User cost customization for a Florida bridge management system


Material Information

User cost customization for a Florida bridge management system
Physical Description:
xi, 144 leaves : ; 29 cm.
Soares, Roberto
Publication Date:


Subjects / Keywords:
Bridges -- Maintenance and repair -- Management   ( lcsh )
Bridges -- Data processing -- Florida   ( lcsh )
Civil Engineering thesis, Ph.D   ( lcsh )
Dissertations, Academic -- Civil Engineering -- UF   ( lcsh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1999.
Includes bibliographical references (leaves 129-142).
Statement of Responsibility:
by Roberto Soares.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 021577915
oclc - 43707747
System ID:

This item is only available as the following downloads:

Full Text







Dedication, to

my wife,



my three children,






I am extremely grateful to Dr. Najafi the chairman of my committee who has

given me his continuous support, guidance, and encouragement throughout my career at

the University of Florida. The proposal of the Bridge Management System to Florida

Department of Transportation was originally initiated by Dr. Najafi which has provided

me with an opportunity to conduct this dissertation work. I am also grateful to Dr.

Thompson for his guidance and constant support throughout my career at the University

of Florida. A sincere and wholehearted appreciation is extended to my committee

members Dr. Shrestha, Dr. Glagola, and Dr. Foti for their wise guidance in the work of

this dissertation.

I extend appreciation to the National Highway Institute (NHI) for selecting me as

the 1998 recipient of the Eisenhower Fellowship Grant, for their fellowship during my


I offer special thanks to all library personnel at the University of Florida, at the

Department of Transportation in Washington, DC, and at the American Truck

Association in Virginia, and to the personnel who helped me locate specific material

during my research.

With all my heart, I give very special thanks for the support received from my

family; and, above all, I thank God for providing me protection and care.



LIST OF TABLES ........................................ vii
LIST OF FIGURES ................................... ix
ABSTRACT ....................................... x

INTRODUCTION ................
1.1. BMS Development.......
1.2.What is BMS in Pontis .....
1.3. Focussing on Pontis ......

2.1 Problem .................
2.2 Survey ..................
2.3 User Cost Need ...........
2.4 Surplus Theory on User Cost
2.5 User Cost Weight Factor ...
2.6 Pontis Default Values Origin
2.7 Statement of the Hypothesis .
2.8 Validation ...............

3.1 Accidents ................
3.2 User Costs ...............
3.3 Economic Evaluation ......
3.4 BM S ...................
3.5 Remaining Relevant Entries .

4.1 Pontis Databas ............
4.2 Prediction Models ........
4.3 Cost Models .............
4.4 Program Integration Model .

. . . . . 1
. .. .. .. .. .. ... . .. ... 2
................ ....... ... .. ... ...... 6
. . . . . .. 9

. . . . . 10
. . . . . 10
. . . . . 10
.......................... ............ 11
......................... ........... 12
................... ................ 13
. . . . . 14
. . . . . 15
...................................... 16

. . . . . 17
. . . . . 18
. . . . . 2 1
. . . . . .2 4
. . . . . .2 5
. . . . . 2 6

...................... ................27
. . . . . .2 8
. . . . . 3 1
. . . . . 3 3
. . . . . 3 8
. . . 13
.2 .. ... .. ... ..... .. 1
.6 ..... ........
. . . 9

.. .. .. .. .. ... .. .. .. .. .. .. 107
.. .. .. .. .. ... .. .. .. .. .. .. 108
.. .. .. .. .. .. .. .. .. .210
.. .. .. .. .. 112
.. .. .. .. .. ... .. .. .. .. .. 1.25
.13 .......... .........2

.14 .......... .........2
.15 .......... .........2
. ..................316

.17 .......... .........3
.1 ............. 8

PONTIS USER COSTS MATHEMATICAL MODEL.......................... 41
5.1.Benefits and User Costs ......................... ..... ......... 41
5.2 FDOT Default Values Policy .......................... ........ 42

5.3 User Cost Models ............................................ 44
5.4 Benefit of Widening .......................................... 45
5.5 Benefits of Raising ..................... ........................ 47
5.6 Benefit of Strengthening ....................................... 47
5.7 Benefits of Replacement ......................... ................. 48
5.8 Detour Cost ...................... .......................... .. 50
5.9 Application Example .......................................... 50

TRAVEL TIME COSTS FOR BMS ....................................... 55
6.1 Background ............................... ................. 55
6.2 Theoretical Basis for Travel Time Evaluation ......................... 57
6.3 Percent Wage Index Analysis ..................................... 59
6.4 Comparing Travel Time Values ................................. 64
6.5 Non Business Travel Time .................. ................... 65
6.6 Business Travel Time ............................................ 66
6.7 Value of One Hour Travel Time ................ .................. 69
6.8 Truck Travel Time Cost for Florida ................................ 71

7.1 VOC Related Factors ......................................... 73
7.2 Running Costs Calculation Methodology ............................. 78
7.3 Variable costs .................................................. 78
7.4 VOC Truck Customization ................ ...................... 82
7.5 New CVv Value ............................................. 86

8.0 Background ................................................ 88
8.1 Bridge Accident Cost Evaluation Methodology ........................ 88
8.2 Comprehensive Fatality and Injury Costs ............................ .89
8.3 Conversion of MAIS into ABC Classification ......................... 91
8.4 FDOT Crash Data Preparation .................................... 92
8.5 Bridge Related Accident Average Cost ................... ..... ... 95

PONTIS USER COST SENSITIVITY ANALYSIS .......................... 97
9.0 Background ................................................... 98
9.1 Parameters Definitions ........................................... 99
9.2 Scenarios for Simulation ........................................ 100
9.3 Comparing Results From Basis 1 and Basis 2 ........................ 105

CONCLUSIONS AND RECOMMENDATIONS ............................ 108
10.1 Conclusions ..................................................108
10.2 Recommendations ........................................... 109


PONTIS OUTPUT REPORTS ............................... 111
RESPONSIBILITIES ............................... .118

REFERENCES .................................... 129
BIOGRAPHICAL SKETCH ......................... 143


Table P

1. Comparison of the Main Attributes Between Pontis and Bridgit BMS Models .... 6

2. Comparative Types of User Costs in North Carolina, Indiana, and Pontis BMS
Models ..... .......................................... 8

3. Average Number of Traffic Fatalities per Each Weekday--1975-1995 ......... 19

4. Travel Time Cost--Dollar per Hour-Base 1993--Miller Approach ............. 24

5. Travel Time Cost-Dollar per Hour-Base 1993--Zaniewski Approach .......... 24

6. Pontis Definition of Bridge Element's Environment ....................... 30

7. Basic Inventory Information for Each Bridge ............................. 32

8. Bases for the Need and Benefit Calculations on Improvement Projects ......... 38

9. Cost Matrix Values Adopted by FDOT .............................. 42

10. Composite Listing of Update Travel Time Costs .......................... 65

11. Value of One Hour Travel Time for Business and Non-Business Trips by
Vehicle Category ................................................. 69

12. Research Result Value for TTC ....................................... 71

13. Vehicle Costs Classification Between Fix and Variable Costs ................ 79

14. Variable Costs Assigned at Operating Vehicle Cost Studies ................. 80

15. Estimated Cost for Fuel Per Mile by Truck Category ....................... 84

16. Fuel Costs Distribution by Equipment Type .............................. 84

17. Maintenance Costs for Trucks- CPM/CPK .............................. 85

18. Tire Costs Descriptive Statistics--CPM and CPK ............... ........ 86

19. New VOC Value for Pontis CV, Default Value ......................... 86

20. Injury Costs by Injury Type--Year 1996 ................................ 87

21. Comprehensive Injury Costs, Years 1994 and 1996 ........................ 92

22. Economic Injury Costs, Years 1994 and 1996 ............................ 92

23. Bridge Related Accident Unit Costs, for Florida .......................... 96

24. User Cost Default from Pontis and Research Result Values ................. 99

25. Benefit Percent Change Due to User Costs Percent Change Under Different
Scenarios (Using Pontis Default Values) .............................. 102

26. User Cost Sensitivity Ranking Against % changes in Benefit-- "Basis 1" --
(Using Pontis Default Values) ..................................... 103

27. User Cost Sensitivity Ranking Against % Changes in Benefits-- "Basis 2" --
(Using Research Resulted Default Values) ............................ 104

28. Benefit Percent Change Due to User Costs Percent Change Under Different
Scenarios--Basis 2--(Using Research Resulted Default Values) ............ 105

29. User Costs Variability Under two Different Calculation Basis .............. 106

30. Car, Vans and Small Trucks VOC Values .............................. 110

31. Travel Time Costs for Auto, and Five Truck Types ....................... 110


Figure page

1. Consumers' Surplus For One Period .................................... 13

2. Data Treatment Flowchart ............. .............................. 27

3. Overall BM S Structure ............................. ................... 28

5. Data Flows in Preservation Modeling .................................... 35

6. Flow Chart for BMS TTC Development ................................ 56

7. Flow Chart for BMS VOC Development ................................ 72

8. Flow Chart for the Average Accident Costs Development ..................... 87

9. Flow Chart for User Cost Sensitivity Analysis ................... ....... 97

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Roberto Soares

August 1999

Chairperson: Fazil T. Najafi
Major Department: Civil Engineering Department

Bridge Management Systems (BMS) are tools developed to help decision makers

to prioritize projects that gives the highest benefit/cost ratio, where benefits are road user

cost benefits. They are identified by the savings achieved in not spending travel time (TT)

and vehicle operating costs (VOC) doing a detour, and reducing the risk of bridge related

accidents. The state of Florida plus 38 states adopted the BMS Pontis to manage their

bridge inventory.

The hypothesis of this dissertation is that values benchmarked from other areas

are not suitable to replicate the Florida reality. In order to prove this hypothesis a multi-

step methodology was used including bridge-related accident cost using data from 11,332

bridge related accidents which occurred in 1996, sensitivity analyses, using 524 bridges

under 39 different scenarios, modeled by Pontis-BMS version 3.4. The original user cost

default values in Pontis are: $37,600 for each bridge related accident (Ca); $0.25 per

kilometer for VOC costs (Cv), and $19.34 per hour for travel time (Ct).

The findings of this research show the need to raise the Pontis-BMS, user cost

default value rates to 45.05, 24.0 and 16.6 percent respectively for accident costs, VOC

per kilometer costs and travel time values. The new values are $68,404.39 for Ca;

$0.3138 for C,, and $22.55 for C,.

The Project Attractiveness Index (PAI), measured by the benefit/cost ratio

increased 25.96 % with the use of the new user cost parameters. Sensitivity analyses

showed that detour per kilometer is the most sensitive parameter, followed by the detour

per hour, and finally the average bridge related accident cost parameter. Using the new

default values in place of the old ones, deficient bridges that are related with accidents

receive a higher prioritization order to be fixed, giving to decision-makers the possibility

to reduce bridge accident risk.


The objective of this dissertation is to develop three new users cost default values

to be used in a Bridge Management System (BMS) software known as Pontis. The users

cost default values are named Travel Time Costs (TTC), Vehicle Operating Costs (VOC),

and Average Accident Costs (AAC).

Chapter 1 presents BMS and why Pontis BMS was selected in this research.

Chapter 2 presents reasons why it is necessary to develop a new set of user costs default

parameters to be used in Pontis, applicable to Florida BMS. The reasons are the

framework for the hypothesis set in this dissertation. Chapter 3 is the literature review

focusing on the relevant information related to user cost parameters to be used in the

Florida BMS model. Chapter 4 presents the Pontis basic concepts, and the models and

their interactions used to optimize BMS. Chapter 5 discusses how the user models are

used in BMS to quantify, in economic terms, the potential safety and mobility benefits of

functional improvement and elimination of deficiencies to bridges. A hypothetical

example is offered showing the evaluation of total benefits. Chapter 6 discusses TTC and

the methodology used to develop a TTC parameter under BMS. A predictive

mathematical model and the generation of a new TTC are presented in this chapter.

Chapter seven discusses VOC and the methodology used to develop a VOC parameter. A

predictive mathematical model and the generation of a new VOC are presented in this

chapter. Chapter 8 presents bridge-related accident costs and the methodology used to

develop a Florida bridge related AAC parameter. A predictive mathematical model and

the generation of a new AAC value are presented in this chapter. Chapter 9 presents a

sensitivity analysis study comparing the original set of Pontis user cost default values

with the new set of user cost default values developed in this research. The results of the

sensitivity analysis study confirm or negate the hypothesis in this research. Chapter 10

presents the conclusions and the recommendations derived from this research. The main

conclusion is that the developed default user cost values are suitable to Florida BMS to

prioritize projects and to allocate funding to improve Florida bridges.

1.1. BMS Development

New York State can be considered the pioneer in BMS studies using main-frame

computers. Others states also developed BMS studies. However, the one from North

Carolina can be considered the most documented. BMS developments using personal

computers started by the middle of the 1980's resulting in the Bridgit and Pontis BMS

software. This research focuses on the BMS Pontis software because FDOT uses it. The

AASHTO Transportation glossary defines BMS as a system designed to optimize the use

of available resources for the inspections, maintenance and replacement of bridges

(AASHTO, 1994b).

The United States has by far the largest number of bridges in the world: 600,000

as compared to second-place Germany with 160,000 (National Bridge Inventory NBI,

1998). Considering that most bridges built in the past were designed for a service life of

50 years, this leads one to conclude that bridges constructed before 1960 are at the end of

their service life. This also indicates that there will be growing replacement needs for

those structures, and a growing need for deck repair and replacement for the bridge

population constructed during the 1960s (Amrhein 1977).

State Highway Agencies (SHA) responsible for managing the nation's bridges

must use limited funds as wisely as possible. A Bridge Management System (BMS) can

help SHA evaluate current and future conditions, needs, and determine the best mix of

maintenance and improvement work on the bridge network over time with and without

budget limitations.

According to the American Association of State Highway and Transportation

Officials (AASHTO, 1993) the genesis of BMS is linked with the establishment of the

National Bridge Inventory (NBI) in the 1970s. The purpose of the NBI was to inform

Congress of the status of the nation's bridges, define the magnitude of bridge needs,

support national bridge inspection standards and provide for defense information needs.

The Federal Government required that every bridge on public roads and larger than 6.30

meters (20 feet) in total length be described in a national database. Legislation (CFR,

1988) also required that bridges be inspected and evaluated at regular intervals not to

exceed two years following the National Bridge Inspection Standards (NBIS, 1995).

Scholars (Tuner and Richardson, 1994) and practitioners (Shirole et al., 1994) link the

pre-genesis of BMS with the Silver Bridge collapse in 1967 between Point Pleasant,

West Virginia and Galllipolis, Ohio that killed 46 people (Mair, 1982). This disaster was

highly publicized and drew attention to the aging condition of the nation's bridges.

As of November 1995, the United States had over 590,000 bridges, of which

about 100,000 of these bridges were built prior to 1935. Nearly 187,504 bridges are

classified as structurally deficient or functionally obsolete due to the increase of legal

weight loads and traffic volumes, combined with the effects of weather and chemicals

The budget needed to remedy the national bridge deficiencies is projected to be over $ 80

billion or an average cost of over $ 426,000/bridge (BRM, 1995).

The Florida Department of Transportation (FDOT) Bridge Inventory, 1977

Annual Report, shows a total bridge inventory of 11,156 (Amrhein, 1977) bridges and a

recent publication shows a sample of 941 bridges in Florida with functional needs at an

average projected cost of over $212,000/bridge to overcome these needs (Thompson et

al., 1998). The difference of over $200,000/bridge between the average national cost and

the average Florida cost to overcome the bridge needs, is mainly due to the degree of

uncertainty in the forecasted national value, and the results of an aggressive maintenance

program to extend the useful life of Florida bridges, thereby minimizing the need to

replace a large number of bridges within a short period of time. However, shortage of

funds forced public officials, administrators, and bridge engineers to learn how to manage

limited funds as wisely possible.

By the 1980s many states started to address the problem shown by NBI by

developing new analytical methods and procedures to allocate funds among different

types of problems to overcome the bridge network deficiencies. Wisconsin (Hyman and

Hughes, 1983), North Carolina (Niessner, 1979), Pennsylvania (Krugler,1985), New

York (Wade and Larder, 1973), Kansas (Kulkarni et al., 1984), and Indiana (Youngtae

and Sinha, 1997) were the pioneers in developing customized Bridge Management

Systems (BMS) using mainframe computers. The purpose of a BMS is to combine


management, engineering and economic inputs in to determine the best actions that can

be taken on a network over time.

By the middle of the 1980s, many states had independently come to the

conclusion that they needed better bridge management tools, and the national efforts

began to converge. Two competing projects were formed, one by the Federal Highway

Administration (FHWA), and one by the National Cooperative Highway Research

Program (NCHRP). The FHWA project began with a series of 49 workshops held around

the nation that resulted in the 1992 release of Pontis (Thompson and Shepard, 1994), five

years later the 3.2 version of Pontis was released, and by August 1998 the 3.4 version of

Pontis was released. A new version of Pontis is expected by the year 2000. The NCHRP

conducted a study known as Report 300 (NCHRPR, 1987). From this report, the panel

overseeing the project decided to produce a software package which followed the

principles outlined in Report 300. This resulted in the 1995 release of Bridgit (Lipkus,


The main differences between the features of Pontis and Bridgit with respect to

analysis, policy and optimization are listed on Table 1. The ability of Pontis to provide

analysis at a network level seems to be the main reason why universities and consultants

strongly recommend the use of Pontis to SHAs. Both BMS's use the principle of a

systematic analysis of expected benefits and costs as prescribed by Executive Order

12893, "Principles for Federal Infrastructure Investments" (F.R., 1994). However, at the

national level, 39 states were Pontis subscribers (Pontis, 1998b) as of May 1, 1998,

Florida is in the advanced stage of Pontis implementation in their BMS.

Table 1. Comparison of the Main Attributes Between Pontis and Bridgit BMS Models

Attributes Pontis Bridgit
Analysis Level Network Level and Project Project Level
Policy Policy Optimization Subjective Policy Choices
Optimization Top-Down Application of Botton-up Aggregation to
Approach Policies to Project Needs System Wide

1.2.What is BMS in Pontis

BMS is a complex set of formal procedures for analyzing bridge data, which

provides information from which to recommend project prioritization and schedules

considering budget and policy constraints. The Pontis analytical software is only one part

of the minimum BMS requirements. It requires data input generated from the output of

the bridge inspection process, the physical inventory, traffic and accident data, cost

models, deterioration models, definitions and policies. The software analyzes the data at

different levels of analysis and action categories. The output that monetizes the needs is

presented into two main categories: Agency Costs and User Costs. Agency costs are the

amount of funds required for maintenance and repairs, rehabilitation, and bridge

replacement. User Costs are known as the benefits received by the user when a bridge

deficiency is removed (NHI, 1996).

The cost side of the BMS (Agency Costs), is accepted almost without controversy.

The benefit side of the BMS (User Costs), traditionally, is not considered in the bridge

investment decision process. The scarcity of adequate methodology available to evaluate

bridge user costs with accuracy, and the lack of a clear understanding of the role of the

bridge user costs, are probably the main factors that support this tradition. Although the

user costs generated by bridge deficiencies are not paid or assumed directly by

government, the public is both the user and the ultimate owner of the bridge. Thus, the

user costs generated by bridge deficiencies as well as the ownership costs associated with

bridge maintenance, rehabilitation and replacement should be considered in the decision-

making process for bridge improvement. Today's society has a clear understanding of this

concept and for this reason, they have a legitimate right to demand from public officials

the use of a scientific approach, instead of the political approach, in decisions that involve

taxpayers money.

The term user cost in the BMS output means road users "out of pocket" money

spent to overcome bridge deficiencies. In the BMS economic analysis it is named user

benefits, since the users are saving in vehicle operating costs, travel time costs and

accident costs. Federal Legislation (F.R., 1994) demanded a systematic analysis of

expected benefits and costs where benefits and costs should be quantified and monetized

to the maximum extent possible. SHA (FDOT, 1997) estimates that user costs are ten

times larger than agency costs. If this estimate is true, it is envisioned that future

refinement in BMS decision making process will challenge the actual status quo of the

SHA users.

The theoretical basis for the methodology for economic analysis using user cost

was initially presented in the 1960 AASHTO Report, "Road User Benefit Analyses for

Highway Investments (AASHTO, 1960)," and updated by the 1977 ASHTO publication,

"A Manual on User Benefit Analysis of Highway and Bus-Transit Improvements

(AASHTO, 1977)" known as the "Red Book." Besides the fact that both publications

present a procedure to develop a methodology to conduct highway user economic

analyses for highway improvements, the publications do not address properly the BMS

user cost issue. The later presented more clarity for the BMS design development

observed in the decade of 1980. The Pontis BMS was the only one that followed the

theoretical basis recommended by the "Red Book" as a base to develop the user costs

dedicated for BMS. Its basic model of bridge user costs has three components. These are

travel time costs (TTC), vehicle operating costs (VOC), and average accident costs

(AAC), as shown in Equation 1. The Indiana BMS model and the North Carolina BMS

model do not consider all user cost parameters as indicated in Table 2.

User Costs = TTC+VOC+AAC (1)

Table 2. Comparative Types of User Costs in North Carolina, Indiana, and Pontis BMS

Indiana Carolina Pontis
Types of User Costs (Son et al., (Johnston et (Pontis,
1996) al., 1994) 1997)

Yes No Yes No Yes No

Detour costs due to load restrictions x x x

Detour costs due to vertical clearance x x x

Travel costs due to load restrictions x x x

Travel costs due to vertical clearance x x x

Narrow width x x x*

Accident costs due to narrow width x x

Accident costs due to vertical clearance x x

Accident costs due to poor alignment x x
*Included in the accident cost mathematical model

The implementation processes for Pontis BMS generally creates a need for the

SHAs to re-engineer the active process related to bridges including clear definitions and

required integrated actions between all the stakeholders of the process. As mentioned

before, users are by definition, the ultimate owners of the bridge, and for this reason the

benefit side of the software should always be considered in the decision making process.

However, Pontis allows practitioners to modulate the cost weight in the range of 0-100 %

during the evaluation of the agency costs. This provision is to conform to the level of

certainty that each practitioner or decision-maker has about user costs. The confidence

level in dealing with user costs when compared with agency costs is low.

1.3. Focussing on Pontis

Because Pontis is already installed and operational in Florida, research is

necessary to improve components of the user cost model, providing a roadmap for the

agency to supplement its existing data resources in the future to ensure continued

improvement in the effectiveness of the models. This research has the objective to fulfill

this need, and will focus on user cost parameters to be used in the Florida Bridge Network

(FBN) system, assuming that the agency costs to run BMS software is in good standing.

A recent survey (University of Florida, 1998) was conducted with all 50 states to

try to find ongoing studies that might not yet have been published relating to BMS user

costs. No such studies were found, which reinforces the need for this research.


2.1 Problem

The objective of this dissertation is to develop a new set of user cost default

parameters (TTC, VOC, and AAC) to be used in Pontis BMS in Florida. The main

problem was that the FDOT BMS team during the implementation of PONTIS BMS first

recognized that the default user cost values used in Pontis is not applicable to Florida, and

for this reason they funded UF to conduct a user cost study (FDOT, 1997).

The second motive that led to pursue this research was framed by the Highway

Capacity Manual which states that traffic conditions are a function of the roadway

conditions, the environment and the driver behavior (HCM, 1994). In another words, it is

not safe to state that user cost parameters developed for bridge network from state A is

applicable to the bridge network of state B.

The third motive was the result of the survey between all 50 states, which

confirmed the need to develop new BMS user cost parameters. The following sections

detail the problem and the justifications to develop a new set of user costs default values

for Florida.

2.2 Survey

The Florida Department of Transportation (FDOT) funded the University of

Florida (UF) to perform a research study to develop user cost models for the

Department's implementation of AASHTOWareTM PontisTM One of the tasks of this

research was to conduct a survey of all 50 SHAs to find out how many states are using

user cost in their decision making related to their BMS. From 72% overall response, 83.3

% have not undertaken any work to develop user cost model factors for their BMS

implementation. Fifty percent (50%) of the respondents have requested additional

information on how to formulate and apply the user cost model efficiently, 55.5% are

concerned with truck factors related with BMS, 48.2% are concerned with bridge related

crashes, 13.8% are concerned with bridge work zone related costs, and 11.1 % are

concerned with bridge deck roughness related costs. The results of this survey confirmed

the need for this research with the UF research team. Furthermore, according to Wall and

Smith (1998) user cost rates, and cost rate assigned to user delay (i.e., the value of time)

are by far the most controversial. Gillespie (1998) states that the state-of-the-art

calculation of vehicle operating cost is still ill defined, and that the state-of-the-art in

estimating accident costs is undergoing rapid change.

2.3 User Cost Need

The use of user costs, as a tool to perform economic evaluations for highway

projects, historically was never a popular choice among decision-makers. The need to

select user costs as an economic tool in the economic evaluation of transportation projects

emerged from increasing public scrutiny or hostility, concerns with the environment, and

legal requirements to avoid undesirable effects to the society.

User costs for road users have existed since 1920. However, the user costs

parameter for BMS models came into practice in the 1990's, and today it is still being


refined. There is a real need to increase the knowledge about the importance of user cost

in the economic evaluation of bridge related investments. Pontis employs user cost

models to primarily set priorities, since absolute need is established by the use of level-

of-service standards. The FDOT objective is to remove bias on the existing Pontis user

cost model. The FDOT desires to develop a new FDOT user cost model applicable to

Florida condition.

2.4 Surplus Theory on User Cost

The surplus theory states that bridge improvements bring benefits to the users. In

other words, user costs will be decreased if bridges are maintained regularly and properly.

Service is defined as the design level of service (LOS) which is considered the basis for

an economic engineering analysis. When a bridge develops a deficiency that generates

restriction in traffic flow, this disservice generates a user cost increase. Figure 1 shows

the consumer surplus for one period. The estimate of the net user benefits of a bridge

improvement is represented by the area UoABU1 where a bridge improvement will reduce

the user costs that would have been U0 to U, and the traffic volume is expected to

increase from the base level Vo to V,.

The formula for consumer surplus is (U0 UI) (Vo + V,)/2, or the difference in

user costs times the average traffic volume. This formula can be shown to be the total

benefits to new users plus benefits to present users, as follows:

Benefits to present users (area UoAC U,) = (U0- U ) Vo
Benefits to new users (area ABC) = (Uo- U1) (V, Vo)/2
Total Benefits = (U0 U,) Vo + (Uo U,) (V, V)/2
Total Benefits = (U0 U,) (Vo + V,)/2

,--------- Demand for one Period
U0o ------- "Surplus" Benefits to New Users
Benefits to


Figure 1. Consumers' Surplus For One Period

User costs at the level Uo are the costs that the traveling public is experiencing for

travel over a bridge that is operating with a volume deficit of AV due to bridge

deficiencies. These correspond to a low LOS that causes a reduced traffic volume ofV0.

User costs at the level U, are the costs that bridge users are experiencing after the bridge

improvement, where its LOS has been improved, offering an increase in traffic volume

corresponding to V,.

2.5 User Cost Weight Factor

The consumer surplus theory provides adequate support to user costs benefits, and

BMS practitioners accept this fact. What they question are the user costs parameters used in

the mathematical models developed to evaluate the user costs benefits magnitude. These

parameters are specifically the values assigned to travel time costs, vehicle operating costs,

and accident costs. The lack of an adequate methodology to develop these cost parameters

has led to a large range of user cost values allowing decision makers the possibility of using

vague values which may result in wrongfully prioritizing bridges for maintenance work.

This approach was the main reason in lowering the state highway engineer's and project

annalist's credibility who relied on vague values of user cost parameters. In order to

improve this situation, the BMS AASHTOWareT PontisTM allows practitioners to select

the weight of user costs benefits according to the credibility assigned by each practitioner

for user cost values used. The weight range selection varies from 0 to 100 percent (Pontis


This observed level of variability of user costs in BMS could have several plausible

justifications. One can be the methodology used in the development of user cost parameters

actually used in BMS applications. For example, travel time costs were developed to be

used in urban traffic demand forecast studies where the focus is travel time savings, which

is totally different from BMS. Vehicle operating costs were developed with emphasis on

passenger cars, and BMS requires focus on trucks. Accident costs were developed with

emphasis on roadway accidents as a general rule, not focusing on bridge-related accidents.

2.6 Pontis Default Values Origin

There is a need for the use of user cost parameters in Pontis BMS software to

perform the benefit cost analysis on each bridge in the network on which a deficiency was

located during the inspection. However, Pontis BMS practitioners are not confident about

the correctness of the user cost values used as default values: ($19.34/ hr for detour per

hour; $.25/Km for detour per kilometer; and an average of $37,600.00 per accident). The

Pontis documentation does not outline how these values were developed and from were

they came. According to one officer from the FHWA -Bridge Management Division, Mr.

Romack, the SHA members that participated in the development of Pontis offered these


values, which are the ones that normally are used at their agencies to evaluate transportation

projects. The first version of Pontis used an average accident cost default value of $ 14,000

per accident, which is based on California Department of Transportation (Caltrans) data for

1990 (Pontis 1993). This figure has increased to $17,900 as of 1995, (Pontis 1997a) and to

$37,600 as of 1998 Pontis (1998a). The sources from where the VOC and TTC default

values came from were not identified.

At the time of the Pontis development, the accuracy of the default values to be used

in the Pontis mathematical models was not a main concern of the Pontis developers. Each

SHA representative believed that the values developed by their agency were the right ones,

and for this reason it was provided a factor, weight (W), in which each practitioner adjusted

the default values. Based on this scenario a hypothesis for this research was created.

2.7 Statement of the Hypothesis

Current user costs parameters used as default values in Pontis BMS mathematical

models that were benchmarked from other areas are not suitable to evaluate user cost

benefits for the Florida bridge network. A specific and customized development of user

cost parameters is required for Pontis BMS application, considering local SHA policies.

All mathematical models of the BMS software AASHTOWareTM PontisTM will be

assumed to be appropriate with the exemption of the default values for travel time, vehicle

operating costs, and accident costs. The conceptual framework of BMS will be established

and used as criteria to evaluate the suitability of the user cost parameter to satisfy the needs

of the BMS framework applicable to Florida.

The analysis of the theoretical basis for each parameter is discussed in chapters

6,7,and 8 and the results were used as criteria to establish the development of the user cost


parameter. Each user cost parameter must satisfy BMS needs and the Florida DOT policies

while confirming that systematic bias does not exist when using Pontis and making

decisions for project selections for replacement or repair work and maintenance. The

degree of hierarchy between two available parameter selections will be in the following

order: local data --first choice; other state data--second choice; national data--third choice;

and international data--fourth choice.

2.8 Validation

The mechanics of the BMS Pontis software are not being questioned in this study.

What is being questioned is the quality of the user cost parameters used in the software.

Pontis software will be used to validate the quality of the user cost parameters that will

come from this investigation. This validation will be performed though a sensitivity

analysis where user cost data from this study will be used as new user cost default values for

the software. Another sensitivity analysis will be performed using the original user cost

default values, and then a comparative study will be performed to define the variability of

each set of data.


There is a scarcity of literature sources about BMS, probably due to the fact that

BMS is relatively new. The strategy used to collect data relative to BMS was to search the

literature under eight main entries in a cross-reference with bridges. The entries used were:

Accidents; User Costs; Economic Evaluation; BMS; Inspection; Maintenance and

Rehabilitation; Technical Issues; and Others. A form was created to record each entry with

the abstract of their contents, the importance for the project, comments and evaluation, and

references. During the first six months of the literature review process a total of 72 entries

led to 1,571 references. With this literature review start-up it became evident that only two

BMS systems are in the process of implementation in the USA, and one in the discussion

process. In the world only four BMS systems were located in the phase of discussion (three

in Europe and one in South America). Another fact learned was about the quality of the

abstracts. In the majority of the abstracts related to BMS, the authors overemphasized the

contents of their studies, promising solutions for a common problem in all BMS, that of

calculating the user costs in BMS environments. A majority of the BMS related studies only

mention that it is possible to evaluate the user costs. However, they do not display how, and

do not show the values of user costs parameters.

The abstract summary of each category entry provides an overview of the BMS

related issues leading to the selection of the three main issues of this research. The relevant

issues found under each category are outlined below.

3.1 Accidents

One of the entries classified as highly relevant was the Blincoe (1996) study, "The

Economic Costs of Motor Vehicle Crashes, 1994". This was the first study that presented

evidence to support the value of $2,854,500 for a life. For this research, the higher cost

values for fatality and injury we can use will contribute more to the economic justification

process of the BMS. North Carolina uses the value of $1,500,000 per life for fatality costs,

using the "willingness-to-pay" approach. Additionally, the Blincoe study supports the

highest cost value for life found in the literature. Blincoe points out that it can be even

higher than what he reported.

Kragh, Miller, Reinert (1986), stress the importance of evaluating social costs under

the classification of indirect accident costs. Under the two existing approaches to determine

accident costs (Human Capital and "Willingness-to-pay"), only the later considers psycho-

social costs. It represents 65% of total accident costs. The human capital approach used by

the NHTSA considers only 10% of the willingness-to-pay approach. Comparing the

numbers used by the National Safety Council (NSC) with the ones used by the NHTSA, it

was observed that NSTSA generates cost figures 2.5 times larger than NSC. The authors

mentioned above present a cost of $1.3 million per fatality in the "willingness-to-pay"

approach which is considered conservative. The Occupational Safety and Health

Administration (OSHA) calls for $5 million/life and the Office of Management Budget calls

for $1.5 million/life. A compromised value between the two offices is $2,000,000 per life.

A USDOT (1996) publication presents the number of motor vehicle occupants

killed and injured in the year 1996. A total of 41,907 people lost their lives in 1996 in motor

vehicle crashes, a 2.0% increase from 1995. The fatality rate per 100 million vehicles miles,

was equal to 1.7 with an average of 115 deaths each day (one every 13 minutes). This

publication shows the magnitude of the problem. The report points out that trucks account

for 12 % of all fatalities, and that accident costs are linked to inadequate truck management

in roadway networks. Cerelli (1996) presents the trends in crash fatalities/day for the period

1975-1995 based on National data. Weekends are the period with the highest fatality

average, Saturday having the greatest number. Table 3 shows the average number of

fatalities per each weekday.

Table 3. Average Number of Traffic Fatalities per Each Weekday-- 1975-1995

Sunday Monday Tuesday Wednesday Thursday Friday Saturday
150 100 100 100 110 110 190

The author claims that the Saturday fatality rate is very likely related to alcohol use.

Another relevant issue related to safety on bridges found in the literature is the issue of

"narrow" bridges whose diminished widths may increase the risk of single vehicle

collisions with roadside appurtenances such bridge ends, railings or approach guardrails as

well as collisions with other vehicles.

Brinkman and Mark (1986) used the bridge and roadway inventory from five states

totaling 11,880 bridges and 24,809 accidents during a three-year period. Bridge related

accidents were found to be approximately twice as likely to result in a fatality as a typical

accident. The same result was found in a North Carolina study. For undivided bridges, the

discriminate variables in order of importance are ADT (Average Daily Traffic), roadside

distraction, percent shoulder reduction, degree of bridge curvature, curb presence, bridge

length, degree of approach curvature and demarcation.

An older investigation made by Turner and Rowan (1982), in a sample of 24,000

accidents occurring between 1972-1979 on Alabama state-route highways, found that

roadway accidents rates increase near bridges. One-fourth of the traffic accidents

investigated occurred within 0.33 miles of a bridge. The authors observed that many bridge

accidents are apparently incompletely investigated, not properly identified, erroneously

recorded, misallocated, or ignored due to limited room for identifying information on the

accident report prepared by the police.

Zegeer and Council (1995) reports that bridge widening can reduce total bridge

crashes by as much 80% depending on the width before and after widening.

Rahim and Johnston (1993) report the accident rate relationship between bridge

accidents and roadway accidents for North Carolina, using a sample of 2,000 bridge related

accidents. The relationship for fatality, injury type A, injury type B, and injury type C are

found to be respectively: 2.0, 1.3, 1.05, and 0.87. The rate 2.0 means that bridge related

accidents are twice as severe as other roadway accidents. The predictor equation used by

the authors shows low values for the coefficient of multiple determination (R2 = 0.33, 0.34).

For this reason it is not recommended to use the equation generated by Rahim and

Johnston. They also found that approach roadway alignment is a parameter with no

significance for the accident rate. That is contradictory to the findings of Behnam and

Laguros (1973), and Brinkman and Mak (1986).

3.2 User Costs

Under this category, the Zaniewski, et al, (1982) study was the second

comprehensive study about Vehicle Operating Costs (VOC) found in the literature. This

study was financed by the World Bank and was performed in Brazil. The weakness of this

study is the transposition of the findings in Brazil to the USA assuming similarities between

the roadway networks. However, this study developed tables to update VOC costs against

the roadway grade and speed. Currently, the most updated study about VOC is the HERS--

Highway Economic Requirement System (1996), financed by the USDOT-FHWA, where

VOC values are developed based on an update of the Zaniewski tables. The evolution

observed of VOC studies is as follows.

The genesis of Vehicle Operating Costs (VOC) studies in the USA can be traced

back to the period just after the First World War with studies about fuel performance

conducted by Agg (1923). In the following years, under Agg management, the research

staff at the Iowa State Engineering Station, introduced new parameters in the VOC studies.

These included the effect of the roadway geometry on: VOC (Agg and Carter 1928); truck

operations (Winfrey 1933); tractive resistance; and road surface types (Paustian 1934).

One of the earliest surveys of VOC was reported by Moyer and Winfrey (1939),

who examined the fuel, oil, maintenance and tire costs of rural mail carriers. Moyer and

Tesdal (1945) complemented this study with the results from tire wear experiments. In the

1960's, researchers concentrated on the relationships between highway geometry, vehicle

performance and costs. Saal (1942) extended his experimental fuel consumption data using

survey information, while Claffey (1960) developed models and reported results on speed

and fuel experiments incorporating highway and vehicle characteristics.

By the mid-1960s, only fuel consumption could be predicted accurately. Between

1963 and 1969 the National Cooperative Highway Research Program (NCHRP) sponsored

Claffey and Associates (1971) to conduct a VOC research which resulted in the classic

NCHRP Report 111 (Running Costs for Motor Vehicles as Affected by Road Design and

Traffic) where the radioisotope technique was used to measure tire wear. This technique

was found to be unsatisfactory.

With the growing need for economic appraisal of highway transportation projects in

developing countries, the World Bank concluded that the VOC data developed in the USA

was not appropriate to be used in these appraisals. Funds were generated to conduct studies

in four areas: Brazil, India, Kenya and the Caribbean. The primary data collected in the

VOC study performed by Zaniewski, et, al, in Brazil was then used to update a VOC study

made in the USA. In June 1982, the Federal Highway Administration published a

comprehensive study about VOC performed by Zaniewski et al (1982) with the objective to

update the 1971 study of VOC performed by Claffey (1971). This was the last

comprehensive study conducted on VOC.

The Zaniewski study was performed with the objective to determine the VOC

relationship to roadway characteristics in order to determine the effect of pavement type and

condition on these costs. In order to determine the effect of pavement type and condition on

selected parameters, and develop an adjustment procedure for these performance parameters

as a function of the pavement and condition, five parameters were investigated: VOC,

running speed, fuel consumption, vehicle emissions, and accidents. The VOC included

consumption of fuel and oil, tire wear, vehicle maintenance and repair, and use-related

depreciation. A study made by Harrison et al. (1992) for truck operation costs, using a life

cycle cost approach, found a VOC value of $1.07 per mile for the Pennsylvania 1-80


In 1996 the Highway Economic Requirement Systems (HERS) sponsored by the

USDOT and the Federal Highway Administration published a set of VOCs to be applied to

pavement applications. These costs were estimated as a function of the average effective

speed, average grade and pavement serviceability rating (PSR), excess operating costs due

to speed change cycles and excess operating costs due to curves.

A new approach to evaluate VOC costs was found by Delucchi (1996). The

Delucchi study introduced the concept of social cost analysis into VOC costs dividing it into

two groups: non-monetary and monetary costs. The author states that costs of travel delay

imposed by others (which is the case of a bridge restriction) remain completely unpriced for

the responsible motor-vehicle user.

Another source of VOC was the American Truck Association. In 1988 they reported

a cost of $1.07 per mile where 79.26 cents/mile represented variable costs (ATA 1990). In

1992, the cost reported per mile was $1.20 (ATA, 1992). In 1996, the cost per mile reported

was $1.25 (ATA, 1996). In 1997 the cost per mile reported jumped to $1.92 (ATA-1997).

The main reason for the discrepancies in the VOC reported by ATA was due to the lack of

methodology to evaluate the VOC. Cost components were included or removed without any


For travel time costs, the most important source found in the literature was traced to

the studies by Miller (1996). Almost all relevant work on travel time uses Miller's

methodology. In the HERS (1996) study, the value of one-hour travel time by a vehicle was

developed using the Miller approach. They are listed on Table 4 using 1993 dollar.

Table 4. Travel Time Cost--Dollar per Hour-Base 1993--Miller Approach

Auto 4-tire truck 6-tire truck 3-4 Axle truck 4-Axle Comb. 5-Axle Comb
$11.22 $12.61 $23.69 $27.7 $30.09 $30.26

Software named MicroBencost was identified as a tool to evaluate travel time on

different roadway scenarios. The Tennessee DOT used it to evaluate the travel time value at

nine (9) different scenarios in the case of a closure of the Interstate 155 bridge crossing the

Mississippi River. The vehicle travel time cost in 1993 dollars per vehicle hour is listed in

Table 5. The basic default numbers used in the software are derived from the Zaniewski


Table 5. Travel Time Cost-Dollar per Hour-Base 1993--Zaniewski Approach

Auto 4-tire truck 6-tire truck 3-4 Axle truck 4-Axle Comb. 5-Axle Comb
$10.34 11.74 22.11 25.42 28.16 28.33

3.3 Economic Evaluation

A model used in traffic assessment was developed by Texas A&M University

named QUEWZ-92. It is more oriented to evaluate queue lines for work zones and the total

costs for delay. The default values used for cars and trucks in 1993 dollars are respectively:

$12.64 and $23.09 per hour (Krammes et al., 1993).

Farid, et al. (1994) developed a formula to estimate the annual user cost of an

existing bridge. This formula uses the proportion of vehicles involved in accidents due to

bridge deficiencies and the proportion of vehicles that need to detour, also due to bridge

deficiencies. The weakness of this mathematical formula is the assumption made for the

proportion of each vehicle type that will detour due to load and vertical clearance


limitations. Also, it is not clear how to find the proportion of vehicles that will be related to

an accident.

3.4 BMS

As mentioned before, only the BMS Pontis and Bridgit are being implemented.

However, it was observed that new BMS models are under development around the world.

Countries like Russia (Johnson et al., 1998), Poland (Vegosz and Wysokowski 1995) and

Hungary (Kolozsi 1995) are developing BMS models that do not consider user costs. In

Brazil, an Engineering Company named MCN Engenharia Ltda is developing a BMS

named SIMGO (NHI-1999).

A study made at North Carolina State University by Chen and Johnston in 1987,

resulted in a BMS analysis program which considers owner costs and user costs to

determine the optimum improvement action and time for each individual bridge in a system

under various levels of service. A sample of 17,000 bridges in North Carolina were

analyzed using ADT data from 1974-1984. The coefficients used to measure proportions of

vehicles that incur accidents due to deck width, alignment and vertical clearance

deficiencies are assumed to be constant. The proportion of vehicles that detoured due to

load and vertical clearance is also assumed to be constant. These assumptions are not

consistent with the 4.6 % yearly ADT increase observed in the North Carolina Interstate

System. The vehicle classification distribution is adjusted using data from six different

studies. The truck weight distribution is adjusted based on a study made by the FHWA in

1985 about the bridge structure-loading spectrum. The truck speed at a detour is evaluated

by dividing the driver salary and benefits at union scale of$13.35/hr, by the owner-operator

driver salary of 0.311/mile, resulting in an average speed of 40mph. This approach is

mathematically correct, however, conceptually wrong once the average speed changes

according to the salary variation between the two categories. VOC value for cars are based

on the Internal Revenue Service (IRS) allowance of 20 cents per mile plus 15 cents for

labor. For trucks the VOC value is derived from the data collected by the United States

Department of Agriculture (USDA) at the value of $1.15 per mile.

The BMS study developed by Yongtae and Sinha (1997) uses a new methodology to

estimate user costs. It was developed to address the needs of the Indiana DOT. It considers

detour costs for load and vertical clearance bridge restrictions, and travel time due to load,

vertical clearance and width bridge restrictions. The weakness of this methodology is the

absence of the user costs related to bridge related accidents. One of the main features of this

work is the calculation of traffic proportion used to evaluate user costs. It has more

flexibility than the North Carolina BMS model.

The VOC cost is calculated based on the update of Zaniewski's work, which

produces extremely low values. The study has an innovative approach to evaluate user costs

due to narrow width.

A brief from the NCHRP that lists all projects in progress, indicates a new BMS

model named StratBENCOST which was to be released in 1997. However, this product

could not be found in the literature (Lewis 1992a, 1992b). Keating and Turner (1994)

report a new BMS software named ALBBRIDGE that is the BMS from North Carolina

customized from the Alabama DOT.

3.5 Remaining Relevant Entries

All sources located under the entries Inspection, Maintenance and Rehabilitation,

Technical Issues and Others were of no importance to the objective of this research.


BMS is an analytical tool that empowers decision-makers with the ability to make

effective decisions for optimal use of available resources in the bridge business. This

chapter has the objective to discuss the BMS basic components, their capabilities, and their

decision management support output to accent the importance of user costs into the system.

Shirole et. al. (1994) states that in any BMS, there are only three basic components: data,

data analysis, and decision support. They are expanded in Figure 2. The BMS of focus in

this dissertation is the AASHTOWareTM PontisM.

Input- Processing Output
(Data) (Data analysis) (Decision support)
Data that are necessary The analysis routine to The results of the analysis
for the decision process which the data is routines that will assist the
subjected user in making decisions

Figure 2. Data Treatment Flowchart

Pontis is the Latin world meaning "pertaining to bridges." According to Thompson

(1994), the overall BMS structure involves the input of condition prediction and cost

models to the database. The database is then used to optimize preservation and determine

improvement strategies. Along with additional supporting data, from the database, these are

integrated to form the program. Figure 3 shows this overall BMS structure.


According to AASHTO (1993) the database needs to contain inventory, inspection

and appraisal data as well as complete historical information and codes indicating the dates

and nature of detailed, special and supplemental inspections. The BMS software needs a

capability to edit and update the database as appropriate. The database includes many of the

data items in the NBI database, but also needs to include other items, especially a more

detailed inventory, and condition data on the elements of each structure.

Figure 3. Overall BMS Structure

4.1 Pontis Database

The main objective of the BMS database is to provide identification of needs,

accurate economic forecasts, prediction of physical condition, and continuous

improvement. The Pontis approach to generating a database is compatible with the

modeling operations, and includes: the definition of each bridge by its individual element;


the establishment of particular classification of the condition of a bridge element. The term

"feasible action" means "a defined BMS preservation activity unique to an element's

material composition and condition state". Preservation activities are referred to as all

actions taken to offset the deterioration caused by traffic, weather, or any chemical or

physical process.

The selection and definition of structural elements is a central issue in preparing a

bridge database for successful modeling. Pontis adopted the AASHTO proposed list of

standard elements referred to as Commonly Recognized Elements (CoRe ,1996). There are

a total of 98 structural elements described in the CoRe Bridge Inspector's Field Guide. The

condition state of each element receives a classification in the range 1 to 4, and 1 to 5

integer scales where 1 indicates excellent. Pontis has about 140 different types of bridge

elements defined. Federal funding apportionment to States is still based on sufficiency

ratings derived from the NBI Structure and Inventory Appraisal (SI&A) data. The primary

NBI data item for prevention of failure is still the condition rating on a scale of 0 to 9, where

9 indicates excellent condition. The NBI requires condition ratings for only three major

structural components: deck; superstructure; and substructure. To merge Pontis and NBI

data, a standard conversion program, commonly refereed as the NBI translator or BMSNBI,

was developed by the University of Colorado (Heam et al, 1997) to compute NBI ratings

from Pontis.

The element listing includes a description, a definition, condition state language,

and a unit of measurement for each element. The element descriptions consider material

composition and, where applied, the presence of protective systems (NHI, 1996)

Each element is also categorized in one environment. Pontis defines four types of

environments to which each element can be exposed: benign, low, moderate and severe.

The definition of each environment type is listed on Table 6 (NHI, 1996).

The efficacy of element-level bridge management systems evaluation, such as used

in Pontis, has been confirmed by Hearn and Renn (1999) for eight highway bridges in


Table 6. Pontis Definition of Bridge Element's Environment

Environmental factors and operating practices are not likely to
Benign significantly change the condition of the element over time or their
effects have been mitigated by past non-maintenance actions or the
presence of highly effective protection systems. Example: desert bridges
Environmental factors and operating practices do not adversely influence
Low the condition of the element or their effects are substantially lessened by
the application of effective protective systems. Example: reinforced
concrete bridge in a warm climate
Environmental factors and operating practices are considered to be
Moderate typical for the Agency and any change in the condition of an element is
likely to be normal. Example: Reinforced concrete bridge in the north
with average use of road salt
Environmental factors and operating practices contribute to the rapid
Severe decline in the condition of an element. Protective systems to negate
environmental effects are not in place or are ineffective. Example: bridge
in brackish water, bridge exposed to excessive deicing chemicals.
Source: National Highway Institute, 1997- (NHI, 1976)

4.1.1 Needs Identification

The identification of needs is classified in two classes: functional and preservation

needs, that are also named as MR&R needs. In order for Pontis to identify functional needs,

it requires data on widths and clearances, load capacity, traffic, and accidents. To identify

preservation needs, it requires evidence of deterioration, such as the condition to describe

the physical symptoms of deterioration which can be visually observed by inspection.

4.1.2 Cost Models

Cost models are related to accurate economic forecasts, economic inputs, including

unit costs, user costs, and the transportation service attributes which determine user costs,

such as traffic and accident rates.

4.1.3 Prediction Conditions

To predict the physical condition of bridges, the BMS needs deterioration rates,

which are developed by making use of all available past inspections as well as expert

judgment. In order to distinguish between the effects of deterioration and the effects of

maintenance, the BMS needs to know what past maintenance was done on each bridge.

4.1.4 Continuous Improvement

Continuous improvement prediction models can be continuously improved over

time if there are methods to compare the predictions with what subsequently actually

happened. The cycle between models and outcome is shown in Figure 4. The database

structure contains four sets of basic information about each bridge, describing the bridge

itself, and each of its elements, roadways and spans or structural units. They are listed on

Table 7 (NHI,1996).

4.2 Prediction Models

There are two kinds of models to predict future bridge conditions: deterioration and

action effectiveness. Deterioration models predict what will happen to the bridge if no

maintenance or improvement is performed. It tells how quickly the bridge element will

reach a condition level where some corrective action might be warranted. Action

effectiveness models tell how the condition is changed if a maintenance or corrective action

is actually performed.

Table 7. Basic Inventory Information for Each Bridge (NHI, 1996)

About Structure
About the Bridge About Elements About Roadways Units
Identification Material Identification Design
Age/Service Type Traffic Material
Geometry Environment Ruts
Clearances Quantity Dimensions
Load Rating

Deterioration models can be divided into two groups according to how they handle

uncertainty: Probabilistic (subject to uncertainty) and deterministic (known for certain).

Pontis applies the probabilistic Markov Chain process, which means that they divide time

into discrete, equal periods; forecast next period condition, without regard to earlier

conditions; and perform this prediction by use of transition probabilities among the

condition states. The strength of the Markovian models used in BMS is that it is simple to

use, (requires low data collection, it is easy to update from historical data, and has the ability

to use an inexpensive visual inspection procedure to collect the required data). The

weaknesses are that it is not precise and can not model latent properties. Kleywegt and

Sinha (1994) state that the developers of Pontis suggested an approach to overcome the

Markovian approach weaknesses by using the subjective judgment of bridge maintenance

experts to obtain estimates of transition probabilities. As data are collected through regular

inspections, these initial estimates are updated and improved. Developers of the BMS

software named BRIDGIT (Lipkus, 1994) and Mansino and Pardi (1999) also employ the

Markov Chain Process to calculate the transitional rates for each condition state of a bridge


4.3 Cost Models

Tuner and Richardson (1994) state that BMS are driven by costs. Everything

eventually is compared in terms of costs. Costs are the common denominator in bridge

management systems. The degree of difficulty to estimate bridge related cost was compared

by Son and Sinha (1994) as the same degree to produce deterioration rate estimates for

groups of bridges.

The aim of BMS is to help decision-makers make cost-effective decisions. For this

reason cost models are also an important set of inputs to BMS. There are many kinds of

cost models, each sensitive to a different set of factors. Pontis uses maintenance, repair and

rehabilitation (MR&R) direct costs, functional improvement and replacement direct costs,

indirect costs, and user costs.

MR&R costs primarily depend on structural characteristics of the bridge and the

extent of deterioration which is to be corrected. They are specific to bridge elements

expressed in dollars per physical unit, specific to the type of action and may depend on

condition location and element properties (NHI, 1996).

Functional improvement and replacement costs are normally provided by the SHA.

Those are the costs to widen or strengthen a bridge. Indirect costs are associated with the

decision to perform any work at all on a bridge. They include design, traffic control, land,

environmental mitigation, demolition, and administration (NHI, 1996).

User costs measure the effect of substandard bridges on road users. Although the

inconvenience to each vehicle is small, the numbers add up quickly. User costs are related

to weight limits and clearances in the areas of truck height/weight distributions, detour

lengths/times, truck driver labor cost per hour, vehicle operating cost per kilometer, bridge

related accidents, traffic delay costs in work zones, and environmental costs. User cost

models require some special inputs of their own. Many agencies maintain these data in their

planning departments because they are used for many other purposes. Accident rate

estimation is the most difficult because it is necessary to identify the accidents which are

specifically associated with bridges.

4.3.1 Preservation Optimization Models

The Pontis preservation model is in reality a comprehensive set of models to

optimize the structure preservation policy, to recommend actions, and to set priorities.

Figure 5 is a simplified schematic diagram of how data flows in preservation modeling. The

main steps of the process are:

Step #1: Development of a probabilistic deterioration model (Pontis, 1997a).

Step #2: Update of the probabilistic deterioration model (Pontis, 1997a).

Step #3: Development of a set of unit costs for preservation actions (Pontis,


Step #4: Update of the set of unit costs based on experience (Pontis, 1997).

Step #5: Optimization model that combines the consideration of condition

(Pontis, 1997a) change, action effectiveness, and action cost to

determine the most cost-effective long-term policies.

The guiding principle of the preservation optimization models is to find the long-

term policy for each element in each environment which minimizes the long-term

maintenance funding requirements while keeping the element out of risk of failure.

Another important concept of optimization is steady state. Bridges stay in service

for a very long time, and the overall objective of providing transportation connectivity

means that there is never a point in the future where network-wide preservation policies

might have to be drastically changed or eliminated.

Figure 5. Data Flows in Preservation Modeling

The preservation optimization process is one of the alternative results of the bridge

program. It includes the phases of inspection, policies, programming and project.

The inspection objective is to detect conditions threatening the structural integrity.

Expert Cost
Elicitation (3)

Elicitation (1)

Past bridge collapses due to failure of fracture-critical members, failure of underwater

members, and scour of foundations soils, have attracted attention to these failure modes.

There are three basic requirements which BMS policy should meet. They are: experience--

the policy and its predicted impacts should be consistent with reasonable expectations of

how bridges will deteriorate and how much proposed actions will cost; feasibility--

recommended actions should be feasible for the agency to accomplish; sustainability--the

policy should be sustainable over a long period of time, because the bridges will be in

service for a long period of time. Automated support for defined objectives, and ways to

quantify them, guide policy decisions. There are four network level objectives: minimize

agency costs; minimize user costs; maximize service life; and maximize progress toward

optimal conditions. There are also constraints: budget and level of service.

With objective quantitative policies and criteria, it becomes possible to generate

project lists automatically as a way to approximate quickly the composition of an objective,

budget-constrained program. Pontis uses the network level policies and standards to

generate projects. A project is a collection of preservation and functional improvement

actions on a single bridge. The basic decision criteria is to accomplish as much as possible

with the current budget, minimize the long term cost of keeping each bridge in service,

minimize inconvenience to road users, and act in a consistent manner across all projects.

The focus should be to find an optimal balance and combination of all alternatives.

Preservation policies are developed with the help of a model which specifies action

selection rules, unit costs and calculated benefits. An action selection policy, for example,

aids in the choice of project level action by specifying the action which gives the lowest

long-term costs, based on condition and environment. Network-level unit costs provide an

initial rough estimate of project costs as an interim step until a more detailed project cost

estimation can be performed. When a network-level policy analysis tool, such as Pontis,

calculates life cycle costs, it can compare the annualized long-term cost if the recommended

action is taken, to the annualized long term cost if no action is taken. The difference can be

expressed as the benefit of the action, and can be allocated to the project on a unit basis.

Functional improvement policies have a similar relationship to the definition of

projects. Standards typically determine the type of improvement action to be taken. Design

standards are generally higher than level-of-service standards, because a bridge, which is

only slightly below design standards, probably would not merit the expense of improving it.

In Pontis, a level of service standards is typically determined by an interactive process

where the budget requirements of the standard are determined, and then the standard is

adjusted so its budget requirements more closely match likely funding availability. Initial

cost estimation is typically handled in the same way as with preservation projects, and

project benefits are typically expressed as savings in user costs.

According to ASSHTO (1993), preservation actions (MR&R actions) should be

evaluated based on their necessity to keep a bridge open and serviceable to users, and a

MR&R program should be formulated to minimize the agency cost of maintaining a

standard. Improvement actions on the other hand, should be evaluated on the basis of

potential user cost savings in travel time, vehicle operating costs and accidents.

Results of the optimization analysis include the long-term percent of each element

in each condition state, the long-term annual cost of each action, and the low cost

recommended action.

4.3.2 Improvement Optimization Models

The ASSTHO BMS guidelines (ASSHTO, 1993) establish a difference between

preservation and improvement actions. Pontis follows the same criteria. Improvement

actions considered by the Pontis improvement models include widening, raising, and

strengthening. Functional needs are determined primarily by design and LOS standards.

Standards are included for lane and shoulder widths, vertical clearances, and load limits.

The original source of level-of-service standards is the June 4, 1991-FHWA Notice of

Proposed Rule-Making on level-of-service standards (F.R., 1991). The design standards are

based on Caltrans practice, which in turn is based on the AASHTO Geometric Design

Policy "Green Book"(AASHTO, 1994b). The bases to perform the decisions in the

preservation model are listed in table 8.

Table 8. Bases for the Need and Benefit Calculations on Improvement Projects

Widening Need Based on comparison of current roadway widths to the width
standards set in the Pontis policy matrix
Widening Benefit Based on estimated reductions in accident costs
Strengthening Based in the structure level operating rating and design load, and
Need LOS standards in the network level policy matrix
Strengthening Based on reduced truck traffic detour costs. These have both a time
Benefit and a distance component.
Raising Need Based on comparison of current roadway vertical clearance to the
clearance standards set in the policy matrix
Raising Benefits Based on reduced traffic detour costs. These have both a time and a
distance component.
Replacement Based on presence of vertical clearance or width deficiencies
Replacement Based on reductions in accident and truck detour costs
Reference: Pontis User Manual (Pontis, 1997a)

4.4 Program Integration Model

The integrated project programming model which selects the most cost effective set

of structure projects is described by Pontis User's Manual (Pontis, 1997a) in the following


"The Pontis project programming model performs a simulation of structure
condition change and traffic growth for up to a 30-year time horizon, and selects the
most cost-effective set of projects which meet the constraints of a user-defined set of
budget limits. The procedure integrates the results of the preservation optimization

and the functional improvement models, developing and evaluating project
alternatives, which incorporate both preservation and functional actions. The
simulation is designed to develop a set of candidate projects to meet the identified
needs, but it will also accommodate a set of projects, which have been manually
defined by the user."

This module basically defines and selects a set of projects which optimize expenditures

with established budget constraints for a given scenario. The determination of needs

changes from year to year over a specific selected planning horizon. The generation of

project alternatives includes preservation only, preservation and functional improvement,

and replacement.

4.5 Pontis Results

Pontis BMS produces more than 60 reports and it has the capability to generate a

customized report. Appendix A shows a list of the reports available from Pontis, covering

the areas of inspection, preservation results, network-level programs, project-level

programs and historical projects.

From the perspective of user cost models the first point is to define costs and

benefits. Agencies use the terms "costs" and "benefits" in many different ways for many

different purposes. In BMS, the definitions are chosen so that they accurately reflect what is

at stake in bridge management decisions. Seven different types of cost were found in the

agency vocabulary. They are direct costs, avoidable costs, budgetary requirements, "full"

costs, first costs, life-cycle costs and social costs. Seven different types of "benefits"

definitions were found. They are avoided costs, budgetary reductions, social benefits,

payback period, internal rate of return, net present value and extended "service life." Some

of these terms will be discussed in the following chapters. Pontis adopts for "cost" and

"benefits" the following definitions:


Costs are defined as budgetary requirements, but not as additions to the budget.

Usually when a cost is incurred, it is treated as a reduction of funds available to other

projects. Direct and indirect budgetary first costs are incurred. "Benefits" are defined as

avoided costs to the agency and to road users, minus the first costs. Thus they may be

referred to as "net benefits". As a result of these definitions, any benefit/cost ratio greater

than zero represents an attractive investment. These concepts will be expanded in the

following chapters.


This chapter discusses user cost mathematical models used to quantify the annual

benefits resulting from widening, raising, strengthening, replacing, and detouring a deficient

bridge, which are evaluated with respect to the user cost default parameters. Florida default

parameters that are related to the benefits listed above are also presented. This information

is extracted from a FDOT user cost study (Thompson, Soares, Najafi, Choung, 1998). One

illustrative example is presented showing the savings in accident costs (SAC), vehicle

operating costs (SVOC) and in travel time (STT), using the original Pontis cost default


5.1 .Benefits and User Costs

One of the Pontis requirements is that benefits of functional improvements are

assessed in terms of user cost savings (Glolabi et al., 1992). Benefits are the value of taking

actions to address the functional improvement needs. Standard functional improvement

actions include widening, raising, and replacing a bridge. The benefits of addressing

functional improvement needs are calculated by user-modifiable formulas which are based

on user cost (travel time, accident) reductions gained from eliminating detours and

improving safety (Pontis, 1997a). A functional need is related to the structure's ability to

accommodate user demands. Examples of functional needs include the need to increase

clearances, increase widths, or raise weight limits in order to serve more traffic and/or

improve safety.

5.2 FDOT Default Values Policy

The user cost model is used only for functional improvement and replacement

projects. There are 64 data items that are included in the Pontis user cost model. Seventeen

of those items are related to bridge data variables, and 47 items are related to

mathematical model parameters.

The 47 mathematical model variables resulted from SHA policy. They are defined

to input to seven different models: user costs; traffic; widening; raising; strengthening;

replacement; and detour. The model parameters are defined by the FDOT. Appendix B

presents the widening, raising, replacement, strengthening, detours, and bridge data

adopted by the FDOT to be used in Pontis. The Pontis cost matrix that relate to user costs

as defined by FDOT policies is shown in Table 9.

Table 9. Cost Matrix Values Adopted by FDOT

Model Variable Model Florida Default
Detour cost per hour detour $19.34
Detour cost per kilometer detour $0.25
Cost per Accident widening $14,247*
Weight given to user cost user cost 100 %
Source : Thompson, Soares, Najafi, Choung, (1998)
At the time of the report FDOT was using this value. Current value is $37,600

Bridge data variables are bridge functional class, detour distance, detour speed,

roadway functional class, roadway speed, truck fraction, vertical clearance, operating rating,

future volume, future volume year, traffic volume, traffic volume year, bridge length,

approach alignment rating, approach road width, number of lanes, and roadway width. The

values adopted by FDOT are listed in Appendix B, Table B6.

A user cost model is fed with the weight given to user cost that is expressed in a

percent index. In Florida's case it is 100%. The traffic model is fed with the default traffic

growth period, normally 30 years. The widening model is fed with cost per accident, high

approach alignment rating, low approach alignment rating, approach width factor, design

lane width, design shoulder width, short bridge threshold and two regression constants. The

raising model is fed with height detour default, and 10 different height detour points. The

strengthening model is fed with four weight detour points. The replacement model is fed

with five height eligibility points. The detour model is fed with detour cost per hour, detour

cost per kilometer, detour speed factor, default truck percent and 12 levels of default road


There are three different tables where these data values are located. They are: cost

matrix; policy matrix; and improvement model tables.

The Cost matrix items include agency functional improvement unit costs, user

detour and accident costs, agency costs for special improvements such as seismic retrofit

and scour protection, and associated agency and user benefits for making special

improvements. For this dissertation, the focus is on the four user cost models defined as


Detour per hour--user costs per hour of additional travel time incurred by vehicles

that would normally use a structure but cannot due to clearance or load restrictions. This is

used to calculate user cost reductions associated with raising or strengthening a structure

(Pontis 1997a).

Detour per kilometer-user costs per kilometer of additional travel distance incurred

by vehicles that would normally use a structure but cannot due to clearance or load

restrictions. This is used to calculate user costs reductions associated with raising or

strengthening a structure (Pontis 1997a).

Average per accident--the user costs per accident. This is used to calculate the

accident reduction benefits of widening or replacing a deficient structure (Pontis 1997a).

Weight--determines the relative impact ( a percent) of user costs to actual costs in

the benefit-cost calculations. If the user cost weight is 100, user costs are treated on a par

with agency costs. If the user cost weight is 50, user costs would be cut in half (Pontis


The policy matrix contains standards when different types of improvement actions

should be applied. These standards can vary for different combinations of ADT class,

functional class, structure funding responsibility, and NHS status. The improvement model

parameter table contains improvement benefit model parameters defined for each

improvement matrix in the programming module.

5.3 User Cost Models

When a deficient NBI approach alignment or roadway width exists on a bridge, road

users are theoretically subjected to a higher accident risk. To evaluate a functional

improvement or replacement which corrects the deficiency, the user cost model predicts a

reduction in accident risk which then is multiplied by an accident cost to yield a user cost

saving. When a bridge has a substandard vertical clearance or load capacity certain trucks

are unable to go on or under the bridge, and must detour, thus incurring higher labor costs

and vehicle operating costs. The user cost model estimates the volume of detoured traffic

and the resulting user costs which would be avoided if the deficiency were corrected. The

total user benefit of the functional needs in a project is therefore:

User benefit B,= W, / 100 x V, ( BW, + BR, + BSr) (5)

W, is the weight given to user cost benefits, in percent (Pontis cost
V, is the forecasted average daily traffic volume for the program year
being analyzed.
BW, is the annual benefit of widening per unit average daily traffic
(calculated below)
BR, is the annual benefit of raising per unit average daily traffic
(calculated below)
BS, is the annual benefit of strengthening per unit average daily traffic
(calculated below)

In the notation for all equations, subscripts indicate either the level of resolution of the

variable, or the entity which the variable describes. These are defined as follows:

b indicates a bridge attribute (corresponds to bridge or inspection event table)
indicates a roadway attribute (corresponds a roadway table)
c indicates a cost matrix parameter (linked to the bridge table)
p indicates a policy matrix parameter (linked to the bridge table)
y indicates a program year within the planning horizon (Thompson et al. 1998).

Variables without a subscript are systemwide parameters. Approach alignment

rating is the only attribute of this type. When a bridge-level attribute is taken from the

inspection event table, it is taken from the most recent inspection for the bridge.

5.4 Benefit of Widening

Pontis estimates the user benefit of widening as the savings in accident costs. The

method for estimating accident user costs in Pontis is derived from the North Carolina BMS

using the following formula:

Benefit of widening BW, = CA ( R, R',) (6)


CA, is the average cost per accident (Pontis cost matrix)
R, is an estimate of the current annual accident risk per vehicle (calculated
R', is an estimate of the current annual accident risk per vehicle after
improvement (calculated below) (Thompson et al. 1998).


This result is calculated only for roadways on a bridge. It is zero for roadways under

a bridge. It is also set to zero if R, < R',. The parameters R and R' can, in principle, be

estimated from actual accident studies. However, no such studies were found in the

literature or from the questionnaire survey. The North Carolina system offers an

approximate way to estimate R based on bridge attributes as follows:

Current accident risk:

R,= 365 x 200 x (3.2808W,)65 [1 + 0.5 (9-Ab) / 7] (7)


W, is the roadway width (curb to curb) in meters (Pontis roadway table, NBI
item 51)
Ab is the approach alignment rating (typically 2 to9, Pontis inspection event
table NBI item 72)

If the approach alignment rating is missing, it is taken as zero. It would be more

appropriate to take it as nine so it does not add to the accident risk. If roadway width is less

then zero, it is treated as zero. Some of the numeric constants in this formula are user-

modifiable in Pontis in the improvement model parameter table. They are defined as


365 is the number of days in a year
200 is a regression constant
3.28084 is the constant Pontis uses to convert from meters to feet
6.5 is regression constant
0.5 is a model specification constant
9 is the highest approach alignment rating
7 is the range of allowed approach alignment ratings (Thompson et al. 1998).

The 200 and 6.5 are regression constants derived from the North Carolina study, so

they should be modified only if another statistical analysis of accident data is conducted.

The 0.5 constant arose because of the practice in North Carolina of assigning only even

numbers for approach alignment ratings. It is not important for the model framework, but


must be used with North Carolina regression constants. The final two constants are artifacts

of the NBI approach alignment scale, which range from 2 to 9.

The formula for accident risk after improvement is similar to (7), but depends on

the width of the improved roadway.

Improved accident risk:

R', = 365 x 200 x (3.2808W',)-65 [ 1 + 0.5 (9-Ab) / 7 ] (8)


W', is the roadway width (curb to curb) in meters (Pontis roadway table, NBI item 51)
Ab is the approach alignment rating (typically 2-9, Pontis inspection event table
NBI item 72) (Thompson et al. 1998).

5.5 Benefits of Raising

Pontis calculates the vehicle operating cost and travel time cost associated with

traffic on a detour route, and assumes that this entire cost is saved if a functional

improvement is made. Only trucks are assumed to be affected. Raising is considered only

for roadways under the structure.

Benefit of raising is: BR, = 365 x DC, x PT, / 100 x P, /100 (9)


DC, is the detour cost per truck for this roadway (calculated below)
PT, is the percentage of the traffic stream occupied by trucks (Pontis roadway
table, NBI item 109)
PH, is the percentage of trucks detoured by the bridge.

If the truck percentage is missing or zero, it is given the value of the improvement

model parameter default truck percent whose default value is 5 percent (Thompson et al.


5.6 Benefit of Strengthening

Pontis calculates the vehicle operating costs and travel time costs associated

with traffic on a detour route, and assumes that this entire cost is saved if a functional

improvement is undertaken. Only trucks are assumed to be affected. Strengthening is

considered only for roadways on top of a structure.

Benefit of strengthening is: BS, = 365 x DC, x PT, / 100 x PW, /100 (10)


DC, is the detour cost per truck for this roadway (calculated below)
PT, is the percentage of the traffic stream occupied by trucks (Pontis roadway
table, NBI item 109)
PWb is the percentage of trucks detoured by the bridge

If the truck percentage is missing or zero, it is given the value of the improvement

model parameter default truck percent, whose default value is 5 percent (Thompson et al.


It is possible that some fraction of trucks exceeds the operating rating, but ignores

any posted signs. Also, many states post bridges at levels different from the operating rating.

The model makes assumptions about these factors since it describes only the percentage of

trucks which are actually detoured at each operational rating level.

5.7 Benefits of Replacement

The user costs model for replacement benefits is very similar to the combined effect

of all of the separate functional improvements. An analysis of the source code reveals just a

few differences as discussed in this section. When a bridge is replaced, Pontis recognizes

the benefits of widening for all roadways on and under the bridge. All roadways are

assumed to have the approach alignment rating of the bridge before the project, and all are

assumed to have an approach alignment rating of 9 after the project.

Pontis assumes that bridge replacement eliminates all operational rating

deficiencies. As a result, the project benefit includes the benefit of strengthening, calculated

in the same way as described above in equation (10).

The replacement benefit model for height- related detours in Pontis is formulated to

allow for the possibility that, when both height and weight restrictions exist, certain trucks

may be affected by both restrictions.

Replacement height benefit = BR,

BR, = 365 x DC, x PT, / 100 x [(1-PWb /100) x PG /100 x P, /100 ] (11)


DC, is the detour cost per truck for this roadway (calculated below)
PT, is the percentage of the traffic stream occupied by trucks (Pontis roadway
table, NBI item 109)
PWb is the percentage of trucks which are detoured by the bridge due to weight
(Pontis roadway table, NBI item 109)
PGb is the percentage of those trucks not detoured by the weight limit, which are
potentially subjected to height restrictions.

There is a subtle logical inconsistency in the use of PH, in the raising and

replacement models. In the raising model, PH, is the percentage of the entire truck traffic

stream which is detoured since the percentage detoured by weight restrictions is zero. In the

replacement model, on the other hand, PH, is the percentage of only the lighter-weight duals

and tractor-trailers. The (1-PWb) term restricts PH, to lighter-weight vehicles, and the PGb

term restricts PHr to only duals and tractor-trailers (Thompson et al. 1998).

Part of this inconsistency can be removed by setting all the percentages in the PGb

model to 100 so the definition of PH,, is not limited to duals and tractor-trailers. There is no

easy way, however, to remove the effect of (i-PWb) (Thompson et al. 1998). Considering

the Pontis user community as a whole, it would be worthwhile to consider eliminating the

PGb factor and simplifying the definition of PH to conform to its usage in the strengthening

model. This could cause some minor double counting of benefits in cases where both

clearance and weight restrictions exist on the roadway on top of the bridge, but the number

of cases where this is a problem is likely to be small in most states. The benefit of the

change would be to make the user cost model smaller, more consistent, and more

understandable (Thompson et al. 1998).

5.8 Detour Cost

Each time a truck is detoured, it experiences vehicle operating costs associated with

the added detour distance and travel time costs associated with the added detour time.

Pontis uses a model of these factors for raising, strengthening, and replacement.

Detour cost per truck DC, = CV, x D, + CT, x (D, / DS,) (12)


CV, is the average vehicle operating costs per km of detour (Pontis cost matrix)
CTc is the average travel time cost per hour of detour (Pontis cost matrix)
D, is the detour distance for the roadway in km (pontis roadway table, NBI
item 19)
DS, is the speed on the detour route, kph, (Ponts roadway table)

Since detour speed is not in the NBI data item, many SHA lack this information.

When missing, Pontis estimates the detour speed from the roadway speed (Pontis roadway

table) using the improvement model parameter DetspeedFactor. The default value of this

factor is 80 percent. Since roadway speed is not in the NBI item, Pontis has a set of default

speed values, DefaultRoadspeed FCnn, where nn is the roadway functional class in the

improvement model parameters table. Since these defaults are very rough, it is better to

collect the actual detour speed or at least the bridge roadway speed, if possible (Thompson

et al. 1998).

5.9 Application Example

To clarify how the Pontis user cost mathematical model is integrated into the BMS,

one example of one deficient bridge is presented. This example is an adaptation of the

Blundell (1997) technical notes for a two lane concrete arch bridge.

One two lane reinforced concrete arch bridge has a deficiency that forces 46% of the

truck traffic to detour 38.6 Km. The actual Average Daily Traffic (ADT) is 7,166 and the

percent trucks composition is 14%. The BMS analysis result indicates a need to spend $ 1,

850,000 to replace the bridge as the best option to overcome the bridge deficiency. What is

the benefit cost ratio for this action assuming the use of the Pontis default values for user


5.9.1 Savings in Accident Cost (SAC)

SAC = 365 x V (R- R') Ca (13)


V= 7,166 (Average Daily Traffic)
Ca = 37,600 (Average Accident Costs)
R= 0.000035 (current annual accident risk)
R'= 0.0000028 (Current annual accident risk after improvement), explained below

The values for the current accident risk for the current year is R= 0.000035 (this value

can be taken from project simulation log files for this bridge). The after improvement

widening accident risk is listed by Pontis as R'= 0.0000028 (this value is evaluated by the

BMS using the new values of the roadway width Wr, and the new approach alignment rate,


The difference between the two accident rates is the reduction in accidents.

R-R' = 0.000032

Since the values of R are adjusted for year, the total SAC will be:

SAC = 7,166 x 0.000032 x 37,600

SAC = $ 8,622

5.9.2 Savings in Vehicle Operating Costs (SVOV)

SVOC = 365 x VD x C, x D (14)


VD = 7,166 x 0.14 x 0.46 = 461 (trucks detoured per day)
Cv = 0.25 (average vehicle operating costs per kilometer)
D = 38.62 (detour distance in kilometers)

The percent of trucks detoured is determined by the user settings in the

improvement module. For this particular bridge the percent of trucks detoured was 46%

(0.46). The equipment unit cost (Vehicle Operating Costs to overcome the detour length) is:

38.62 x 0.25 = $9.66/truck.

The number of trucks detoured per year is then trucks detoured per day times 365

days: 461 x 365 = 168,265 trucks detoured per year. Then saving in Vehicle Operating

Costs will be:

SVOC = 365 x 461 x 0.25 x 38.62

SVOC = $1, 623,757

5.9.3 Savings in Travel Time Costs (STTC)

STTC = 365 x VD x Ct x D/S (15)


VD = 461 (number of trucks detoured per day)
C, = 19.34 (average travel time cost per hour of detour)
D = 38.62 (detour distance in kilometers)
S = 70.24 (speed on the detour route, km/hr) explained below


This step is to include the labor cost for the drivers making the detour. The labor cost is

the detour distance divided by the detour speed (in this case Pontis uses as a default value of

80% of the posted road speed, that is 70.24 Km/hr) times the cost of labor per hour.

Detour length (38.62 Km)/ detour road speed (70.24 Km/hr) $19.34/hr labor cost =

$10.62/truck. Then the labor unit cost per truck is $10.62/truck. The resulted STTC is:

STTC= 365 x 461 x 19.34 x 38.62/70.24

STTC= $1,789,279

5.9.4 Total User Cost Benefits (UCB)

UCB = SAC + SVOC + STTC (16)

UCB = 8,622 + 1,623,757 +1,789,279

UBC = 3,421,658

5.9.5 Benefit Cost Ratio

Total User Cost benefits/ Agency Costs

3,421,658 / 1,850,000 = 1.85

The BMS follows this methodology for all deficient bridges found in the network,

and then ranks all the benefit cost ratios. The bridge that receives the highest benefit cost

ratio generates the highest savings for the user when its deficiency is fixed. Assuming that

the bridge from the example is ranked in third place in one hypothetical bridge network

means that it will receive priority number 3 in the budget allocation process if the SHA

decides to use the road users benefit approach.

The main point in this example is that detoured trucks are responsible for 99.75%

of the total user cost benefit. That is, 7.6 % of the ADT volume is charged a total of

$3,413,036.00 per year for not using the bridge. This represents an additional expense of

$20.28 for each truck.

The penalty of $20.28 ($9.66/truck as equipment cost + $10.62/truck for labor

cost) paid for each truck for not using the bridge facility is directly related to the detour

length, the detour road speed, the VOC selected, and the travel time selected. The total

detouring cost is directly related to the number of trucks detoured each year at the bridge.

The lost savings of $ 0.35 cents incurred for each vehicle that uses the bridge is

related to the current ADT the current accident risk rate, and the future accident risk rate

after improvement. Each accident rate is related to roadway width, the approach alignment

rating, and the average cost per accident.

From the 10 inputs data used in the evaluation of user costs, only three are supplied

by the NBI source. Different offices in the SHA organization are responsible for generating

the seven remaining data input. Maintenance, inspection, safety, law enforcement,

engineering and administration offices are normally those involved in data generation for

BMS user costs. Four parameters related to policy decisions and one related to law

enforcement present the most difficulty in the process of selecting the correct ones. They are

the average costs per accident, the average vehicle operating costs, the average travel time

costs the speed in the detour costs, and the proportion of trucks detoured. The implications

involved in selecting each one of these parameters will be discussed in the next three



The Travel Time Cost (TTC) developed by this research is $22.55/hour. This

chapter describes how this value is developed, the theoretical basis to support its

development, and why TTC values benchmarked from other places are not useful for BMS

applications in Florida. The structure of this chapter is displayed in Figure 6 where the

solid line represents the critical path used to develop the new TTC value, and the dashed

line represents reasons why TTC benchmarked from other areas are not suitable for BMS in


6.1 Background

The value of travel time savings is usually considered by transportation analysts as

the most important category of benefits for major highway investments. According to

Strand (1993), for the average road project, 70 to 80 percent or more of the total benefits are

attributed to the time savings of the project. The way in which time is converted into

money is becoming more and more decisive in the calculations to profitability and

feasibility of road projects. Depending on the approach used to monetize travel time, one

can make a project profitable or unprofitable. Chui and McFarland (1986) state that before

1965, the estimated value of time was based more in intuition and non-behavioral estimates

than on a reliable, theoretical model. However, there are several conceptual problems in the

evaluation of time; pertinent questions about research issues and bottlenecks related to a

credible, practical application of time utilization theories. These issues can turn the use of

6.2 Theoretical Basis for TTC Evaluation

6.2.1. Behavioral Approach 6.2.2 Survey Approach 6.2.3 Resource Approach

16.3 Labor Wage Index Analysis

6.3.1 AASTHO I 62 S DIT 6.3.3 HERS 6.3.4 BMS

6.5 Non-business Approach 6.6 Business Approach

Equation 21 MODEL Equation 22 MODEL

6.8 New TTC Value $22.55/Hr

Figures 6. Flow Chart for BMS TTC Development

time both meaningless and misleading.

There is a fundamental conceptual difference between the value of travel time for

transportation planning and for BMS applications. For the former, drivers are offered

several alternatives to choose from on how to move from point A to point B. Road users

have the freedom to select the mode, the route, and also how much they are willing to pay

for time savings and better driving conditions. On the other hand, for BMS applications the

drivers do not have the freedom to select the mode or the route to move from point A to

point B, if point A and point B are connected by a bridge. If the bridge is closed or posted, it

is mandatory to remain in the same mode and take the bridge detour that has a specific

length and characteristic. In summary, planners are interested in knowing what costs an

individual is willing to pay in order to save one unit of trip time. With BMS applications,

the BMS practitioner is interested in evaluating what costs an individual is forced to pay in

order to overcome the extra time incurred in the journey. In the first case, the focus is in

travel time savings, and in the second case, the focus is in travel time costs that will be

considered as savings once the bridge deficiency is removed.

6.2 Theoretical Basis for Travel Time Evaluation

One major area of application of the macroeconomic approach to the evaluation of

travel time lies in the appraisal of improvements in transportation systems. The problem

consists essentially of an efficient allocation of resources in the economy, and how time in

transport is applied in benefit cost analyses. IF the principle of the welfare theory is used

(Bruzelius, 1978; Serpa,1971;), that is, on the individual maximization of benefits, the

marginal value of time is the one the consumer is willing to pay for a marginal reduction in

travel time. According to Strand (1993) this theory has been severely questioned, mainly by

the existing imbalance between theoretical and empirical calibration (Heggie, 1983), and

the way small savings and aggregate problems are treated.

Alternatively, a conceptually identical approach is used to evaluate the economic

losses due to unproductive travel time which is addressed in economic theory as

opportunity cost. For one category of time consuming human activity, one can establish a

value based on the marketing mechanism. A market for labor exists so that the time saving

in trips undertaken during working time (on-the-clock trips) can be assigned a value related

either to the wage rate or overhead costs. According to Miller (1996), on-the- clock trips

have values equal to the wage rate, plus fringe benefits and vehicle/inventory costs.

Behavioral models give the value of time from the perspective of the person whose

behavior is being modeled. Thus, models estimated from decisions made by travelers give

the value of time saved to the traveler. Models estimated from decisions made by travelers'

employers give the value of time to the employers (Morrison, 1996). Because of this, the

results from behavioral models are useful when predicting travel demand. However,

because of taxes, among other things, these values will most likely diverge from the

"resource" values appropriate for use in cost benefit analysis. The following are the

description of each of these approaches.

6.2.1 Behavioral Approach

The behavioral approach is an extension of conventional consumer theory that

illustrates why time in general and travel time, in particular, is valuable. At its core is the

tradeoff between time and money.

The state-of-the-art behavioral models indicated at least five approaches for

estimating the value of time. They are identified mainly by the data collected on the choices

that the sample of travelers make, e. g., modal choice model ( McFadden and Reid 1975,

Ben-Akiva 1973, Charles River Associates 1972, Lave 1968, Lisco 1967, Becker 1965,

Morhring 1960), route choice model (Guttman 1975, Thomas and Thompson 1970, Claffey

et al. 1961), speed choice model (Chui & McFarland 1986, Winston & Associates 1987),

usage of safety belt (Blomquist et al. 1996), and housing price approach (Nelson 1977,

Wabe 1976).

6.2.2 Survey Approach

According to Miller (1996) the survey approach is the most effective way to

evaluate the travel time value. He suggests that surveys should be done offering each

respondent just one or two randomly generated choices per question, with the values

implicit in the choices. Another alternative is to find an unbiased starting point then to let

the respondents choose a value from a menu that uniformly covers the entire response


6.2.3 Resource Approach

The resource approach is used to place value on business travel time for project

appraisal. In its simplest form, travel time savings are valued according to the "resources"

that are saved, i.e., the work time for "labor", and vehicle and equipment time for "capital"

(Hensher 1976).

6.3 Percent Wage Index Analysis

Value time estimate modelers use wages as a reference to monetize travel time. For

business travel time they use wages plus fringe benefits. According to Miller (1996), the

theoretical basis for the result that the value of work time (to society) is equal to the gross

wage plus fringes, is based on the assumption that the labor market and the output market

are competitive, that firms are able to substitute capital for labor, that there are no positive

or negative externalities in production, and that firms maximize profits. Miller (1996)

considers that this practice provides an automatic adjustment for inflation and facilitates

comparison across different currencies and cost of living. However, he questions the

validity of the theoretical foundation to support the use of wage to express value of time.

Miller (1996), after analyzing over 30 travel time studies recommends the use of

conservative estimates due to a poor fix on the value of travel time. According to him, for

project evaluation purposes, a value of travel time of 55% of the wage rate is recommended

for drivers with an uncertainty range from 50% to 75%. A value of 40% of the wage is

recommended for car and public transport passengers, and a value of 100% of the wage rate,

plus fringe benefits and vehicle/inventory costs is recommended for on-the-clock travel.

As can be seen, there is a consensus about the percent wage index to be used in the

evaluation of travel time for business. However, for non-business travel there is no

consensus. This is a direct reflection of the contribution level provided by the behavior

models at the present stage in evaluating travel time. In fact, according to Miller (1996),

analysis of highway investments should use travel time values from route choice or speed

choice models, and not from modal choice models. The weakness of mode of choice

models is that they underestimate the value of personal vehicle travel time savings,

comfort, convenience and privacy while ignoring capital investments in a private passenger

vehicle. This observation confirms the critics of Hensher (1976) and Gronal (1976) in the

use of modal choice models for a wide range of decision variables.

The Texas Transportation Institute (TTI 1993) pointed to concerns raised about the

validity of the assumptions made in the speed choice model, which assumes a driver's

knowledge of the relation of driving costs to driving speeds.

For a BMS application the route choice has a relatively low importance, because it

deals with the choice of routes (often a toll road vs. a parallel non-toll road) and in reality


under the BMS scenario there is no route selection. There is only the detour route prescribed

in the NBI item 19 related to each bridge which is fed into the BMS mathematical models.

These values do not change; consequently there is no choice involved.

6.3.1 TTC Under AASHTO Methodology

The traditional approach used for analysis of transportation investments is based in

the AASHTO 1977 Manual on User Benefit Analysis of Highway and Bus-Transit

Improvements. This manual is an extension and replacement of the 1960 AASHTO report,

Road User Benefit Analyses for Highway Improvements (AASHTO 1960), and the

National Cooperative Highway Research Program Report 133 (NCHRP 1972). Kinboco

and Henion (1981) critically evaluated the AASHTO 1977 report a few years after it was

published. However, travel time was not evaluated.

6.3.2 TTC Under USDOT Methodology

In 1997, the Office of the Secretary of Transportation (OST) released a

memorandum intended to provide DOT staff with the best current available procedures and

empirical estimates for calculating the value of time. The focus was on transportation

investments, not on BMS. This memorandum recommended the use of nationwide statistics

for income/wage rates of the traveling population, using 100 percent of the wage (plus

fringe benefits) for all local and intercity business travel, including travel by truck drivers.

This equals 50% of the wages for all local personal travel and 70% of the wages for all

intercity personal travel (OST, 1997).

6.3.3 TTC Under HERS Methodology

In 1999, the FHWA received the latest version of the HERS (Highway Economic

Requirement System) model, which is the result of DOT's efforts to better examine costs,

benefits and national implications associated with highway investment options. The


approach used to estimate the travel time value is to apply a value of 100% of the wage rate,

plus fringe benefits and vehicle/inventory costs for on-the-job travel, and 60 % of the wage

rate exclusive of fringe benefits for other trips (HERS, 1999).

6.3.4 TTC Under BMS Methodology

When a bridge has a deficiency, and this deficiency demands load and height

restrictions in the traffic flow, the SHA is faced with two traffic management alternatives:

post the bridge and divert a proportion of the traffic to a detour route; or close the bridge

and divert 100% of the traffic to a detour route. If the bridge is posted it is assumed that

only heavy vehicles (trucks) are diverted.

For BMS objectives, trucks and buses are always treated as being on business trips

and the rest of the traffic flow is treated as non-business trips. This assumption generates

some bias by ignoring the fact that besides trucks and buses, no other vehicles are engaged

in business trips. However, this assumption is adopted due to the lack of data about the

composition of business trips existing in the traffic stream. Business trips, also known as

on-the-clock trips, have travel time valued on the basis of savings to the employers as

discussed before. The savings include wages, fringe benefits, vehicle costs and inventory

costs (Miller 1996, HERS 1999). This approach, which includes vehicle costs and inventory

costs in the composition of the travel time estimation is not shared by the Canadian

Department of Transportation (Culley and Donkor, 1993). For travel time savings for

business travel, the Canadians use the approach of "equivalent hourly wage rate" that is

calculated dividing the average annual individual earnings by 2,101 hours (52 weeks of

work per year at an average of 40.4 hours of work per week). On top of this "equivalent

hourly wage rate" is added 19.5% for employee fringe benefits, plus 10% for unreported

overtime, plus 14.0% for fringe benefits. As an example a typical "equivalent hourly wage

rate" in 1990 was $16.39 resulting in a final value of $24.01 due to the benefits included

(HERS, 1999).

The rationale to include vehicle costs and inventory costs is the following: For

vehicle costs, capital invested is depreciated for the lifetime of the vehicle assuming a

certain salvage value at the end of the life. The resulting average vehicle cost per year is

then divided by a number of vehicle working hours in service per year per vehicle category.

The weakness of this assumption is that the traffic composition will be treated as having

only vehicles aged to the limit of the selected depreciation time. For example HERS

assumes a five-year life for autos and four-tire trucks. That is, if in the traffic composition

there is a vehicle older than five years, it is treated as a newer vehicle, inflating the

depreciation value of the traffic composition. Data from the 1995 National Transportation

Survey indicates that the average age by vehicle type has been increased since 1977. The

1995 average age for cars and trucks were 8.24 and 14.93 years respectively. These numbers

show that if one uses the HERS approach of five- years life for vehicles to determine their

hour value, one is attributing, to a proportion of old vehicles, the value of a newer vehicle.

The same survey discloses that for 1995 only 37.7 percent of all vehicles are in the 0 to 5

years age range.

For inventory costs the rationale is to select one discount rate and multiply its hourly

value by the average value of the truck cargo. The problem here is knowing what type of

cargo each truck is carrying. HERS assumes that 35% of all combination trucks carry low

value natural resources and agricultural products, and the remaining 65% of trucks carry

manufactured products, including goods of medium to high value, processed foods, building

materials and paper products. The weakness of this approach is due to the assumptions

made. It basically divides the truck population in two groups in which one will be sensitive

to the inventory cost and the other group will not. Using the values published by the 1977

transportation census, we have a median value for manufactured commodities of $2.29 per

pound and a value of $ 0.04 per pound for non-manufactured goods. One more assumption

that is necessary to be made is the assignment of the class of cargo (manufactured versus

non-manufactured) with the type of truck. These assumptions introduce a bias factor in the

evaluation of the inventory cost.

Non-business travel time savings are valued through surveys and behavior models.

Under the classes of business and non-business travel it is necessary to investigate the

implications for the value of time in personal and business travel.

6.4 Comparing Travel Time Values

Walls and Smith (1998) used the All Items Component of the CPI to update the

travel time costs to August 1998 of four different travel time studies (NCHRP 1972,

NCHRP 1990, OST 1997, and HERS 1996). Comparing the values of the NCHRP 1972

(1970 dollar) with the values developed by HERS 1996, increases of 377% for the

passenger and 526% for trucks travel time cost were observed. Table 10 shows the update

to 1996 dollars of the values of travel time from four different sources.

This procedure of updating travel time costs is one of the factors that contribute to

the poor fix on the value of travel time. This procedure freezes all the technological

improvements implemented in the vehicles and in the transportation infrastructure over the

years. Further, it produces a static view of assuming that driver behavior has not changed

since the former travel time value was established. Another factor that contributes to the

low quality of the value of travel time is the practice of using the results of travel time

studies performed outside the United States for applications within the United States. This

practice assumes, for example, that Brazil's roadway network is similar to the United

States' network; it also assumes similarity with vehicles and driver behavior between the

two countries. In fact, according to the Highway Capacity Manual (HCM 1994), weather,

pavement conditions, user's level of familiarity with the facility, and incidents in the traffic

flow are the factors considered in the evaluation of the roadways.

Table 10. Composite Listing of Update Travel Time Costs

Source Units Autos Trucks Combination
USDOT-OST* $/Person-Hr $10.80 $16.50 $16.50
MicroBENCOST $/Veh-Hr 11.37 17.44 24.98
NCHRP $/Veh-Hr 11.78 19.64 19.64
HERS $/Veh-Hr 14.30 25.99 31.30
Escalator Factor used: travel time cost- August 1996
*USDOT-OST= US Department of Transportation-Office of Secretary of Transportation.

6.5 Non Business Travel Time

From the perspective of economic theory, non-business trips are those trips that are

not on-the-clock trips. The market value assigned for these trips is not by the employer, but

by the driver. This assumption creates some difficulties in allocating the appropriate value

for each class of drivers. The classical example is the housewife time cost allocation

problem. The work performed in the home by the housewife is not compensated by the

actual exchange of money, and consequently it is not reported to the Internal Revenue

Service, generating the false concept of work performed with no monetary value. However,

if a professional provides the same work previously performed by the housewife, this work

now has a monetary value. The traditional approach used to solve this problem is to select a

percentage of the wage rate and assign this percentage to all non business trips, without

considering fringe benefits.

As mentioned before, a value of 55% of the wage rate is recommended for drivers,

40% for passengers with an uncertainty range from 50% to 75%. Using the 1995 census

data of 1.59 average vehicle occupancy for all purpose trips the following general formula

can be established:

TTNB = ( WPD x FWR) + (WPp x FWR) (OR -1) (17)


TTB- = travel time for non business
WPD = wage proportion for drivers (%)
FWR = full wage rate ($)
WPp = wage proportion for passengers (%)
OR = vehicle occupancy rate

6.6 Business Travel Time

Considering that one of the objectives of this research is to develop user costs for

BMS applications that includes no bias in its development process, it is important that the

model use concepts that can be supported by a theoretical base accepted by the

transportation community. Concepts supported with preponderance of assumptions should

be avoided or minimized.

In the case of travel undertaken by employees in the course of work, the economic

theory states that the value of travel time consists of the resource value of the time itself,

plus the monetary equivalent of whatever utility or disutility results from spending time

traveling. Assuming time spent traveling would otherwise be spent working, and that no

productive use could be made of travel time, the employee's pre-tax wage rate plus the

monetary value of fringe benefits represent the value of time.

Based on these premises the conceptual mathematical model to evaluate business

(on- the- clock) travel time, can be expressed in the following way:

TTB= (FWRD + FB) + (FWRH +FB) (OCR -1) + VCC + VIC (18)


TTB = travel time for business
FWRD = full wage rate for drivers
FWR H = full wage rate for helpers
FB = fringe benefits
OCR = vehicle occupancy rate
VCC = vehicle capital costs
VIC = vehicle inventory costs

The values to be used for wage rate and fringe benefits can be found in the US

census for each category of driver. In our case we are interested in truck and car drivers.

Trucks and heavy vehicles (military vehicles) are the main source of bridge structure

degradation. Extensive research has demonstrated that single unit tandum and triaxle dump

trucks have a high potential for overstressing bridge structures. Further, illegal overloading

of trucks is now to a level such as to cause significant overstressing of bridge members

(Schelling, 1985). The Highway Performance Monitoring System (HPMS) lists thirteen

different vehicle categories (Mactavish and Neumann, 1982). HERS uses one classification

for cars and four different classifications for trucks: 4-tire truck; 6-tire truck; 3-4 axle truck;

4-axle combination truck; and 5-axle combination truck. The FDOT Weigh-in-motion

(WIM) study shows that a 5-axle combination truck represents 50% of the truck traffic

composition in the Florida roadway network, and 67% of the truck traffic composition in

the Interstate Highway System. Since cars do not affect the bridge structure, and normally

they are not required to detour due to bridge limitations, the use of one classification for

cars seems appropriate. The classification used by HERS is considered to be equivalent to

the truck composition in the Florida Network, and for this reason it will be used to evaluate

the basic values of each vehicle. After this evaluation, these values will be adjusted


according to the FDOT weigh-in motion (WIM) truck percentages in the traffic composition

(WIM, 1998).

Another factor that must be considered is the appropriate vehicle occupancy rate

index to be used in the business travel time evaluation. There are two methods that measure

vehicle occupancy: the travel method and the trip method. The travel method computes the

vehicle occupancy as person miles of travel per vehicle mile, and the trip method computes

the vehicle occupancy as persons per vehicle trip. Because longer trips often have higher

occupancies, the travel method generally yields a higher rate. For BMS applications the

travel method seems more appropriate. The average vehicle occupancy, measured as person

miles per vehicle mile, has decreased consistently over time. This trend is related to an

increase in vehicle ownership, and decreases in household size. Data from 1977 indicates a

vehicle occupancy rate for all purposes of 1.9; in 1995 this rate decreased to 1.59. The

vehicle occupancy rate for trucks is generally accepted to be 1.0 for heavy trucks and 1.1

for small trucks and pick-ups that use helpers on business trips (HERS 1996). Data from

Florida Truck Permit Office shows an occupancy rate for trucks of 1.0, and this value will

be used for heavy trucks. For small trucks and pick ups the occupancy rate of 1.1 will be

used based on HERS data.

6.7 Value of One Hour Travel Time

Table 11 presents the value of one-hour travel time by benefit category and vehicle

type. Labor, fringe benefits, vehicle capital costs, and vehicle inventory costs are the TTC

constituents. The values are in 1997 dollar.


Table 11. Value of One Hour Travel Time for Business and Non-Business Trips by Vehicle
Category (1997 dollars)

Vehicle Type
Category Auto 4-Tire 6- Tire 3-4 Axle 4-Axle 5-Axle
Truck Truck Comb. Comb. Comb.
Labor 15.33 14.19 12.9 12.9 12.9 12.9
Fringe Benefits 2.97 5.10 4.6 4.6 4.6 4.6
Vehicle Capital Cost 0.42 0.50 0.86 2.79 1.95 2.00
Vehicle Inventory Cost 0.00 0.00 0.00 0.00 0.50 0.50
Total 18.72 19.79 20.15 22.58 22.24 22.29
Wage 9.27 N/A N/A N/A N/A N/A
Vehicle Occupancy 1.59 1.1 1.0 1.0 1.0 1.0

6.7.1 Labor/Fringe Benefits

For autos, the hourly wage per vehicle occupant for on-the-clock trips in 1997 was

assumed equal to the Florida median household income of $31,900 (1997 dollar) divided by

2,080 hours/year. The main reason for selecting the Florida median value in place of the US

median value for household income was to follow the criteria established previously for

prioritizing the use of local data. The fringe benefits value used is the US average of 19.73

% published in the Statistical Abstract of the United States (STA 1989). The US and the

Florida Statistical Abstracts use the median to report the household income. Median income

is the amount that divides the income distribution into two equal groups, one having

incomes above the median and the other having incomes below the median. For households,

the median income is based on the distribution of the total number of units including those

with no income (FSA, 1998). The median better represents the reality than the mean income

that is obtained by dividing the total income of a particular statistical universe by the

number of units in that universe.

For trucks, the value used for labor is the one published by the US Census Bureau

for weekly median full-time wages for truck drivers, that is $516/week. For truck helpers

the value assigned was the median weekly earnings of full- time wages for freight, stock,

and material handlers of $399/week. The fringe benefit for trucks was the value reported by

HERS of 36% which was taken from the Teamster's Union contract for the Central United


6.7.2 Vehicle Capital Cost

The evaluation of vehicle capital costs made by the HERS model is made assuming

a five-year life with a 15 percent salvage value at the end, and with initial costs taken from

the Motor Vehicle Manufacture Association. This approach is appropriate only if all

vehicles in the traffic composition work only five years. As mentioned before, the average

vehicle age existing in the traffic composition is higher. For this reason a factor will be

introduced to lower the value of each vehicle capital cost. This factor is calculated dividing

5-years by the average vehicle age reported in the 1995 summary of travel trends. The

values are 0.60, 0.57 and 0.33 respectively for cars, four wheel trucks and heavy trucks.

For heavy trucks, the cost per hour was computed as the average vehicle costs per

year divided by the number of hours in service ( assumed to be in service only 1600 hour

per year). For cars, four and six tire trucks it is assumed that they work 2000 hours per year

(HERS, 1999).

6.7.3 Vehicle Inventory Costs

The methodology to calculate vehicle inventory costs, as mentioned before, is based

on applying one discount rate to the value of the payload carried by the truck. The HERS

model uses a discount rate per hour of 0.0011 percent. This rate is derived from a discount

rate of 9.8 percent plus 1 percent. The average payload of a five-axle combination was

assumed to equal 30,000 pounds. The type of cargo carried by all combination trucks was

spilt to 35% carrying natural and agricultural products, and 65% carrying manufactured

products. The median value per pound of the manufactured products indicated in the

Commodity Transportation Survey data is $2.29, and for agricultural and natural products

$ 0.04 per pound. With these values, the average payload is valued at roughly $45,000,

yielding a time value of $0.505 per hour (HERS, 1999).

6.8 Truck Travel Time Cost for Florida

Using the truck composition data from the WIM study for the FDOT that is in

progress at the Civil Engineering Department, University of Florida, only the last three

vehicle types listed on Table 12 are statistically relevant for BMS applications. The research

result values are showed in Table 12.

Table 12. Research Result Value for TTC

Truck Type $/Hr Percent Total (CPK)
3-4 Axle Truck 22.58 92.0 20.7736
4-Axle Combination 22.24 7.0 1.5568
5-Axle Combination 22.29 1.0 0.2229
Total 100.0 22.5533


The Vehicle Operating Costs (VOC) developed by this research is 31.3843

cents/Km. This is the marginal cost for a truck run for one extra kilometer per trip. This

chapter describes how this value is developed, and the methodology used to evaluate the

VOC components (fuel, maintenance, tire and oil change). A VOC for cars and light trucks

is presented in Appendix H. The structure of this chapter is shown in Figure 7, where the

solid line represents the critical path to develop the new VOC for trucks, and the dashed

line represents variable costs not included in this VOC development.

7.1 VOC Related Factors

7.1.1 Weight Limits
7.1.2 VOC Value Ranges
7.1.3 Equipment, Carrier, and Capacity Optimization Types
7.1.4 Political Factors
7.1.5 Road Network Compatibility

7.2 Running Cost Calculation methodology

7.3 Variable Costs]

7.3.1 Depreciation 7.4 VOC Truck Customization
7.3.2 Labor Costs
7.3.3 Miscellaneous Costs
7.3.4 Accident Costs 7.4.1 Fuel Costs
7.4.2 Maintenance
7.4.3 Tires
7.4.4 Oil Changes

7.5 New CV, Value 31.3834 Cents/Killometer

Figure 7. Flow Chart for BMS VOC Development

7.1 VOC Related Factors

According to Equation 19 the motor vehicle operating cost (VOC) is evaluated as a

function of average vehicle operating costs per kilometer of detour (CVv) and detour

distance (D,). CV, is the cost incurred for a truck to run one kilometer. Since the detour

distance is a fixed parameter in the Pontis matrix, the parameter that is variable and

sensitive to user costs is the CV,.

VOC= CVv x D, (19)


VOC= motor vehicle operating costs
CVv =average vehicle operating costs per kilometer of detour
D, =detour distance for the roadway in km

The term motor vehicle for Pontis BMS means a truck. The reason why Pontis

focuses only on trucks is due to the fact that bridges are designed with the constraints of

bridge formula B, which defines the maximum weight that may be carried on two or more

truck axles. This weight is evaluated as a function of the spacing between any two

consecutive axles and the number of axles according to the following mathematical


W= 500 [(LN/(N-1) + 12N +36] (20)


W =Maximum weight that may be carried on two or more axles, in lbs.,
L =Spacing, in feet, between the outer axles of any two or more consecutive
N =Number of axles being considered (Francher and Gillespie, 1997)

7.1.1. Weight Limits

Current federal limits on truck weight, length and width are defined by the Surface

Transportation Assistance Act (STAA) of 1992. These limits apply to all vehicles using the

Interstate system and other designated federal-aid highways. The weight limits are

nominally defined as

9,000 Kg (20,0001b) on a single axle
15,000 Kg (34,0001b) on a tandem axle group
36,000 Kg (80,0001b) maximum gross weight (with a few exceptions) (Fancher and
Gillespie, 1997).

Since bridge designers use bridge formula B with the maximum constraint value,

that is the maximum vehicle load allowed by legislation, it becomes clear that only heavy

trucks cause stress to bridge structures, with high concentrated loads being more stressful.

Cars, pickups, and light trucks normally do not cause stress to bridge structures, unless the

bridge has severe limitations like historical wood bridges. For this reason, Pontis considers

that under normal circumstances, when a bridge is posted, only a percent of trucks must be

detoured. However, if the bridge is closed, all traffic must be detoured, and for this reason

the evaluation of truck and car operating costs should be considered. The VOC for cars,

vans and light trucks are displayed in Appendix H.

7.1.2 VOC Value Ranges

The evaluation of operating cost is a tough challenge faced by the truck industry.

According to Cox (1996a), it is impossible to come up with an industry "average" cost per

mile, because it depends on a number of variables: the cost of fuel, the number of miles run

per year, the truck age, the driver lifestyle, and so on. However an average cost per mile

(operations cost) is one of the most useful pieces of information necessary for the business.

Truck operating costs will depend on the type of truck and the type of operation in which

they are used. According to Elgin (1998), it is estimated that only ten percent of owner

operators know their cost per mile, and that the old proverb of $1.00 per mile just does not

apply in today's environment.

Comparing the data from a population of 104,561 tractor trucks associated with

332,838 trailers, distributed into 98 carriers in 1997, the operating costs range goes from

14.25 to 111.04 cents per mile (TFM, 1998). Comparing the operation cost of all Florida

truck companies reported in the 1997 Motor Carrier Annual Report a range of 9.3 to 100.9

cents per mile was observed. There are several reasons why there is a large value range in

the operating cost reported. The main reason is a lack of a uniform methodology to evaluate

operating costs associated with the diversity of equipment and carrier types.

7.1.3 Equipment. Carrier and Capacity Optimization Types

According to the 1999 National Motor Carrier Directory Record Counts, there are

11 equipment types and 19 carrier types hauling commodities. They are listed on Appendix

C. To increase complexity, the FHWA uses a classification where vehicles are grouped into

5,000 pound weight categories ranging from 5,000 to 150,000 pounds generating a matrix

of 20 vehicle classifications. That is, there is not a uniform truck classification that covers

all users.

Other factors that contribute to the difficulty in evaluating truck-operating costs are

related to the capacity of each truck. If the truck runs at full capacity, it is classified by the

truck industry as truckload-TL If the truck runs with partial capacity it is classified as Less-

Than Truckload LTD. The ownership type of the truck is also responsible to create new

classifications: owner-operators and fleet operators.

7.1.4 Political Factors

Besides the factors listed above that contribute to the difficulty in evaluating truck

operating costs, there is also a political issue to be considered. That is the old "political

war" between FHWA and the truck industry that is related to the truck maximum load

capacity allowed by law, and the appropriate tax share that the truck industry should pay to

use the roadway network.

At the same time that the truck industry has a strong lobby to increase the limits on

truck weight and length based on the premise that these will improve productivity in the

transportation sector, the FHWA feels that the tax share paid by the truck industry is not

adequate to cover the related pavement damage. Heavier trucks produce more damage and

consequently require more taxation for repair and replacement. As a result of this "political

war" an increase of 9.17 percent in the maximum vehicle load has been achieved during the

last five decades (Noel et al. 1985), and large discrepancies are observed in the tax share

paid by truck owners. In relation to the tax share paid by each FHWA vehicle classification,

the 1997 Federal Highway Cost Allocation Study results show a negative equity ratio

between revenues and costs. The study presents the estimated cost burden for different

vehicle classes and registered weight groups for the Federal related program cost funded

from the Highway Trust Fund (HTF) in 2000. Comparing these costs with the Federal user

fees paid per mile of travel using the same vehicle class and same year, it was found that the

ratio of the shares of revenues contributed by each vehicle class to the share of highway

costs are unbalanced. Pick-ups and vans have the largest over payment of any vehicle class.

Other vehicle classes in the aggregate that pay more than their fair share are 2-axle single

unit trucks, all truck-trailer combinations, and 5- and 6 axle twin trailer combinations.

However, 5-axle tractor semitrailers have the largest underpayment of any vehicle class,


followed by automobiles and 3-and 4-axle single unit trucks (USDOT, 1997). Appendix D

shows the breakdown of the 2000 Federal highway cost responsibilities and user fee

payments by vehicle class and weight group in cents per mile, and the percent cost share.

The truck traffic composition in Florida identifies 68% of the trucks as 5 -axle tractor

semitralers (WIM, 1998).

Traditionally, truck-operating costs are not disclosed by the truck industry on the

premise that it is internal information that is part of the negotiation process. Harrington

stated in his 1994 article that the University of Tennessee found a growing desire among

shippers for simpler, more predictable carrier pricing structures. Many are tired of the

mystery surrounding rates and pricing, and in fact find that mystery a hindrance in serving

their customers effectively (Harrington 1994). Some government agencies like the USDA

require a disclosure of the truck operating costs in their contracts. Other sources of

information are the income statement published by commercial carriers, and data from truck

associations. Probably part of this behavior of not displaying clear operating costs is to

protect themselves from higher taxation and extra labor costs.

7.1.5 Road Network Compatibility

The majority of the VOC studies made around the world are concerned with

discovering the relationship between highway design, pavement condition and road user

costs to be used in a transportation investment feasibility analysis. If the findings of these

studies are assumed to be applicable to the USA, with some simplifications of the formulas

developed, it is possible to generate costs for fuel, oil, tire wear and maintenance costs.

However, these studies were developed assuming compatibility between the roadway

network from developing countries with the USA roadway network. Since one of the


objectives of this research is to reduce the bias in the user cost calculation these studies will

not be valid for our investigation.

7.2 Running Costs Calculation Methodology

Truck running costs are those costs which are incurred solely when a vehicle is

operating. They have nothing to do with the costs of owning the vehicle or with the

expenses involved in running that transport business as a whole. They are classified as a

variable costs and they are comprised of the following four items: fuel, tires, maintenance,

and lubricants.

Under the vehicle costs classification there are costs which relate directly and solely

to individual vehicles, like licenses, insurance, drive labor, rent, interest and depreciation,

that are labeled as "standing costs"(Lowe, 1974).

All those expenses incurred in running a transport business, which cannot be

directly attributed to any individual vehicle, are labeled "overhead costs". All those

expenses incurred when the vehicle is running are labeled "running costs". The combination

of standing costs plus overhead costs and running costs are recognized by some sectors of

the truck industry as operating costs (Lowe, 1974).

7.3 Variable costs

Running costs are under the classification of vehicle variable costs as listed in Table

13. In reality what Pontis BMS requires is the marginal costs which is the additional cost of

running an extra kilometer, e.g. the variable costs.

In some cases, authors use the term "operating costs" with the same meaning as

running costs. There is not a consensus in the literature about the classification of variable

costs parameters for evaluation of vehicle operating costs.

Table 13. Vehicle Costs Classification Between Fix and Variable Costs

Fixed Costs Variable Costs
Licenses Fuel
Insurance Oil
Depreciation Tires
Rent Maintenance
Interest Tolls
Overhead Labor

7.3.1 Depreciation

Comparing old studies with recent ones, the main change observed is that

depreciation which was considered a variable cost in old studies, is being treated as a fixed

cost in recent studies.

Early studies include depreciation as a variable cost. However, it is noticed that it is

not included in the comprehensive VOC studies after 1987. Table 14 presents the

MicoBENCOST study as a recent study since it was published in 1993. However, the data

used to develop its VOC was based on an old study performed by Zaniewsk in Brazil for

the years 1975-1982. Today some carriers generate two weights for the same measurement.

One hypothetical example can clarify this point. Assume that company A starts a

transportation business with 100 new trucks with a policy to depreciate by mile, and to

replace the trucks when they reach the 400,000 miles. After a period of time (t) that is

defined by the minimum time allowed by the Internal Revenue Service for depreciation and

50% of the trucks really reached of 400,000+ miles and 50% reached only 200,000+ miles

in time (t). Technically speaking company A accrued depreciation costs for all 100 trucks as

expenses in order to maximize its profit. Further 50% of the fleet would subsequently have

zero depreciation in their operating costs.

Table 14. Variable Costs Assigned at Operating Vehicle Cost Studies

VOC study Misc. Labor Fuel Oil Tires Maint. Depr. Acc.
Winfrey, 1962 X X X X X
Welle, 1966 X X X X X
Clafey, 1971 X X X X X X
Zaniewski, 1982 X X X X X
OOM, 1986 X X X X X X
Chester & Harison, 1987 X X X X X
Witconis & Stadden, 1988 X X X X X
FHWA, 1991 X X X X X
USDA, 1991 X XX XX
MicroBENCOST, 1993 X X X X X
MCAR, 1997 X X X X X
Berwick, 1997 X X X X

Hypothetically if a company can extend the life of its trucks by careful maintenance,

working for a period of time with a totally depreciated vehicle, the level of profit will

increase assuming that the maintenance cost are keep under control. According to Ryder, a

large carrier, the operating costs at 500,000 miles were close to those at the million-mile

mark (Romba, 1995). In other words, trucks that experience large mileage per year generate

lower depreciation per mile, and trucks that experience low mileage per year generate

higher depreciation per mile. This policy increases operating costs if the truck is not

running, which is contrary to the standard classification of variable costs. For this reason

depreciation cannot be considered as part of the running costs. The National Accounting &

Finance Council classifies depreciation as a fixed cost (NAFC, 1994).

7.3.2 Labor Costs

The cost for labor was considered in the travel time cost evaluation, and for this

reason it will not be included in this section.

7.3.3 Miscellaneous Costs

Expenses under miscellaneous and other costs are those that do not fit in any other

accounting category. An investigation of log sheets designed to record cost per mile

expenses revealed that miscellaneous and other costs are related more to the driver's life-

style than with the equipment (Witconis and Stadden 1988). The cost per mile log sheet

published by the Overdrive Book Division & Randall publishing Co. provides a breakdown

of the variable costs and divides it into two parts. Variable costs (I) that are related to the

truck, and variable cost (II) that are not related to the truck. When a cost per mile log sheet

does not break down the variable costs into two parts, they list the costs associated with the

truck, and then list miscellaneous costs. From this observation we can conclude that

miscellaneous costs are those costs that are not related to the truck running costs.

7.3.4 Accident Costs

Only one study lists accident costs under vehicle operating costs. For Pontis BMS,

accident costs will be given a specific treatment. These costs are not treated as part of the

VOC costs. They will be discussed in the next chapter.

7.4 VOC Truck Customization

As mentioned before for PONTIS only the truck VOC is important. The values of

fuel, maintenance, oil and tires will be calculated. The weigh-in-motion (WIM) for Florida

during calendar year 1997 shows that 61.61% of the trucks that use the Florida road

network is class 9. Trucks under class 9 are identified as a 3-axle tractor +2-axle trailer, and


3-axle truck +2axle trailer. These vehicles have a Gross Vehicle Weight (GVW) of 80,000+

pounds. Vehicles, class 5 to 8 have a maximum GVW of 68,000 pounds. For the evaluation

of the average operating costs, it was assumed that the vehicles under class 5 to 8 would

not be required to detour at any bridge site. This is to say that 33.73% of the trucks from the

WIM study will not be classified as heavy trucks. Vehicle classes 10 to 13 are combination

tractors with a number of axles between five and more than seven with GVW ranging from

80,000 to 140,000 pounds. The percent of heavy trucks considered in this study was

calculated to be 92% as a five-axle -80,000 lbs., 7% as a seven axle- 87,000-97,500 lbs.,

and 1% as nine-axle- 80,000 to 113,500 lbs. (WIM, 1998).

7.4.1 Fuel Costs

Fuel costs are by far the trucker's largest variable expense. Cox indicates that the

driver can directly control the fuel consumption in four ways: reduce highway speed, reduce

idling, proper tire inflation and smart driving (Cox, 1996b). Engineers from Donaldson Co.

add to the Cox list some maintenance items like excessive exhaust back pressure or air

intake restrictions that can cause as much as 4% drop in the fuel economy. As a rule of

thumb (since it was not possible to trace the source of the named "fuel tests") for each 1-

mph increase above 55 mph, 1/10 gallon more fuel per mile is consumed. A truck at 65

mph, on the average, will get 1-mpg less than it does at 55 mph. Idling, particularly fast

idling, consumes up to 1 /2 gallon per hour, and according to the Maintenance Council

tests, under- inflated tires (70 psi vs. 100psi) can reduce fuel mileage between 1.5% and 3%

(HDT, 1990). Campbell points out that fuel purchased under credit plans can cost 6 to 12

cents more per gallon than fuel purchased under cash plans (Campbell 1991). According to

MacDonald (1993) if the drivers can run their trucks in the highest possible gear at all times

they will minimize the revolutions per minute, and will maximize mpg at the rate of 1-

gallon less fuel per hour (Macdonald 1993).

Kanapton (1981) developed an empirical equation for estimating unit vehicle inter-

city freight fuel consumption. It is a linear function of weight in which the constant term

represents the fuel consumed in moving the empty vehicle, and the variable term represent

the fuel required moving the payload:

GPVM= FFC + VFC (P) (21)


GPVM = is the average line-haul fuel consumption in gallons per vehicle mile.
FFC = is the fixed component of fuel consumption which is required to
overcome resistance of the tare weight and aerodynamic drag.
VFC = is the variable component of fuel consumption and is the additional fuel
required to move the payload.
P = is the payload

Jack Faucett Associates (1991) updated the Kanapton formula using a 65-mph speed

limit, and diesel fuel at $1.25 per gallon. One change was made to Kanapton's formulas:

they were modified so that they would be a function of the gross vehicle weight (GVW)

instead of a function of payload. Coefficients for fixed and variable costs were obtained for

four vehicle types (vans, flatbeds, tanks and dump trailers). The fuel cost per vehicle mile

was derived from 1988 dollars. The CPI index was used to update the data to 1998 dollars.

These are listed in Table 15 according to the number of axles and GWV.

Table 15. Estimated Cost for Fuel Per Mile by Truck Category

Axle/L GVW 000 TL Van Refrigerated Flat Bed Tank Hopper Dump
(lbs.) (CPM) Van (CPM) (CPM) (CPM) (CPM) (CPM)
5 axle twin 80 23.0 27.1 22.0 21.3 15.1 15.1
7 axle twin 80-97 23.7 28.8 23.1 22.8 16.1 16.1
9 axle twin 89,2 23.9 30.9*, 24.2*, 22.8*, 15.1*, 16.1*4

Calculated at GVW of: 1=13,500 lbs. *2=90,000 lbs. *3=80,000 lbs. *4=81,000 lbs.
Source: Adapted from Jack Faucett Associates (1991)

Using these percentages it was possible to evaluate the fuel cost per kilometer. The

methodology used was to calculate the average weighted value of each fuel cost listed in

table 15 and then split these values into six selected equipment types shown in Table C 1-

Appendix C. The results are show in Table 16.

Table 16. Fuel Costs Distribution by Equipment Type

Equipment type Percent Fuel Cost Totals
Per mile Cents/ Mile Cents /Km
TL Van 39.19 23.058 9.03643
Refrigerated 12.87 22.289 2.86859
Flat bed 23.19 22.099 5.12475
Tank 7.26 21.420 1.55509
Hopper 10.48 15.180 1.59086
Dump 7.21 15.131 1.06068
Grand Total 1 (1988 dollar value) 21.23642
Grand total 2 (1998 dollar value) 28.88506
Grand Total 3 (1998 cents-per-kilometer) 18.05316

7.4.2 Maintenance

The maintenance cost was calculated based on a formula from Jack Faucett and

Associates (1991) where a scaling procedure was used. The formula is weight sensitive and

is based on a gross vehicle weight of 58,000 pounds. At this level, a value of 10 cents per

vehicle mile is used, and for each 1,000 pounds increase in the GVW .097 cents is added.

The formula for maintenance is:

Maintenance Costs = 10 + [(GVW-58, 000)/1000] 0.097 (22)

Using the same percent of heavy trucks mentioned in section 7.10, the total of 12.4598

cents per mile was calculated as listed in Table 17.

Table 17. Maintenance Costs for Trucks- CPM/CPK

% truck GVW Equation 22 Total
Cents/mile Cents/Km
Composition (000) Cents/mile Cents/e
92.0 80 12.13 11.1596
07.0 97 13.78 1.1207
01.0 113.5 15.38 0.1795
Grand total 1 (1988 dollar value) 12.4598
Grand total 2 (1998 dollar value) 16.7001
Grand total 3 (1998 dollar value) 10.4377

7.4.3 Tires

Tire costs were derived from the Pace Report where truck expenses were reported

by 98 carriers in 1997 (TFM, 1998). A value of 3.5557 cents-per-mile was derived using

1987 dollars. Using the CPI index for 1998, the update value is 3.612303, which is

equivalent to 2.25765 cents-per-kilometer. The statistics with a 95% confidence level is

shown in Table 18.

Table 18. Tire Costs Descriptive Statistics--CPM and CPK

Descriptive Values Results
Mean 3.55570(CPM) 2.2576 (CPK)
Standard Deviation 3.37248
Standard Error Mean 0.34971
Upper 95 % Mean 4.25026
Lower 95 % Mean 2.86114
Number of entries 93

7.4.4 Oil Change

The majority of vehicle operating cost studies found in the literature includes oil

change under fuel. Cox (1997) reports the cost of 1 cent-per-mile after running 125,000

miles in a year. Using the CPI index for 1998, the update value is 1.016119, which is

equivalent to 0.6350 cents-per kilometer.

7.5 New CVv Value

Adding the costs of fuel, maintenance, tire and change oil in cents- per-kilometer,

using 1998 dollars, the sum in 31.3834 cents-per-kilometer which is the new CVv value for

Pontis. The composition of the total cost is listed in Table 19.

Table 19. New VOC Value for Pontis CV, Default Value

Variable Costs Cents per Kilometer
Fuel 18.0531
Maintenance 10.4377
Tire 2.2576
Oil Change 0.635
Total 31.3834


The new Average Accident Cost (AAC) default value developed by this research is

$68,404.39 per accident under the comprehensive approach (including social costs), and

$27,365.75 per accident under the economic approach. A total of 10,115 crashes were

considered which occurred on 4,505 bridges in Florida in 1996. The injury costs developed

by this research range from $3,014,525 to $8,815.72 per injury. They are shown in Table

20. The structure of this chapter is shown in Figure 8.

Table 20. Injury Costs by Injury Type--Year 1996

Injury Type Cost in 1996 Dollars
Approach Fatality Injury A Injury B Injury C
Comprehensive $3,014,525 $211,515.4 $45,927.2 $29,844.7
Economic $871,697.2 $49,294.19 $12,289.65 $8,815.72

8.1 Bridge Accident Cost Evaluation Methodology

8.2 Comprehensive Fatality and Injury Costs

8.2.1 Fatalities 8.2.2 Nonfatal Injuries

8.3 Conversion ofMAIS into ABC Classification
S 8.4 FDOT Crash Data Preparation

8.4.1 Matching Crashes to Bridges 8.4.2 Injuries in the Database

8.4.3 Eliminating Questionable Bridges

8.5 New Bridge Related Accident Average Cost: $68.404.39/Accident

Figure 8. Flow Chart for the Average Accident Costs Development

8.0 Background

According to several studies, bridge related accidents are more severe than other

roadway accidents (Chen and Johnston 1987, Turner 1984, Mitchie 1980, Hilton 1973).

The literature shows indices that measures the severity of bridge related accidents varying

from 2 to 50 times the severity of general roadway traffic accidents. It is difficulty to point

to the main cause of a bridge-related accident. One of the former studies on this subject

found that the average daily traffic (ADT), sharp curvature at approaches, and bridge width

had major effects on bridge related accidents (Raff, 1953). Another study correlates narrow

bridge roadway, width and curved approaches as the most important factors contributing to

accidents at bridges sites (Hilton, 1973). Empirical observations also have confirmed these

results. Recent example was the death of Princess Diana in a car accident that hit a bridge

structure (underpass) in France. The main point in establishing that bridge related accidents

rates are more severe than roadway accidents is that the higher the accident rate the higher

the user cost. According to Chen and Johnston (1987) the average cost of accidents

involving bridges was estimated to be 5 to 8 times the costs of general motor vehicle


8.1 Bridge Accident Cost Evaluation Methodology

The number of bridge related accident occurring in Florida was evaluated using the

FDOT bridge accident database for 1996. The data was arranged in a spreadsheet using each

line to record an accident. If the accident occurred in a scenario where two or there bridges

were involved, an allocation factor was used to split the costs. For two bridges an

allocation factor of 0.5 was used, for three bridges the value was 0.33, and for four bridges

the value was 0.25.

There are four types of injury classifications in the database selected from the KABCO

classification listed in Table E-l, Appendix E. The KABCO injury scheme is designed for

police coding at the crash scene. The American National Standards Institute (ANSI) in

standard D-16.1 defines it.

Besides the injury classification for each accident, a correspondent property damage

is also reported in the FDOT database. A total of 10,015 bridge accidents were found for the

year 1996 in the FDOT database involving 4,505 bridges. The average cost of each accident

was calculated using Equation 23.

ABAC= (FC K + IAC A + IBC B + ICC C + PDO) AF (23)


ABAC = Average Bridge Accident Cost ($)
FC = Fatality Cost ($) (Research Resulted Value)
K = Number of fatalities (FDOT database)
IAC = Injury Type A cost ($)(Research Resulted Value)
A = Number of Injury A (FDOT database)
IBC = Injury Type B cost ($)(Research Resulted Value)
B = Number of Injury B (FDOT database)
ICC = Injury type C cost ($) (Research Resulted Value)
C = Number of Injury C (FDOT database)
POD = Property damage Only value ($) (FDOT database)
AF = Appropriation Factor (FDOT database)

8.2 Comprehensive Fatality and Injury Costs

The latest comprehensive study about fatality and injury costs is based on 1994 data

prepared by Blincoe (1996), which has its roots in the studies prepared by Miller (1991,

1993a, 1993b, 1994, 1995, 1995b, 1995c). The Blincoe study covers 1994 motor vehicle

crashes, where 40,676 people were killed, 5.2 million were injured and 27 million vehicles

were damaged. The estimated cost of these motor vehicle crashes was $150.5 billion.

Death, injury and property damage caused by these crashes were the major

contributions to the financial loss to victims, their families and to society at large.

Included in these costs are: lost productivity; medical costs; legal and court costs;

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EZPTL4Y6U_CNP2HE INGEST_TIME 2013-10-09T21:29:43Z PACKAGE AA00017646_00001