, and 80 to 120 km/h for moist tunnel crest
vertical curve.
This impact of a reduced vertical curve length triggers the
possibility of shorter tunnel construction length, lower tunnel
construction cost and maintenance cost including components
of complementary systems, and exibility in tunnel construction
due to geotechnical, groundwater, and other infrastructure
constraints. Still, the selected tunnel alignment, even if it is
shorter than other alternatives, requires eld tests such as
geological investigation, vibrations and building inspection, and
hydrologic balance.

1. Background: Design and infrastructure components of tunnels versus


open highways
The design of road tunnels is an essential component in the transportation net-
work and specically in the highway system. The need for roadways construction
along difcult topography including overcoming natural conditions is the major
motivation for selecting the road tunnel alternative solution. Road tunnels

CONTACT Shy Bassan bassans@netvision.net.il Amy Metom Engineers & Consultants, Ltd., 55A Yigal Alon
St., Tel Aviv 67891, Israel.
Color versions of one or more of the gures in the article can be found online at www.tandfonline.com/utss.
2016 Taylor & Francis Group, LLC and The University of Tennessee
2 S. BASSAN

solution minimizes the damage to the environment and land, preserves land
resources, and reduces trafc congestion and air pollution. Generally, the design of
road tunnels should be based on the geometric design principles of open highways.
The differences between tunnels and open roads are caused by (1) construction
cost considerations, (2) lighting, (3) structural requirements, (4) cross-section
implications, (5) friction coefcients and driver perception reaction time adjusted
to tunnel surface properties, (6) the impact of ventilation design on the longitudi-
nal gradient, and (7) the need to locate complementary elements inside the tunnel
envelope in addition to the trafc envelope, transport of dangerous goods, and
signs installations (for trafc and re safety guidance). These issues are based on
information acknowledged by road tunnel design guidelines and highway geomet-
ric design guidelines from several countries (2010; American Association of State
Highway and Transportation Ofcials [AASHTO], 2011; Austroads, 2009; Design
Manual for Roads and Bridges [DMRB], 1999; Federal Highway Administration
[FHWA], 2009; Norwegian Public Road Administration, 2004; PIARC, 2001,
2004, 2003, 2008 Road and Transportation Research Association [RAA], 2006),
and practical experience of recent road tunnel projects constructed in Israel.
The main differences inuencing the design and infrastructure components of
tunnels versus open roadways are documented as follows:
1. Lighting: Tunnels have permanent lighting for 24 h except the entry zone.
The lighting plan during daylight is different than during night hours. The
drivers entering the tunnel, immediately after daylight, have a short time to
adapt their eyes to the relatively dark surrounding inside the tunnel because
the distance traveled during this adaptation process is relative to the travel
speed. The slow adaptation of eyes from daylight to the tunnel dim environ-
ment necessitates gradual reduction of tunnel lighting in the threshold and
transition zones of the tunnel (Figure 1; DMRB, 1999). The conventional sag
vertical curve safety criterion of night driving and headlight beam is not valid
in road tunnels.
2. The design of road tunnels requires components of complementary systems
(re safety, re detection, ventilation, communication systems) that are not
critical or do not exist in open roadways. These components are crucial for
tunnels design. The design of these components depends on the tunnel cross
section dimensions, the tunnel length, and so on.
3. The accessibility of rescue vehicles, ambulances, and heavy vehicles due to
road accidents has to be taken into consideration in the geometric design
process of road tunnels.
4. The driver perception reaction characteristics are different in road tunnels.
On one hand, the driver nds it hard to get used to the restricted environ-
ment of the tunnel. He feels bunged and is unable to connect the natural
environment in the open area. Nonetheless, the tunnels exhibit a better acci-
dent record (Austroads, 2010; Lemke, 2000) than open roadways because
drivers (especially commuters or regular drivers) become more alert in the
JOURNAL OF TRANSPORTATION SAFETY & SECURITY 3

Figure 1. Luminance change curve example (data from DMRB, 1999).

changed environment of the tunnel; specically, the absence of roadside


obstacles, narrow shoulders, the standards of construction, and the safety
features (trafc control and re safety).
5. The tunnel walls and the bounded cross section are physical obstacles, which
should be considered during the design process. Heavy good vehicles (HGV)
might be restricted from passing through the tunnel section. These vehicles
have a potential inability to perform a U-turn maneuver.
6. Intersections and branch connections (forks) are not advisable for tunnels
design. These geometric design elements increase signicantly the construc-
tion cost and may also confuse the drivers along the bunged environment of
the tunnel.
7. The construction cost of road tunnels is signicantly higher than open high-
ways due to the possible use of boring machines, the amount of concrete,
and the complementary systems.

1.1. Highway design perspective and article objective


Horizontal and vertical curves may be necessary to align the tunnel with its
approach roadway and to avoid obstacles on the ground. The same considerations
and geometric design elements apply in determining the horizontal and vertical
curve radii of road tunnels as in surface roadways: design speed, equivalent decel-
eration or friction factor, driver perception reaction time, centrifugal force, super-
elevation, sight distance and line of sight.
The current study implements reasonable criteria for obtaining the sight distance
of road tunnels versus open roadways to evaluate the design of tunnels vertical
curves. These criteria are based on interrelated stopping sight distance components:
perception reaction time (PRT) and longitudinal friction coefcient (fT).

1.2. General aspects of driving behavior in road tunnels


Driving along the tunnels dark narrow environment may cause anxiety, uncer-
tainty, and even fear of hitting other vehicle or tunnel walls and of other dangerous
4 S. BASSAN

circumstances such as re or a tunnel collapse (Caliendo, De Guglielmo, & Guida,


2013; PIARC, 2008). Drivers in tunnels generally reduce their speed and increase
their lateral distance to the tunnel wall (Caliendo et al., 2013) that can be inter-
preted as an increased alertness while driving in road tunnels. This change in driv-
ing behavior generally occurs while approaching the tunnel portals. Amundsen
and Ranes (2000) believe that driving in road tunnels causes an uneasy feeling
resulting from the darkness and safety concerns.
Therefore driving in tunnels requires supplemental attention and mental work-
load (PIARC, 2008) that cause the drivers to increase their vigilance.
According to Lemke (2000) and Yeung and Wong (2013) when the tunnel part
in the roadway network is small drivers, in general, tend to drive more carefully in
tunnels and at a lower speed.

2. Stopping sight distance components in road tunnel versus open


roadways
Stopping sight distance (SSD) is the distance that the driver must be able to see
ahead along the roadway while traveling at or near the design speed and safely
stop before reaching a stationary object. SSD can be limited by vertical and
horizontal curves. The fact that it affects the design radius of both curves makes
SSD so fundamental in the geometric design process.
The stopping sight distance has two components: (1) the distance traveled
during the driver PRT and (2) the distance traveled during braking.
The SSD can be determined by using the following formula:

PRT Vd 2
SSD D  Vd C (1)
3:6 2  3:62  d

where SSD D minimum stopping sight distance (m), Vd D design speed (km/hr),
d D deceleration of passenger car (m/s2), equivalent to the longitudinal friction
coefcient (f) multiplied by the acceleration of gravity (g), d D fTg, and PRT D driver
perception reaction time (s).
The formula assumes level terrain. Ascending grade decreases the SSD and
descending grade increases the SSD.
The two sensitive parameters in the SSD formula that are prospective to be dif-
ferent in road tunnels versus open roadways (as described in the background sec-
tion) are the PRT and the coefcient of longitudinal friction (fT).

2.1. Driver perception reaction time


Driving in tunnels requires more attention and mental workload than driving on
open roads. However, the closed environment that possibly causes anxiety and
psychological stress to a certain amount of drivers might compel the typical driver
JOURNAL OF TRANSPORTATION SAFETY & SECURITY 5

to be more alert and cautious by reducing the driving speed and keeping a lateral
distance from the tunnel walls.
Based on extensive literature review of PRT results in open roadways, surprised
drivers will start braking after 1.5 sec on average whereas a completely expected
event requires only 0.75 sec (Green, 2000). If the expectancy is considered normal,
that is, the need to stop the vehicle or slowing down due to vehicle brake lights, the
expected PRT is 1.25 sec. Figure 2 (Summala, 2000) shows the PRT normal distri-
bution of young drivers (age 1840 years) and older drivers (age 5084 years) in a
surprise condition. Drivers on their rst time came upon an obstacle in their path
after a hill crest. The distribution shows approximately maximum PRT of 1.75 s
for older drivers and 1.45 s for young drivers.
Durth and Bernhard (2000) examined braking performance at a crest vertical
curve. Drivers were not informed about the required braking maneuver, that is,
surprise condition. The road section was closed for trafc during the study experi-
ment. The results of PRTs as a function of the initial speeds are presented in
Figure 3. The lower values of PRT were 0.6 sec. The trend of observed data points
in Figure 3 shows a reduction in PRT as the initial speed grows from 75 km/h to
105 km/h. The upper bound graph (95th percentile) of observed data points indi-
cates PRT of 1.85 sec for 75 km/h and PRT of 0.8 sec for 105 km/h. This implies
the strain of drivers as the initial speed increases. Durth and Bernhard (2000) pro-
poses 2.0 sec as a design PRT value for the analytical calculation of stopping sigh
distance. Because drivers were still in an outstanding watchfulness during the
experiment (drivers recognized that they participated in an experimental driving

Figure 2. Perception time, response time and PRT in the surprise condition (Olson & Sivak, 1986;
redrawn by Summala, 2000, 2002).
6 S. BASSAN

Figure 3. relationship between perception brake reaction times and initial speeds (data from Durth
& Bernhard, 2000).

situation) the PRT recommended value by Durth and Bernhard (2000) can reect
the capabilities of most drivers in open roadways and essentially in road tunnels.
Fambro, Fitzpatrick, and Koppa (1997) documented results of several PRT stud-
ies. A brief summary of these results is presented in Table 1. The upper part of the
table refers to surprise conditions, and the bottom part refers to alerted driver condi-
tions. The covert category under surprise condition indicates that the drivers
stopped at trafc signals without knowing that these trafc signals were test subjects.
The perception brake reaction times (PRTs) for unalerted drivers (by passing
unexpected signals or objects) were longer than PRTs of alerted drivers. The ratio
between average PRT of surprise condition and anticipated condition is 1.75 ( D
1.28/0.73).
Dozza (2013) indicated that response times for speeds of 25 to 45 mph
(4075 kph) were signicantly lower than for speeds that are lower than 25 mph
(40 kph). The median PRTs range between 1.0 sec to 1.70 sec and the 75th percen-
tile PRTs are approximately 2.0 sec. Drivers reduce their response time during

Table 1. Summary of surprise and alerted perception brake reaction time studies (Fambro et al.,
1997).
N (observations) Age Mean PRT (s) SD (s)

Surprise (unsuspecting driver)


Covert (unexpected signal): Sivak, Olson, & Farmer (1982),
Wortman & Matthias (1983), Chang, Messer, &
Santiago (1985)
Mean estimate 3062 Mix 1.27 0.66
Surprise (unexpected object): Olson and Sivak (1986),
Lerner, Huey, McGee, & Sullivan (1995)
Mean estimate 120 Young, old, mix 1.28 0.20
Anticipated (alerted driver): Olson & Sivak (1986),
Johansson & Rumar (1971)
Mean estimate 385 Young, old, mix 0.73 0.16

PRT D perception reaction time.


JOURNAL OF TRANSPORTATION SAFETY & SECURITY 7

Figure 4. Response times outcome for night driving with lighting and certain ranges of speeds
(data from Dozza, 2013).

night with lighting conditions: median percentile of 1.0 sec and 75th percentile of
1.5 sec (Dozza, 2013). These results are presented in Figure 4. Night conditions
(with lighting) are analogical to tunnel driving conditions (except the closed envi-
ronment) compared to daylight.
Arndt, Cox, and Lennie (2011) proposed that the PRT that represents typical
drivers that travel at the 85th percentile speed around barriers on interchange
ramps should be 1.5 sec for passenger cars and trucks. For typical freeway sections,
PRT of 2.0 sec could be used in geometric design for determining the SSD. Longer
unidirectional (freeway and multilane) tunnels with higher design speed could
match driving PRT of 2.0 sec assuming a lower drivers alertness and a diminished
concentration due to relatively monotonous al driving (Caliendo, Guglielmo, &
Guida, 2013) in spite of the tunnel closed environment. In shorter tunnels with
reduced driving speed (Calvi, De Blasiis, & Guattari, 2012) driving vigilance may
be more robust without the effect of dull driving. Calvi et al. (2012) showed that
drivers reduced their speed signicantly and increased their lateral position from
the right tunnel wall while driving inside the road tunnel. Practically, and based on
the PRT overview, drivers are capable of a shorter PRT than the acceptable value
of 2.5 sec for determining the SSD in highway design. The international road tun-
nel design guidelines of Australia (Austroads, 2010), United Kingdom (DMRB,
1999), and PIARC (2004, 2001) recommend a PRT of 2 sec for calculating SSD.
The Australian tunnel guidelines recommend a lower PRT value (1.5 sec) if the
design speed is 90 km/h or lower.

2.2. Dry longitudinal friction coefcient relevance and its implementation for
road tunnels
The implementation of reduced longitudinal friction coefcients for road tunnels,
that is, dry friction coefcients is practical if we assume that the inner tunnel zone is
8 S. BASSAN

practically dry during the whole year and the asphalt concrete quality is highly supe-
rior inside the tunnel compared to the open roadway. The reasons for better quality
are periodical maintenance activities such as pavement cleaning and washing and
pavement stratication and scrubbing control. These activities reduce the skidding
component of pavement, presuming that the solar radiation effect is avoided and the
asphalt concrete mixture is of high quality such as stone mastic asphalt (SMA).
Several examples of dry friction coefcient are presented in Table 2. Table 2
shows that most international implemented dry friction coefcient values are fairly
consistent and vary between 0.60 and 0.70.
Amundsen (2009) showed that most tunnel accidents occur in dry surface condi-
tion. Therefore, the assumption of using dry tunnel friction coefcients could be prac-
tical and reasonable for SSD considerations and highway design in terms of road
safety. His correlation between tunnel surface conditions and road accidents inside
the tunnel showed that two thirds of the accidents occur anyway in dry surface condi-
tions and only 2.3% occur in slippery conditions other than wet, bared, or ice-covered
pavement conditions. Other slippery conditions could be based on an unclean surface
due to oil, fuel, and other ammable and toxic liquids of dangerous goods transport.
PIARC (2003) provides dry surface longitudinal friction coefcients that could
be applicable to either dry weather or road tunnels (Table 3). These results are
based on a sideway-force coefcient routine investigation machine (SCRIM)
method for tire-pavement friction measurements (W.D.M. Limited, Transport
Research Laboratory, UK; http://www.wdm.co.uk/history). This test is conducted
with a testing tire moving at a constant speed over the road surface often at speeds
that do not impede regular trafc.

2.3. Typical measurement of tunnel pavement wet friction coefcient in Israel


Most work on skid resistance relates to wet pavement friction and not to dry pave-
ment friction (locked wheel). Recent measurements of tunnel friction coefcient

Table 2. Typical examples of longitudinal dry friction coefcients.


Source Dry Friction Coefcient Comment

Japan (Thanesuen, Kagaya, & Uchida, 2005) 0.70


Australia (Austroads, 2009) 0.61
Kosasih, Robinson, & Snell (1987), Average dry pavement curve: 0.560.68 Figure 5
AASHTO (USA; 1984)
Public roads (USA; Chowdhury, 0.65 (dry friction for modern cars before skidding)
Warren, D. L., & Howard, 1991)
NCHRP 400 (USA; Fambro et al., 1997) ABS, Vd D 65 kph: f D 0.62 (176 observations)
No ABS, Vd D 80 kph, f D 0.65 (216 observations)
ABS, Vd D 80 kph, f D 0.71 (203 observations)
Average dry friction coefcient: 0.662
Oregon DOT (USA; Layton, 1997, 0.6
Layton & Dixon, 2012)
Denmark (Greibe, 2008) Deceleration: 8.79.5 m/s 2
Figure 5
Equivalent dry friction: 0.880.968
PIARC (2003) 0.550.70 (Speed > 50 km/h) Table 3

ABS D anti-lock braking system.


JOURNAL OF TRANSPORTATION SAFETY & SECURITY 9

Table 3. Coefcient of friction of asphalt concrete and Portland cement in dry surface condition
(PIARC, 2003).
Dry

Less than 50 km/h More than 50 km/h

Road surface From To From To

Portland cement
New, sharp 0.80 1.20 0.70 1.00
Traveled 0.60 0.80 0.60 0.75
Trafc polished 0.55 0.75 0.50 0.65
Asphalt or tar
New, sharp 0.80 1.20 0.65 1.00
Traveled 0.60 0.80 0.55 0.70
Trafc polished 0.55 0.75 0.45 0.65
Excess tar 0.50 0.60 0.35 0.60

were performed in Israel on Highway 60 (Gillo Tunnel; Figure 6). Gillo Tunnel is a
bidirectional tunnel, located in southern Jerusalem where the weather can be rainy
during the winter season.
The measurement equipment (DYNATEST 6875H) locks the wheel until 14%
sliding while a control ow of water is sprayed for imaging wet pavement. The
measurement wheel is installed as a fth wheel in the experiment dedicated vehicle.
As the vehicle moves forward, the force generated by resistance to sliding is con-
verted to the wet pavement friction (skid resistance). The pavement condition
index (PCI) along the tunnel segment was considered very good (PCI D 9596).
The tunnel length is 0.9 km (km post: 74.074.9). Measurements were per-
formed also along 0.9 km prior to the tunnel entrance for comparison purposes.
Complete results are introduced in the Appendix. The average travel speed along
the road segment is 65 to 75 km/h. Table 4 presents a summary of the friction coef-
cient measurement results. The Gillo Tunnels average friction result is 0.56 and
the open roadway (preceding the tunnel) average friction result is 0.41.
The results emphasize the impact of the tunnel superior pavement condition
(no solar radiation effect and no exposure to rain during winter) even though they
refer to wet pavement condition and not to dry pavement condition.

3. Stopping sight distance for road tunnels: Recommended concepts and


evaluation
The assessment of stopping sight distance for road tunnel is performed by the fol-
lowing assumptions based on the research overview presented.

Table 4. Roadway 60 (Israel): Gillo tunnel (Israel) and open roadway preceding: summary of longitu-
dinal friction measurements (2009, 2010).
Average Friction 95th Percentile
Coefcient SD Condence Interval

Gillo tunnel (km post 74.074.9) 0.56 0.089 0.0309


Open roadway (km post 73.274.0) 0.41 0.10 0.0358
10 S. BASSAN

3.1. PRT
The perception reaction time (PRT) is 1.5 s for the design speed range of 50 to
80 km/h and 2.0 s for the design speed range of 90 to 120 km/h that possibly
matches longer tunnels (freeway or divided multilane tunnels). The reason for these
reduced values compared to open roadways (2.5 sec) is the consistent PRT results
observed in several studies presented in subsection 3.1 and drivers awareness and
vigilance along the bounded cross-section of road tunnels with narrow shoulders.

3.2. Friction coefcient


The friction coefcient values are based on two possibilities of tunnel surface
situations: dry tunnel and moist tunnel.

3.3. Situation 1: Dry tunnel


The tunnel walls and ceiling are watertight or usually there is no rain where the
tunnel is designed. The range of friction coefcient values for dry tunnels is
adopted by Table 3 (PIARC, 2003); between 0.55 for the design speed of 120 km/h
and 0.70 for the design speed of 60 km/h or less.

3.4. Situation 2: Moist tunnel


Tunnels that are not fully watertight are considered moist tunnels. This situation is
an intermediary situation between wet and dry road surface. The design values of

Figure 5. Dry pavement ction and deceleration rates as a function of vehicle speed (data from
Greibe, 2008; Kosasih, Robinson, & Snell, 1987).
JOURNAL OF TRANSPORTATION SAFETY & SECURITY 11

Figure 6. Gillo tunnel, Highway 60, Israel.

friction coefcient are assumed to be the average values between wet asphalt con-
crete surface (generally used for open roadways) and dry asphalt concrete surface.
These values ranged between 0.464 for the design speed of 120 km/h and 0.569 for
the design speed of 60 km/h, are similar to the wet pavement friction coefcient
values measured in Gillo Tunnel in Israel (0.56 0.03).
These measurement results as well the dry surface longitudinal friction coefcients
results documented in PIARC (2003) support the equivalent friction coefcients rec-
ommendations in this study for moist tunnel and dry tunnel pavement conditions.
The tunnel friction coefcients are suitable for powerful but comfortable braking
(not emergency braking) and periodical maintenance of SMA concrete material.
The proposed tunnel pavement conditions are coherent with the regular climate
conditions in Israel and in other Mediterranean countries (i.e., Greece, Cyprus, etc.).

3.5. EOT zone


The end of tunnel (EOT) zone is assumed to be sited along 150 m from the tunnel
entrance or exit (tunnel portal). The desirable SSD for this zone shall be based on
wet asphalt concrete surface friction coefcients (as used for open roadways). The
perception reaction time values of this zone are the adopted tunnel PRT values
because these zones are still located in a tunnel environment. Short tunnel lengths
of 300 m or less should match the surface characteristics of the EOT zone. The rec-
ommendation of desirable wet friction coefcients (like open roadways) along the
EOT zone can be supported by a research conducted by Amundsen and Ranes
(2000). This research indicated that relatively high percentage of accidents occurred
at the entrance of the tunnel, that is, 50 m prior to the tunnel portal (access zone)
and 50 m past the tunnel portal (EOT zone). Also short tunnel lengths (which are
expected to have in general a wet surface during rainy weather) were found to entail
higher accident rates than long tunnels (Amunsden & Ranes, 2000; Lemke, 2000).
The design values of friction and deceleration of dry, moist, and wet (EOT case)
road tunnels as a function of the design speed are introduced in Table 5. The SSD
design values for the three tunnel surface situations are presented in Table 6 and
12 S. BASSAN

Figure 7. Also included in Table 6 are SSD values of open roadways. Table 6 and
Figure 7 depict SSD values for trucks drivers in tunnels and in open roadways too.
These design values are essential for vertical curves analysis in the following
section.
The deceleration and friction coefcients of wet road tunnels (i.e., open road-
ways) are based on German guidelines for the design of motorways (AASHTO
(2011); Bassan (2012); RAA, 2008). These wet friction coefcients for open road-
ways (ft D 0.340.43) are lower than the measurement results of wet friction coef-
cients in Gillo Tunnel in Israel (ft D 0.56 0.03).
The SSD values for trucks in tunnels are similar to open roadways (Bassan, 2012)
by assuming reduced PRTs that are implemented for road tunnels and EOT zone
and by considering lower equivalent friction coefcient for trucks (0.29 for 50  Vd
 90 km/h, 0.28 for Vd D 100 km/h, 0.26 for Vd D 110, and 0.25 for Vd D 120).
Typical or experienced truck drivers would not exceed the speed limit of 100 km/h.
The dry and moist tunnel friction effects are counterbalanced by the lower level
functional deceleration characteristics of trucks versus passenger cars.
The differences of the SSD values between open roadways and dry and moist tun-
nels are signicant. The ranges of percentage reduction compared to open roadways
are: 28% to 42% for dry tunnels, 19% to 33% for moist tunnels, and 6.5% to 25% for
EOT zone, depending on the design speed. As the design speed increases the SSD per-
centage reduction decreases, however, the SSD absolute difference generally increases,
apart from EOT zone from the design speed of 90 km/h. Also, in dry and moist tun-
nels the absolute difference slightly decreases from Vd D 80 to Vd D 90 km/h. The
reason is the adjustment of driver PRT (from 1.5 sec to 2.0 sec) in this transition.

3.6. The use of reliability analysis for stopping sight distance parameters:
Authors perspective
Probabilistic design approaches have been applied in several studies related to
highway design and highway safety by using reliability concepts (Navin, 1990,

Table 5. Tunnel design values of equivalent deceleration and longitudinal friction, and perception
reaction time (PRT).
Design Speed (km/hr)

50 60 70 80 90 100 110 120

Tunnels PRT (sec) 1.5 1.5 1.5 1.5 2.0 2.0 2.0 2.0
Dry tunnel fT 0.7 0.7 0.675 0.650 0.625 0.600 0.575 0.55
d 6.867 6.867 6.622 6.377 6.131 5.886 5.641 5.396
Moist tunnel fT 0.569 0.569 0.552 0.534 0.516 0.499 0.481 0.464
d 5.584 5.584 5.411 5.238 5.066 4.893 4.720 4.548
End of tunnel (wet) fT 0.438 0.438 0.428 0.418 0.408 0.398 0.387 0.377
d 4.3 4.2 4.1 4.0 3.9 3.8 3.7 3.7
Trucks fT 0.29 0.29 0.29 0.29 0.29 0.28 0.26 0.25
d 2.845 2.845 2.845 2.845 2.845 2.747 2.551 2.453

fT D longitudinal friction coefcient; d D equivalent deceleration.


JOURNAL OF TRANSPORTATION SAFETY & SECURITY 13

Table 6. Tunnel stopping sight distance (SSD), m.


Design Speed (km/hr)

SSD (m) 50 60 70 80 90 100 110 120

Dry tunnel 35 50 60 75 105 125 145 170


% reduction from open road 41.7 33.3 36.8 37.5 27.6 26.5 27.5 27.7
(absolute reduction)a (25) (25) (35) (45) (40) (45) (55) (65)
Moist tunnel 40 50 65 85 115 135 165 190
% reduction from open road 33.3 33.3 1.6% 9.2% 20.7 20.6 7.5% 19.1
(absolute reduction) (20) (25) (30) (35) (30) (35) (35) (45)
End of tunnel (EOT) 45 60 75 95 130 155 185 220
% reduction from open road 25.0 20.0 21.1 20.8 10.3 8.8 7.5 6.4
(absolute reduction) (15) (15) (20) (25) (15) (15) (15) (15)
Open roadway 60 75 95 120 145 170 200 235
Trucks: tunnels, EOT 55 75 100 125 160 200 245 295
Trucks: open roadway 70 95 120 145 175 210 260 310
a
The numbers in parentheses signify absolute reduction of SSD design values from open roadways.

1992). The reliability analysis approach was also implemented in other studies such
as Ismail and Sayed (2012) and Ibrahim, Sayed, and Ismail (2012) to select a suit-
able combination of cross-section elements with restricted sight distance to result
in reduced collisions and acceptable risk levels.
Hussein and Sayed (2014) applied reliability analysis (probabilistic approach) to
take into consideration uncertainty associated with geometric design parameters.
They calibrated design charts for Horizontal Sightline Offset (HSO) at different
probabilities of noncompliance (Pnc) levels. The HSO is the distance between the
edge of median barriers and the centerline of the left trafc lane.

Figure 7. Stopping sight distance (SSD) for tunnels (dry, moist, end of tunnel) and for open
roadways.
14 S. BASSAN

In highway geometric design the use of Pnc characterizes the probability that the
design does not meet the standard design requirement. Hussein and Sayed (2014)
used a limit state function that is the difference between available stopping sight dis-
tance (ASD) and the demanded SSD. The Pnc was generated from this function.
The preferred target Pnc is the road designers choice based on the design policy.
SSD is presented in Equation 1. ASD is the horizontal sight distance formula
assuming SSD is shorter than the horizontal curve length (L). Its approximate form is:

p
ASD D 8  HSO  RHSD for ASD < L (2)

HSO D free horizontal sightline offset between the middle of sightline and the
centerline of the inside lane (m).
RHSD D the horizontal curve radius along the centerline of inside lane (m).
ASD D Available stopping sight distance along the sightline (m).
The calibration of horizontal radius and HSO for different Pncs showed that
the conventional design guides (AASHTO, 2011) are conservative especially at
sharp radii and high design speed (Hussein & Sayed, 2014). It appears that such
results are somehow subjective because increasing the risk (i.e., increasing the tar-
get Pnc) can be performed by deterministic modications such as reducing the
PRT or increasing the friction coefcient (Equation 1) based on the prevailing traf-
c and pavement conditions and specic driving behavior characteristics. The
PRT and deceleration rate design parameters used in Hussein and Sayed study
(2014) were anyway less conservative (mean PRT of 1.5 sec and SD of 0.4 sec and
mean deceleration rate of 4.2 m/sec2 and SD of 0.6 m/sec2) than the conventional
design parameters. This means that they include a certain deterministic element
behind their selection as an input for the reliability analysis. The typical values
used for highway design (based on AASHTO, 2011) are PRT D 2.5 sec and deceler-
ation rate D 3.4 m/sec2. Such values are considered more conservative, and because
they presume larger variability, they might justify implementation of reliability
analysis.
Overall, these parameters (that determine the demanded SSD), implemented for
road tunnels in the current study, are conceptually not random variables but have
certain causal aspect behind their selection due to physical, climate, and driving
behavior considerations that are crucial for geometric design analysis. These con-
siderations reduce the variability and the uncertainty component of the demanded
SSD design parameters and therefore rebut the necessity of probability techniques.
Instead of calibrating a target Pnc for each project (complex procedure), the high-
way engineer can directly select and rene the design parameters based on certain
characteristics related to driver behavior, driving environment, pavement condi-
tions, and trafc, especially in road tunnels design. The proposed procedure is
more practical in the highway design process.
JOURNAL OF TRANSPORTATION SAFETY & SECURITY 15

4. Crest vertical curve insights for road tunnels


4.1. Safety criterion
The radius of tunnel crest vertical curve depends on the SSD, driver eye height,
object height, and the algebraic difference in grade (Equations 3 and 4). Equations 3
and 4 that dene the safety criterion of crest vertical curves are adopted from open
roadways design characteristics (AASHTO, 2011).
The equations are different if SSD is smaller or equal to the curve length (L), or
if SSD is larger than the curve length (L):

SSD2
Rv D p p2 for SSD  L (3)
2  h1 C h2
p p2
200  SSD 20; 000  h1 C h2
Rv D for SSD > L (4)
A A2

where Rv D radius of crest vertical curve (m), A D algebraic difference in grade


(percents), SSD D stopping sight distance for design (m), h1 D design value of
driver eye height (1.05 m for passenger cars), h2 D design value of the object
height: 0.6 m for road tunnels. For EOT zone, this design value is assumed to be
0.15 m for 60  Vd  90 km/h (two-lane road tunnel or one-way tunnel ramp)
and 0.6 m for 90  Vd  120 km/h (multilane or freeway tunnel).
The principal is that in higher design speeds drivers might not be able to see an
object shorter than 0.6 m (AASHTO, 2011). In road tunnels, the most common
lowest obstacle is the vehicle tail lights. Most unidirectional road tunnels are
designed in Israel for the higher range of design speeds (90  Vd  120), that is,
multilane highways or freeways.

4.2. Driving comfort criterion


The accepted centrifugal acceleration in vertical curves is 0.5 m/sec2 (Lamm, Psar-
ianos, & Mailaender, 1999) to meet the comfort criterion.
The mechanical expression of this control is:

Vd 2
Rv D D 0:2572Vd 2 (5)
0:53:62
where Rv D minimum vertical radius based on driving comfort criterion (m) and
Vd D design speed (km/h).

4.3. Crest vertical curve analysis for road tunnels


The minimum vertical radius is the larger computed value based on safety or com-
fort criterion. Table 7 summarizes the results for dry tunnel, moist tunnel, EOT
zone, and open roadway. The gray shaded cells are values that were determined by
16 S. BASSAN

Table 7. Crest vertical curve minimum radius for road tunnels (moist, dry, end of tunnel [EOT]) ver-
sus open roadway.
Design Speed (km/h)

Site Characteristics 50 60 70 80 90 100 110 120

Dry tunnel 390 560 760 990 1710 2420 3250 4470
% reduction from open road 57.1 60.6 66.5 55.6 47.4 45.9 47.4 47.6
(absolute reduction)b (520) (860) (1510) (1240) (1540) (2050) (2930) (4060)
Moist tunnel 390 560 760 1120 2050 2820 4210 5580
% reduction from open road 57.1 60.6 66.5 49.8 36.9 36.9 31.9 34.6
(absolute reduction)b (520) (860) (1510) (1110) (1200) (1650) (1970) (2950)
EOT (h2 D 0.15) 510 910 1420 2270 4240
EOT (h2 D 0.6) 1400a 2620 3720 5290 7480
% reduction from open road 44.0 35.9 37.4 37.2 19.4 16.8 14.4 12.3
(absolute reduction) (400) (510) (850) (830) (630) (750) (890) (1050)
Open roadway (h2 D 0.15) 910 1420 2270 3620 5280
Open roadway (h2 D 0.6) 2230a 3250 4470 6180 8530
a
Supplemental value for h2 D 0.60 (Vd D 80 km/h).
b
The comparison is valid for open roadway: Vd D 5070 kph, h2 D 0.15 m, and Vd D 80120 kph, h2 D 0.60 m.

safety criterion. The other cells results were generated by comfort criterion. Figure 8
presents the results of tunnel crest vertical radii versus open roadways graphically.
The ranges of percentage reduction compared to open roadways are: 48% to 67%
for dry tunnels, 35% to 67% for moist tunnels, and 12% to 44% for EOT zone, depend-
ing on the design speed. These differences are substantial and affect the tunnel length
where crest vertical curves are needed for the design of tunnel vertical alignment.
The relationship between the curve length and curve radius is as follows:

Rv A
LD (6)
100

The resulting lengths of crest vertical curves (L) as a function of the algebraic
difference, for dry tunnels and moist tunnels based on safety and comfort criteria,
are presented in Figures 9a, b correspondingly. Each line in the graph depicts a
specic design speed between 50 and 120 km/h. The radii are bounded to prevent
too short vertical curve length due to esthetical reasons. The lower bounds of verti-
cal curve length are determined by Equation 7. Also included in Figures 9a, b is the
boundary line, SSD D L.

L.min; esthetics criterion/ D 1Vd (7)

where L D vertical curve length (m) and Vd D design speed (km/h).


Figure 10 presents a comparison of typical design values of crest vertical curve
lengths between dry tunnel, moist tunnel, and open roadway. These typical vertical
curves design values are presented for the design speeds of 80, 100, and 110 km/h
for comparison purpose. The dashed lines signify open roadways and the solid
lines signify road tunnels (dry and moist). The dashed line (open roadway) for
Vd D 80 km/h signies vertical radius computed for h2 D .15 m.
JOURNAL OF TRANSPORTATION SAFETY & SECURITY 17

Figure 8. Graphical presentation of crest vertical curve minimum radius for road tunnels.

5. Sag vertical curve insights for road tunnels


5.1. Safety criterion
Tunnels have permanent lighting for 24 h except the entry zone. Therefore, safety
criterion of sag vertical curve for night driving in open roadways is not applicable
to road tunnels. The safety limitation for sag vertical curve in road tunnels is based
on the existence of overhead obstacle inside the road tunnel (e.g., overhead sign, ven-
tilators, etc.) or the vertical clearance in the entrance to the road tunnel after the por-
tal. This limitation is applicable to trucks (or high vehicles) when a certain
percentage of trucks is expected to use the tunnel (e.g., more than 5%). The formulas
for calculating the sag vertical radius in such case are presented in Equations 8 and 9
(derived from Drew, 1968 and PIARC, 2003).

SD 2
Rv D   for SSD  L (8)
8 H h1 C2 h2
 
200  SD 80000 h1 C h2
Rv D H for SSD > L (9)
A A2 2

where Rv D radius of sag vertical curve in road tunnels (m), A D algebraic difference
in grade (percents), SSD D Trucks stopping sight distance for design (m),
h1 D design value of driver eye height for truck (2.40 m), and h2 D design value of
object height: 0.6 m for road tunnels. For simplicity object height of 0.60 m is
18 S. BASSAN

Figure 9. (a) Design values for dry tunnels crest vertical curve length (m) based on safety, comfort
and esthetics criteria. (b) Design values for moist tunnels crest vertical curve length (m) based on
safety, comfort and esthetics criteria.

assumed for EOT zone for sag vertical curve analysis. The chance of truck driver (eye
height of 2.40 m) to see a lower object (height of 0.15 m) is negligible.
H D vertical clearance between pavement surface and overhead obstacle in the
middle of sightline.
JOURNAL OF TRANSPORTATION SAFETY & SECURITY 19

Figure 10. A typical comparison of crest vertical curves length design values between open road-
way and tunnels. SSD D sight stopping distance.

Examples of tunnel infrastructure overhead obstacles are: ventilators, xed sign,


or variable message sign (Figure 11). The value of vertical clearance is assumed to
be 4.5 m (tunnel trafc envelope height (Road and Transportation Research Asso-
ciation, 2006).

5.2. Driving comfort criterion


The comfort criterion for sag vertical curve is identical to the comfort criterion of
crest vertical curve. The effect of centrifugal acceleration is more crucial in sag
curves because the gravitational and centrifugal forces are combined, that is, both
forces function downwards (in the same direction).

5.3. Sag vertical curve analysis for road tunnels


Table 8 summarizes the results for dry tunnel, moist tunnel, EOT zone, and open
roadway. The results for tunnels (moist, dry) and EOT zone are identical. The
results for open roadways are based on night driving safety criterion (Bassan,
2012) that governs the driving comfort criterion. The geometric characteristics of
sag vertical curve reveal larger vertical radii after considering the object height as
opposed to crest vertical curves. The pertinent safety criterion equations for open
roadways are introduced in the Appendix.
The gray shaded cells are values that were determined by safety criterion. The
other cells results were generated by comfort criterion. Figure 12 presents the
results of tunnel sag vertical radii graphically. The implementation of safety
20 S. BASSAN

Figure 11. Typical hiding sightline for high vehicle in sag vertical curve due to overhead obstacle in
road tunnel.

criterion of tunnel vertical curves (crest and sag) assumes that SSD  L (Table 8,
Figure 12).
The ranges of percentage reduction compared to open roadways for sag curves
radii are 46% to 71% for tunnels and EOT zone, depending on the design speed.
The resulting lengths of sag vertical curves (L) as a function of the algebraic dif-
ference, for tunnels (moist and dry) based on safety and comfort criteria, are pre-
sented in Figure 13. Each line in the graph depicts a specic design speed between
50 and 120 km/h. The radii are bounded r to prevent too short vertical curve length
due to esthetical reasons (Equation 7).
Figure 14 presents a comparison of typical design values of sag vertical curve
lengths between tunnels and open roadways. The typical vertical curves design
values are presented in Figure 14 for the design speeds of 80, 100, and 110 km/h

Table 8. Sag vertical curve minimum radius for road tunnels (moist, dry, end of tunnel [EOT]) versus
open roadway.
Design Speed (km/h)

Site Characteristics 50 60 70 80 90 100 110 120

Tunnel (dry, moist, EOT), H D 4.5 m 390 560 760 990 1250 1670 2510 3630
% reduction from open road 67.8 65.0 64.5 71.2 70.0 65.7 56.2 46.1
(absolute reduction)b (820) (1040) (1380) (2450) (2910) (3200) (3220) (3110)
Open roadway (h2 D 0.15) 1210 1600 2140 2830 3530
Open roadway (h2 D 0.6) 3440a 4160 4870 5730 6740
a
Supplemental value for h2 D 0.60 (Vd D 80 km/h).
b
The comparison is valid for open roadway: Vd D 5070 kph, h2 D 0.15 m, and Vd D 80120 kph, h2 D 0.60 m.
JOURNAL OF TRANSPORTATION SAFETY & SECURITY 21

Figure 12. Graphical presentation of sag vertical curve minimum radius for road tunnels.

for comparison. The dashed lines signify open roadways and the solid lines signify
road tunnels (dry and moist). The dashed line for Vd D 80 km/h signies vertical
radius computed for h2 D 0.15 m.
Overall, the results of crest and sag vertical curves show signicant reductions of
vertical radii due to properties of tunnel friction and driving behavior in road

Figure 13. Design values for tunnels sag vertical curve length (m) based on safety, comfort, and
esthetics criteria.
22 S. BASSAN

Figure 14. A typical comparison of sag vertical curves length design values between open roadway
and tunnels. SSD D Sight stopping distance.

tunnels, which reduce the SSD. The possibility for smaller radii design (i.e., shorter
vertical curves) enables suppleness in tunnel mining due to exible design of
tunnel vertical alignment, and minimization of conicts between infrastructure utili-
ties during tunnel construction in addition to a potential reduction in tunnel length.

6. Summary and conclusion


The design of road tunnels is an essential infrastructure component in the highway
system and specically in highway design. Integrating tunnels in the highway sys-
tem assists to alleviate trafc congestion, maintain driving safety, reduce vehicle
emissions and air pollution, and protect landscape appearance. These benets are
especially worthwhile in the long run due to the relatively high construction cost
of the tunnel and its complementary systems.
The study implements reasonable criteria for obtaining sight distance and its
resulting vertical alignment of road tunnels versus open roadways based on friction
characteristics and driving behavior that are unique for road tunnels.
The assessment of SSD for road tunnel is performed by its adjusted parameters
to tunnel driving as follows:
1. Higher values of friction coefcient depending on the tunnel surface condi-
tions: moist or dry tunnel, based on dry friction properties and recent longi-
tudinal friction measurements for wet pavement condition in Israel (Gillo
Tunnel). Still, this nding is based on PIARC (2003) results of dry pavement
surface and tunnel pavement friction measurements in Israel. It may not be
suitable for other regions. Specically, friction eld tests are essential for
JOURNAL OF TRANSPORTATION SAFETY & SECURITY 23

every road tunnel project. A different location may have a different climate,
different geological characteristics, and different properties of pavement
asphalt concrete. Also, road tunnel surface requires maintenance (e.g., pave-
ment renewal, pavement surface cleaning, dust control, and winter related
tasks such as removal of ice if relevant to the climate characteristics), and
periodic friction testing.
2. A lower PRT compared to open roadways (1.5 sec for 50  Vd  80, and
2.0 sec for 90  Vd  120) based on research results presented in section 2.1
and the assumptions of drivers awareness and driver vigilance along the
bounded cross-section of road tunnels with narrow shoulders (Arndt et al.,
2011; Calendo et al., 2013; Calvi et al., 2012; Lemke, 2000; Yeung & Wong,
2013).
The SSD and vertical curve design values were generated for two tunnel surface
situations (moist and dry pavement) and for the EOT zone assuming wet pave-
ment for desirable design.
The impact of sight distance tunnel characteristics of the design of crest and sag
vertical curve radii is based on safety and driving comfort criteria similarly to open
roadways. Results of crest and sag vertical curves show signicant reductions in
vertical radii and vertical curve lengths in road tunnels compared to open road-
ways. The safety limitation for crest vertical curves is conceptually identical to
open roadways. The safety limitation for tunnels sag vertical curves is applicable
to trucks due to overhead obstacle when a certain percentage of trucks is expected
to use the tunnel (e.g., more than 5%). This limitation is critical for the design
speed range of 100 to 120 km/h in road tunnels (moist or dry). The safety criterion
of crest vertical curve is critical for the design speed range of 80 to 120 km/h and
90 to 120 km/h in analyzing moist and dry tunnels, respectively. The driving com-
fort criterion is critical for the lower range of design speeds in road tunnels. This
outcome is different from open roadways where safety criterion generally governs
the critical radii of crest and sag vertical curves.
The ranges of tunnel crest vertical radii percentage reduction compared to open
roadways are 48% to 67% for dry tunnels, 35% to 67% for moist tunnels, and 12%
to 44% for EOT zone, depending on the design speed. The ranges of tunnel sag ver-
tical radii percentage reduction compared to open roadways is 46% to 71% for tun-
nels as well as EOT zone
The reductions of tunnel vertical radii are caused by modications of tunnel longi-
tudinal friction and driving behavior in road tunnels, which reduce the stopping sight
distance to reasonable design values. However, road tunnel should be designed and
constructed by taking into account catastrophe potential related to re incidents and
re crashes. Components of the complementary systems such as design of re and
smoke resistance, re safety equipment, re detection devices, ventilation, structural
properties in the tunnel cross-section, communication systems, spoil removal, emer-
gency lay bays, and turning points for vehicles and pedestrians by connecting tunnels
should be integrated in the design process. Such components directly affect or are
24 S. BASSAN

affected by the geometric design of road tunnels and specically the tunnel cross-sec-
tion, the tunnel length, and the longitudinal grade. Still, the risk of a crash in tunnels
is lower than in open roadways, however, tunnel crash severity (when it occurs) is
higher (Lemke, 2000; Nussbaumer, 2007; Nussbaumer & Nitsche, 2008).
Returning to the major content of this study, the possibility for smaller vertical
radii design (i.e., shorter vertical curves) enables suppleness in tunnel mining due
to exible design of tunnel vertical alignment, and minimization of conicts
between infrastructure utilities during tunnel construction.
In particular, the resulted reduced tunnel vertical radii analyzed in this study
trigger the possibility of shorter tunnel construction length, lower tunnel construc-
tion cost and maintenance cost including components of complementary systems,
and exibility in tunnel construction due to ground and groundwater constraints.
Still, the selected tunnel alignment, even if it is shorter than other alternatives,
requires eld tests such as geotechnical characteristics, geological investigation,
vibrations and building inspection, and hydrologic factors.

References
American Association of State Highway and Transportation Ofcials. (2011). A policy on geo-
metric design of highways and streets. Washington, DC: Author.
American Association of State Highway and Transportation Ofcials. (1984). A policy on geo-
metric design of highway and streets. Washington, DC: Author.
Amundsen, F. H., & Ranes, G. (2000). Studies on trafc accident in Norwegian road tunnels.
Tunneling and Underground Space Technology, 15(1), 311.
Anundsen, F. H. (2009). Studies on Norwegian road tunnels II. An analysis on trafc accidents in
road tunnels 20012006. Road and Trafc Department, Trafc Safety Section. (Nr TS4-
2009). Oslo. Norway: Statens Vegvesen.
Arndt, O. K., Cox, R. L., Lennie, S. C., & Whitehead, M. T. (2011). Provision of sight distance
around concrete barriers and structures on freeways and interchanges. Transportation
Research Record, 2262, 2230.
Austroads. (2009). Guide to road design, part 3: Geometric design, AGRD03/09. Sydney, New
South Wales: Austroads Ltd.
Austroads. (2010). Guide to road tunnels, part 2: Planning design and commissioning. Sydney,
New South Wales: Austroads Ltd.
Bassan, S. (2012). Review and evaluation of stopping sight distance designCars vs. trucks.
Advances in Transportation Studies-An International Journal. Special Issue 1, 516.
Calvi, A., De Blasiis, M. R., & Guattari, C. (2012). An empirical study of the effects of road tun-
nel on driving performance. 5th International CongressSustainability of road infrastruc-
tures. ProcediaSocial and Behavioral Sciences, 53, 10981108.
Caliendo, C., De Guglielmo, M. L., & Guida, M. (2013). A crash prediction model for road tun-
nels. Accident Analysis and Prevention, 55, 107115.
Chang, M. S., Messer, C. J., & Santiago, A. J. (1985). Timing trafc signal change interval based
on driver behavior. Transportation Research Record, 1027, 2030.
Chowdhury, M. A., Warren, D. L., & Howard, B. (1991). Analysis of advisory speed setting cri-
teria. Public Roads, 55(3), 6571.
Design Manual for Roads and Bridges. (1999). Highway structures design, Vol. 2, Section 2, Spe-
cial Structures, Part 9, Design of Road Tunnels, BD 78/99. UK: HMSO.
JOURNAL OF TRANSPORTATION SAFETY & SECURITY 25

Dozza, M. (2013). What factors inuence drivers response time for evasive maneuvers in real
trafc? Accident Analysis and Prevention, 58, 299308.
Drew, D. R. (1968). Trafc ow theory and control. New York, NY: McGraw Hill.
Durth, W., & Bernhard, M. (2000). Revised design parameters for stopping sight distance. In R.
Krammes & W. Brilon (Eds.) Proceedings, 2nd International Symposium on Highway Geo-
metric Design, pp. 410421
Fambro, D. B., Fitzpatrick, K., & Koppa, R. J. (1997). Determination of stopping sight distances
(NCHRP Report 400). Washington, DC: Transportation Research Board, National Research
Council.
Federal Highway Administration. (2009). Technical manual for design and construction of road
tunnelsCivil elements (FHWA-NHI-10-034). Washington, DC: Federal Highway
Administration.
Green, M. (2000). How long does it take to stop? Methodological analysis of driver perception
Brake times. Transportation Human Factors, 2(3), 195216.
Greibe, P. (2008). Determination of braking distance and driver behaviour based on braking tri-
als. TRB 87th Annual Meeting. Compendium of papers. Washington, DC: Transportation
Research Board.
Hussein, M., Sayed, T., Ismail, K., & Van Espen, A. (2014). Calibrating road design guides using
risk-based reliability analysis. Journal of Transportation Engineering. ASCE, 140(9), 16.
Ismail, K., & Sayed, T. (2012). Risk-optimal highway design: methodology and case studies.
Safety Science, 50(7), 15131521.
Ibrahim, S. E., Sayed, T., & Ismail, K. (2012). Methodology for safety optimization of highway
cross-sections for horizontal curves with restricted sight distance. Accident Analysis and Pre-
vention, 49, 476485.
Johansson, G., & Rumar, K. (1971). Drivers brake reaction time. Human factors (vol. 13, pp. 23
27). Santa Monica. CA: Human factors Society.
Kosasih, D., Robinson, R., & Snell, J. (1987). A review of some recent geometric road standards
and their application to developing countries (Digest of Research Report 114). Crothworne,
Berkshire, UK: Transport and Road Research Laboratory, Department of Transport.
Lamm, R., Psarianos, B, & Mailaender, T. (1999). Highway design and trafc safety engineering
handbook. New York, NY: McGraw-Hill.
Layton, R. D. (1997). Stopping sight distance and decision sight distance (Discussion paper No. 8.
A). Corvallis, OR: Oregon State University, Transportation Research Institute. Retrieved
from http://www.oregon.gov/odot/hwy/accessmgt/docs/stopdist.pdf
Layton, R., & Dixon, K. (2012). Stopping sight distance (Discussion paper #1.). Corvallis, OR:
Oregon State University, Kiewit Center for Infrastructure and Transportation.
Lemke, K. (2000). Road safety in tunnels. Transportation Research Record, 1940, 170174.
Lerner, N., Huey, R., McGee, H., & Sullivan, A. (1995). Older driver perception reaction time for
intersection sight distance and object detection (FHWA RD-93-168). Washington, DC: US
Department of Transportation, Federal Highway Administration.
Navin, F. P. D. (1992). Reliability indices for road geometric design. Canadian Journal of Civil
Engineering, 19(5), 760766.
Navin, F. P. D. (1990). Safety factors for road design, can they be estimated? Transportation
Research Record, 1280, 181189.
Norwegian Public Road Administration. (2004). Road tunnels manual 021, standard.
Nussbaumer, C. (2007). Comparative analysis of safety in tunnels. Young Researchers Seminar, Brno,
Czech Republic. Retrieved from www.ectri.org/YRS07/Papiers/Session-9/Nussbaumer.pdf.
Nussbaumer, C., & Nitsche, P. (2008). Safety of road tunnels. Trafc safety in highway and
expressway tunnels. Vienna, Austria: Austrian Road Safety Board.
26 S. BASSAN

Olson, P. L., & Sivak, M. (1986). Perception reaction time to unexpected roadway hazards.
Human Factors, 28(1), 9196.
PIARC. (2001). Cross-section geometry in unidirectional road tunnels, technical committee on
road tunnels operation (Report 05.11B). Paris, France: World Road Association.
PIARC. (2003). Road safety manual. Recommendations from the world road association.
PIARC technical committee on road safety (C13) version 1. Retrieved from http://www.
piarc.org/en/order-library/6286-en-Road%20Safety%20Manual.htm
PIARC. (2004). Cross section design of bi-directional road tunnels, technical committee on road
tunnels operation (Report 05.12 B). Paris, France: World Road Association.
PIARC. (2008). Human factors and road tunnel safety regarding users. PIARC Technical Com-
mittee C3.3, Road Tunnel Operation (Report R17, ISBN 2-84060-218-0). Paris, France:
World Road Association. Retrieved from http://www.piarc.org/en/order-library/6173-en-
Human%20factors%20and%20road%20tunnel%20safety%20regarding%20users.htm
Road and Transportation Research Association, Working Group Trafc Routing and Road
Safety. (2006). Regulations for the equipment and operation for road tunnels. Koln, Germany:
FGSV-Verlag.
Road and Transportation Research Association. (2011). Guidelines for the design of motorways.
FGSV. RAA. Edition 2008. Translation 2011. Cologne, Germany.
Sivak, M., Olson, P. L., & Farmer, K. L. (1982). Radar measured reaction times of unalerted driv-
ers to brake signals. Perceptual and Monitor Skills, 55(2), 594.
Summala, H. (2000). Brake reaction times and driver behavior analysis. Transportation Human
Factors, 2(3), 217226.
Summala, H. (2002). Behavioural adaptation and drivers task control. In R. Fuller & A. J. Santos
(Eds.), Human factors for highway engineers (Ch. 13, pp. 189200). Oxford, UK: Pergamon,
Elsevier Science. Ltd.
Thanesuen, S., Kagaya, S., & Uchika, K. (2005). Trafc management in terms of speed limit in
Hokkaido area. Journal of the Eastern Asia Society for Transportation Studies, 6, 22472262.
Wortman, R. H., & Matthias, J. S. (1983). Evaluation of driver behavior at signalized intersec-
tions. Transportation Research Record, 904, 1020.
Yeung, J. S., & Wong, Y. D. (2013). Road trafc accidents in Singapore expressways tunnels.
Tunneling and Underground Space Technology, 38, 534541. 2013.
JOURNAL OF TRANSPORTATION SAFETY & SECURITY 27

Appendix

Table A1. Roadway 60, Israel: Gillo tunnel wet pavement friction measurements: 2009, 2010.
Start KM End KM Direction of Travel Wet Friction Coefcient Date

73.959 74.066 3 0.62 September 2, 2010


73.959 74.066 4 0.65 March 17, 2009
73.959 74.066 3 0.46 March 17, 2009
73.959 74.066 4 0.66 September 2, 2010
74.066 74.173 4 0.63 September 2, 2010
74.066 74.173 3 0.63 September 2, 2010
74.066 74.173 4 0.64 March 17, 2009
74.066 74.173 3 0.43 March 17, 2009
74.173 74.28 4 0.64 September 2, 2010
74.173 74.28 3 0.63 September 2, 2010
74.173 74.28 4 0.62 March 17, 2009
74.173 74.28 3 0.44 March 17, 2009
74.28 74.386 3 0.63 September 2, 2010
74.28 74.386 3 0.43 3 March 17, 2009
74.28 74.386 4 0.66 September 2, 2010
74.28 74.386 4 0.63 March 17, 2009
74.386 74.493 4 0.66 September 2, 2010
74.386 74.493 3 0.46 March 17, 2009
74.386 74.493 3 0.64 September 2, 2010
74.386 74.493 4 0.62 March 17, 2009
74.493 74.6 4 0.63 September 2, 2010
74.493 74.6 3 0.55 September 2, 2010
74.493 74.6 3 0.4 March 17, 2009
74.493 74.6 4 0.59 March 17, 2009
74.6 74.707 4 0.6 March 17, 2009
74.6 74.707 3 0.43 March 17, 2009
74.6 74.707 3 0.53 September 2, 2010
74.6 74.707 4 0.6 September 2, 2010
74.707 74.813 4 0.45 September 2, 2010
74.707 74.813 4 0.6 March 17, 2009
74.707 74.813 3 0.48 March 17, 2009
74.707 74.813 3 0.58 September 2, 2010
74.813 74.92 4 0.39 September 2, 2010
74.813 74.92 3 0.57 September 2, 2010
74.813 74.92 4 0.59 March 17, 2009
74.813 74.92 3 0.44 March 17, 2009
Average 0.5614
Standard deviation 0.0891
Condence interval 0.0309

3: Northbound (downward); 4: Southbound (upward)


28 S. BASSAN

Tabls A2. Roadway 60, Israel: Open road prior to Gillo tunnel, wet pavement friction measurements:
2009, 2010.
Start KM End KM Direction of travel Wet friction coefcient Date

73.1 73.207 3 0.53 September 2, 2010


73.1 73.207 3 0.47 March 17, 2009
73.207 73.317 3 0.48 March 17
73.207 73.317 4 0.57 March 17
73.207 73.317 3 0.51 September 2, 2010
73.207 73.317 4 0.58 September 2, 2010
73.317 73.423 4 0.55 September 2, 2010
73.317 73.423 3 0.38 September 2, 2010
73.317 73.423 4 0.49 March 17
73.317 73.423 3 0.32 March 17
73.423 73.531 4 0.5 September 2, 2010
73.423 73.531 3 0.33 September 2, 2010
73.423 73.531 4 0.34 March 17
73.423 73.531 3 0.28 March 17
73.531 73.639 4 0.51 September 2, 2010
73.531 73.639 3 0.26 March 17, 2009
73.531 73.639 3 0.33 September 2, 2010
73.531 73.639 4 0.32 March 17, 2009
73.639 73.746 4 0.33 March 17, 2009
73.639 73.746 3 0.43 September 2, 2010
73.639 73.746 4 0.59 September 2, 2010
73.639 73.746 3 0.27 March 17, 2009
73.746 73.853 4 0.34 March 17, 2009
73.746 73.853 3 0.44 September 2, 2010
73.746 73.853 3 0.27 March 17, 2009
73.746 73.853 4 0.39 September 2, 2010
73.853 73.959 4 0.33 March 17, 2009
73.853 73.959 3 0.48 September 2, 2010
73.853 73.959 3 0.32 March 17, 2009
73.853 73.959 4 0.4 September 2, 2010
Average 0.4113
Standard deviation 0.1033
Condence interval 0.0358

3: Northbound (downward); 4: Southbound (upward)

Sag vertical curves equations for open roadways derived by safety criterion
(Bassan, 2012):

SD 2
Rv D for SSD  L (A1)
2 h2 C h3 C SD tan1
200SD 20; 000  h2 C h3 C SD tan1
Rv D for SSD > L (A2)
A A2

h3 D the design value for the headlights height: 0.6 m (passenger cars).

Anda mungkin juga menyukai