Lou, Nassif, Su and Truban 1. Effect of Overweight Trucks on Bridge Deck Deterioration Based on Weigh-In-Motion Data

Transcription

1 Lou, Nassif, Su and Truban Effect of Overweight Trucks on Bridge Deck Deterioration Based on Weigh-In-Motion Data Peng Lou, PhD Candidate * Rutgers Infrastructure Monitoring and Evaluation (RIME) Laboratory Rutgers, The State University of New Jersey, Department of Civil and Environmental Engineering, Frelinghuysen Road, Piscataway, NJ penglou@rutgers.edu Hani Nassif, P.E., Ph.D., Professor and Director Rutgers Infrastructure Monitoring and Evaluation (RIME) Laboratory Rutgers, The State University of New Jersey, Department of Civil and Environmental Engineering, Frelinghuysen Rd, Piscataway, NJ, Tel: --, Fax: :-; nassif@rci.rutgers.edu Dan Su, Ph.D., Visiting Assistant Professor Department of Civil and Environmental Engineering Lamar University, Beaumont, TX, Formerly Post-Doctoral Associate Rutgers Infrastructure Monitoring and Evaluation (RIME) Laboratory Rutgers, The State University of New Jersey, Department of Civil and Environmental Engineering, Frelinghuysen Rd, Piscataway, NJ ; dan.su@lamar.edu Paul Truban, PE, PTOE, AICP, Manager Bureau of Freight Planning and Services, New Jersey Department of Transportation, P.O. Box, Parkway Avenue, Trenton, NJ ; Tel: --, Paul.Truban@dot.nj.gov * Corresponding Author Word count: Abstract: < Figures & Tables: x = Total: < Revised and resubmitted on November,

2 Lou, Nassif, Su and Truban ABSTRACT Highway agencies are responsible for the optimal expenditure of taxpayer dollars allocated to highway infrastructure. Truck size and weight are regulated by federal legislation and every state highway agency has its own legal load limits. Additionally, state agencies issue permits for trucks with gross vehicle weights that are above legal load limits. However, the effect of overweight trucks on the service life of bridge structure especially bridge decks, is not explicitly quantified. In this study, a detailed research on the deterioration models for bridge decks was conducted. Condition ratings of bridge decks in New Jersey from National Bridge Inventory were utilized to derive the deterioration of decks over time and the expected service life of decks on different highways were obtained. Furthermore, weigh-in-motion data from stations in New Jersey was used to extract two datasets: ) all trucks and ) legal trucks. All trucks dataset was used to develop a deck deterioration model as a function of equivalent wheel load that could be used to estimate the expected service life. Lastly, Bridge Life Cycle Cost Analysis was conducted on two contrasting scenarios: one with and the other without overweight trucks to quantify their economic impact on bridge decks. The results show that overweight trucks caused more damage on NJ state highways instead of interstate highway due to a larger proportion of overweight trucks, heavy wheel loads from overweight trucks, and fewer axles per truck. Key Words: Overweight Trucks WIM Database NBI Database Deck Deterioration Model Life-Cycle Cost Analysis

3 Lou, Nassif, Su and Truban INTRODUCTION The Federal Bridge (FB) Formula, shown as Equation, was enacted by Congress in and updated in. It has been used by many local agencies to legislate and enforce the legal truck weight limits. LN Equation W N N where W is the overall gross weight on any group of two or more consecutive axles to the nearest pounds, L is the distance in feet between the outer axles of any group of two or more consecutive axles, and N is the number of axles in the group under consideration. In addition to the FB formula, any single axle is limited to, pounds, a tandem is limited to, pounds, and the gross vehicle weight is limited to, pounds. In New Jersey, the weight regulation follows the Federal Bridge Formula. Although the truck weight is regulated by legislature, permits are issued for trucks with irregular dimensions and heavy loads. There is a need to evaluate the effect of loads exceeding legal limits on the bridge components. Bridge deck is a component that usually undergoes more deterioration than any other due to its direct exposure to heavy and more frequent truck traffic, environmental conditions, and de-icing salts. Therefore, the deterioration for bridge decks could be quite complicated. Previous studies (,, ) have demonstrated that the transverse cracks and water penetration during service decreased the ultimate punching shear and fatigue strengths of the concrete decks. However, the interaction between the deck deterioration and the overweight loading were not quantified explicitly. The current AASHTO Manual for Bridge Evaluation () indicates that the de-icing salts and truck traffic volume affect the deck deterioration rate, but the current practical evaluation of bridge decks is limited to visual inspections only. In reality, a combination of mechanical loading and environmental factors would lead to the end of service life. Moreover, due to the increasing truck traffic both in weight and in frequency on the highway network, there is a need to correlate the effect of overweight trucks with concrete deck deterioration. FIGURE shows the changes of average daily truck traffic (ADTT) over years on I- (Site ) and I- (Site ) as well as deck condition rating over time for a bridge near WIM Site. The ADTT were collected from WIM sites in New Jersey and the deck condition ratings were taken from National Bridge Inventory (NBI) database. It is clearly shown that while ADTT is increasing over the years, the bridge deck rating is decreasing over the same period. This study, which is part of an ongoing project sponsored by the New Jersey Department of Transportation (NJDOT), was performed to evaluate the relationship between truck weight and the accumulated damage they would induce on concrete decks.

4 ADTT Lou, Nassif, Su and Truban Site Site Deck Rating Linear (Site ) Linear (Site ) Condition Rating FIGURE ADTT of interstate highway and deck deterioration Deck Deterioration from Mechanical Loading and Environmental Factors In AASHTO LRFD Bridge Design Specification (), two limit states are considered for bridge deck design, service I limit state that controls the crack width of reinforced concrete deck, and strength limit state that controls the flexure capacity of the concrete deck. However, previous studies (,, ) indicated bridge decks subjected to concentrated loads does not fail by flexure, as traditionally believed but by fatigue. Starting from late s, laboratory tests were performed to investigate the failure modes of reinforced concrete decks as a whole. The test results showed that the fatigue of reinforced concrete deck is governed by the punching shear failure of concrete. From available test data (,, ), all the fatigue models for bridge decks yielded models such as in Equation. The equivalent wheel load is regarded as an important load parameter for deck fatigue. Since the rear wheel usually has dual tires with a print of inch by inch while steering wheel is usually a single tire, the value for steering wheel weight was increased by /.. N pf Year P alog Pu b Equation. /. i i i P f p p where Pu is the ultimate strength of the deck, Npf is the number of cycles to failure, a and b are the parameters of regression line, P is the equivalent wheel load from wheel weight distribution, pi is the value of wheel weight (kips) in wheel weight distribution, and fi(pi) is the frequency for that weight.

5 Lou, Nassif, Su and Truban In current AASHTO Bridge Design Specification, in order to avoid water and chemical penetration and corrosion of steel reinforcement, the cracking of concrete deck is controlled by limiting the spacing of reinforcement under a limitation that is a function of exposure condition, tensile stress in reinforcement, and thickness of concrete cover. In addition, extensive laboratory work also proved that the crack width of concrete deck at service limit state is proportional to steel stress with significant variation. Another major concern about the deck deterioration comes from the chloride-induced corrosion of steel reinforcement bars, specifically for the bridges that receive chlorides through applications of de-icing salts, or marine environment. The corrosion will lead to a reduction in the cross-sectional area of the reinforcing steel or loss of bond, which may further lead to the loss of strength and unserviceability. Various models (,,,) of this type of failure are developed based on the mechanism of the diffusion of chlorides through the protective concrete cover, showing that the corrosion will be initiated once the chloride concentration exceeds a threshold. In addition, cracking, spalling, and delamination induced by mechanical loading accelerates the attack of chloride and prompt deck maintenance helps preserve the deck from corrosion. Bridge Life Cycle Cost Analysis The Bridge Life Cycle Coast Analysis (BLCCA) has been required by current regulatory. It is important for infrastructure investments including highway bridge program (). In BLCCA, the costs are paid by either the agency or user. The agency costs are usually the direct expenditures of funds for planning, design, construction, operation, and maintenance of a bridge. User costs consider a broader category including vehicles, business and residents nearby which rely on the bridges for access or transportation. In this study, only agency cost was considered. There are several economic indicators related to agency cost such as Net Present Value (NPV), Equivalent Uniform Annual Costs (EUAC) and Salvage Value (or residual value). NPV converts all costs to a single base year costs. For assets having useful life remaining at the end of analysis period, a salvage value should be estimated. After converting all agency costs to NPV or EUAC, the costs of various investment options can be compared. The expressions for mentioned indicators are shown in Equation, Equation and Equation. A cash flow diagram for BLCCA corresponding to the deterioration of decks is shown in FIGURE. T Ct NPV t Equation t r where Ct is the cost occurring at year t due to rehabilitation activity, T is the lifetime of the project or analysis period (years), r is the discount rate. T r( r) EUAC NPV T Equation ( r) where EUAC is the Equivalent Uniform Annual Costs converting from NPV. L A SV C Equation LE

6 Lou, Nassif, Su and Truban where SV is the salvage value of rehabilitation activity, C is the cost due to rehabilitation activity, LA is the analysis life of rehabilitation in years and LE is the expected life of the rehabilitation. Deck Serviceability Deck Service Life Analysis Period Tolerant Serviceability (Condition Rating) Remaining Life Agency Cost Initial Construction Cost (material, labor and machinery) Agency Future Rehabilitation Cost (material labor and machinery) Time (Year) Agency Future Maintenance Cost Salvage Value FIGURE Deck deterioration and cash flow diagram for life cycle cost analysis In light of the above discussions, wheel load is regarded as an important parameter that is affecting the fatigue behavior and stress level of concrete decks. This paper presents the results of an on-going study that used WIM data from WIM stations in New Jersey to extract various truck traffic information including ADTT, axles per day, average axles per truck, and equivalent wheel load for truck populations with and without overweight trucks. Meanwhile condition ratings from NBI database were employed to validate the proposed deck deterioration model. Then the expected service life of decks was obtained for available highways by assuming service life ends when condition rating downgrade to four. Correlation between equivalent wheel load and expected service lives on major highways was quantified to formulate the effect of wheel loading on the service life of bridge decks. At last, BLCCA was performed for two scenarios and the economic impact of overweight trucks on bridge decks was summarized. DATABASE In order to evaluate the effect of overweight trucks on deck deterioration, both deterioration of the deck and load effects need to be considered. Therefore, two databases were utilized to reflect the deterioration and load variation during the service life of bridge deck: ) NBI database storing condition data and ) WIM data representing load effect. NBI Database The NBI database has the most extensive and detailed data on highway bridges in the US. The NBI is a collection of information and general condition covering all of the nation's bridges located on public roads as well as publicly accessible bridges on federal lands. The deterioration

7 Lou, Nassif, Su and Truban level of bridge components including deck, superstructure, and substructure, is quantified using condition rating indices in NBI. This is a numeric ranking system from to, where represents Failed Condition and represents Excellent Condition. In this study, the data from Year to Year for bridges located on three major types of highway including interstate, US numbered and NJ state highways were collected and used for the development of deterioration models. The NBI data also includes the information about the wearing surface type, type of membrane, and deck protection type. WIM Data The New Jersey Department of Transportation (NJDOT) has installed WIM sites throughout the state to monitor long-term trends in truck volumes and weights as shown in FIGURE. One WIM system includes one or more instrumented lanes, a WIM data logger, and a permanent enclosure. The data from these sites is used for pavement design, long-term freight planning, and enforcement. The functional classification of the sites encompasses two-lane country roads, urban arterials, and major interstate highways. The duration of available data varies by site depending on the installation date. FIGURE Map of WIM sites in New Jersey ()

8 Lou, Nassif, Su and Truban DATA PROCESSING AND RESULTS Deterioration of Bridge Decks on Various Highways in New Jersey The actual deterioration curve of a bridge deck is usually implicit and complicated due to both mechanical and environmental effect. This study assumed that the downgrade of deck condition rating represents the deterioration of the bridge deck. The expected trend in this deterioration process is that if there is no improvement made to a bridge member, its condition rating either remains the same or falls to a lower value as the bridge ages. Although there is bridge reconstruction information in the NBI database, there are still lots of unrecorded repair or reconstruction activities on the bridge members due to the sudden increase in condition rating history (,,). In order to obtain a complete life cycle of deck rating curve without the interrupt of repair event, bridge data with unrecorded repair needs to be removed. The flowchart shown in FIGURE describes this data validation procedure. The deterioration data is only captured once when the rating of a bridge is downgraded. Afterward, a criteria developed by Morcous () for the maximum and minimum age number for each condition rating was applied here to remove the outlier. Regression analysis was performed for the validated condition rating data on each highway and method of least squares was used to estimate the parameters. It was found that a third order polynomial regression was the best fit with highest R-squared over linear, second order polynomial, exponential, logarithmic, and power regressions. The regression model used in this study was shown in Equation. CR M M x M x M x Equation where CR is the bridge condition rating, and regressor x is the age of deck in years. M, M, M, and M are the parameters, which are summarized in TABLE. As examples, condition rating data were plotted over age as well as the fitted curves for I-, US-, and NJ- to represent the three types of major highways. FIGURE also shows the deterioration curves of bridges deck on different highways. Previous studies concluded that bridge decks usually experienced replacement when the condition rating downgraded to four (,,). In addition, based on repair events from the bridges on three types of highway, deck repair events were mostly taking place when rating has reached level four (.%). The expected service life of bridge decks on each highway is assumed at the age of deck when the condition rating downgraded to four. The expected service life of decks for each highway is summarized in TABLE. The average expected service life of bridge decks on interstate highways, US numbered highways, and NJ state highways are.,., and. years respectively. The bridge decks on interstate highway deteriorated with the highest rate while the decks on NJ State Highway deteriorated the least.

9 Lou, Nassif, Su and Truban No Next Structure Condition Ratings Data after Latest Rehabilitation Year Yes Sort W.R.T Structural Number and then Year Next Year Rehabilitation History (Year of Built or Replacement) Sort W.R.T Year of Rehabilitation Latest Rehabilitation Year Next Structure No, Next Year Rating Increased from Last Year No Rating Downgraded Output for Analysis Eliminate Data Yes Yes Yes No Unrecorded Rehabilitation Deterioration Data (Rating and Age) kept Satisfy Criteria FIGURE Flowchart for Processing Deterioration Data

10 Condition Rating Condition Rating Condition Rating Condition Rating Condition Rating Condition Rating Lou, Nassif, Su and Truban I-(R=.) I-(R=.) I-(R=.) I-(R=.) I-(R=.) I-(R=.) I-(R=.) I-(R=.) US(R=.) US(R=.) US(R=.) US(R=.) US(R=.) US(R=.) US(R=.) US(R=.) US(R=.) US(R=.) NJ-(R=.) NJ-(R=.) NJ-(R=.) NJ-(R=.) NJ-(R=.) I-(R=.) Data Third order polynomial I- Deck Age (Years) US- Data Third order polynomial Deck Age (Years) NJ- Data Third order polynomial Deck Age (Years) Deck Age (Years) FIGURE Comparison of deterioration models for bridges on various highways

11 Lou, Nassif, Su and Truban TABLE Summary of regression models for each highway Regression Model Summary Highway M M M M R Expected service life I I I I I I I I US US US US US US US US US US NJ NJ NJ NJ NJ NJ Data Processing of WIM Data The WIM data that used was collected from various WIM sites the operated by NJDOT. The raw data contains all traffic including cars and trucks and a significant amount of erroneous data. A refined data processing program was proposed as shown in FIGURE to extracted two datasets: all trucks dataset, and legal trucks dataset. All trucks dataset reflects the actual truck loading on the bridges can be used to correlate with the expected service life obtained from previous section. Then the service life prediction functions based on wheel weight could be obtained. Legal trucks dataset was utilized to predict the service life of deck under legal trucks traffic without overweight trucks. By comparing the service lives of deck under these two truck

12 Lou, Nassif, Su and Truban traffic conditions, the reduction in service life of bridge decks can be calculated. FIGURE shows the comparisons of ADTT, axles per day, effective truck weight, equivalent wheel load, average axles per truck, and proportion of overweight trucks over all trucks on the three types of highways. It is found that interstate highways have significantly higher ADTT and axles per day comparing to the other two highway types. Meanwhile the average axles per trucks on interstate highway were found to be higher than the other two. In addition most overweight trucks on interstate highways are Class trucks with five axles while those on US numbered highway and NJ state highway are Class trucks that has four axles. It is important to note that on average NJ State highway has the highest proportion of overweight trucks. As shown in FIGURE (e) and (f), although the interstate highway has higher effective truck weight than the other two, all three highway types have comparable equivalent wheel loads due to the fact that trucks on interstate highways usually have more axles. The highest equivalent wheel load of. kips was observed on US-. Given the similar wheel load level, the reason that decks on interstate highways deteriorated with the highest rate can be attributed to the highest values of ADTT and axles per day. WIM Data No Cars Unclassified Truck Yes Erroneous Data No Yes Legal Trucks Overweight Trucks No Yes Start Data Processing Yes GVW> kips Class= No Axle Spacings >.ft Yes If Class, S>ft and GVW <kips FIGURE Flowchart for WIM data Processing No Truck Data (All) NJ Overweight Vehicle Criteria

13 Lou, Nassif, Su and Truban ADTT Axles Per Day Interstate US Numbered NJ State (a) Interstate US Numbered NJ State (b) Average Axles Per Truck... Proportion of Overweight Trucks (%) Interstate US Numbered NJ State (c) Interstate US Numbered NJ State (d) Effective Truck Weight (kips) Equivalent Wheel Load (kips) Interstate US Numbered NJ State Interstate US Numbered NJ State (e) (f) FIGURE Statistics of all trucks dataset from WIM data

14 Lou, Nassif, Su and Truban CORRELATION BETWEEN TRUCK LOADING AND DECK SERVICE LIFE ADTT and axles per day are considered important factors that affects the service life of bridge decks since they indicates the frequency of loading on bridge decks. Datasets that have the comparable wheel weight were extracted and plotted in FIGURE. Different types of highways are treated separately in order to exclude the effect of highway types. Each line in the plot shows that a higher value of axles per day corresponds to lower service life. Comparing two lines in the same highway, lower wheel load levels correspond to higher service life, when the number of axles per day is held constant. Therefore, both parameters play roles in determining the service life of decks. In order to consider both parameters, the capacity of bridge decks herein was defined as the lifetime axle count, NA, which represents the total number of axles passing the bridge over the predicted service life span as shown in Equation. The lifetime axle count was plotted versus the equivalent wheel load in FIGURE, and linear regressions were performed for three highway types with the method of least squares. Note each data point represents the expected lifetime axle counts on one highway on which the decks are assumed to be under same equivalent wheel load. The total number of bridges considered are,, and for interstate, US numbered, and NJ state highway respectively. The R-squared for three linear regression are.,., and. respectively, for interstate, US numbered, and NJ state highways. It is found that service life of decks on interstate highways and NJ state highways are more sensitive than on US numbered highways. Based on the correlations, prediction functions are proposed for lifetime axle count based on equivalent wheel load as Equation. If service life and axles per day are taken into account, the functions can be solved as Equation. Note these functions are data driven models and works for interpretation. In addition, beyond certain point, the service year of deck is not govern by the wheel loads. Predicted service life of bridge deck is visualized in FIGURE. The annual traffic increase and average axles per truck are taken as.% and.,.% and., and.% and. for interstate, US numbered, and NJ state highway respectively. y N A APDi** r i Equation where NA is the lifetime axle count, APDi is average axles per day at year i, y is service life in years predicted, r is annual truck traffic growth. Interstate NA.. P R =. US Numbered N.. P R =. A NJ State N.. P R =. A Equation where P is the equivalent wheel load. NA d log ADTT APT y log d Equation

15 Equivalent Wheel Load (kips) Expected Service Life (years) Lou, Nassif, Su and Truban where d is the annual truck traffic increase. Axles/Day Interstate (.<P<) Interstate (<P<.) US Numbered (<P<) US Numbered (<P<) NJ State (.<P<.) NJ State (<P<) FIGURE Effect of axles per day on service life of bridge decks Interstate NJ State Linear (US Numbered) US Numbered Linear (Interstate) Linear (NJ State) Interstate ( bridges) US Numbered ( bridges) NJ State ( bridges).e+.e+.e+.e+.e+ Expected Lifetime Axle Counts FIGURE Correlation between lifetime axle counts vs equivalent wheel load

16 Predicted Service Years Predicted Service Years Predicted Service Years Predicted Service Years Lou, Nassif, Su and Truban ADTT= ADTT= ADTT= ADTT= ADTT= ADTT= P= P= P= P= P= Equivalent Wheel Load (kips) (a) Interstate highway ADTT ADTT= ADTT= ADTT= ADTT= ADTT= ADTT= Equivalent Wheel Load (kips) (b) US numbered highway P= P= P= P= P= ADTT

17 Predicted Service Years Predicted Service Years Lou, Nassif, Su and Truban ADTT= ADTT= ADTT= ADTT= ADTT= ADTT= Equivalent Wheel Load (kips) P= P= P= P= ADTT (c) NJ state highway FIGURE Predicted service life of deck

18 Lou, Nassif, Su and Truban ECONOMIC IMPACT OF OVERWEIGHT TRUCKS ON BRIDGE DECKS After the prediction function was developed for the service lives of bridge decks, two scenarios were considered in this study to quantify the economic impact of overweight trucks: Case all trucks which represents current truck traffic with all overweight trucks and Case legal truck traffic only without overweight trucks. BLCCA was performed for both scenarios. The annual maintenance costs are assumed the same for both scenarios. An analysis period of years was used. The deck replacement cost used here is dollar per square foot (). Percentage increases in annual truck traffic are assumed as.,., and. for interstate highways, US numbered highways, and NJ state highways, respectively. The discount rate is assumed as percent. The results of BLCCA were summarized in TABLE. EUAC is annual deck cost per deck area in square feet due to overweight trucks. The unit cost is the cost of unit weight of overweight. It is in dollar per deck square feet per kip. Note that weight of overweight here is the marginal weight of overweight trucks above the legal weight. Boxplots of service life reduction in percent and Life Cycle Cost Premium in percent for three types of highways are shown in FIGURE. The overweight trucks induced more costs on decks of NJ state highways than the other two. This can be attributed to: ) NJ state highways have the highest proportion of overweight trucks compared to the other two, ) overweight trucks introduced much heavier wheel loads than the legal trucks did, and ) the average number of axles per truck on NJ state highways is relatively less. Service Life Reduction (%) Life Cycle Cost Premium (%) Interstate US Numbered NJ State Interstate US-Numbered NJ State FIGURE Effect of Overweight Trucks

19 Lou, Nassif, Su and Truban TABLE Summary of life cycle cost analysis Highway Service Life in years () Case () Case () Annual Weight EUAC in dollar () Service Life in years () EUAC in dollar () of Overweight in kips () Service Life Reduction [()-()]/() Unit Cost [()-()]/() I-...E+ %.E- I-...E+ %.E- I-...E+ %.E- I-...E+ %.E- I-...E+ %.E- I-...E+ %.E- I-...E+ %.E- I-...E+ %.E- US-...E+ %.E- US-...E+ %.E- US-...E+ %.E- US-...E+ %.E- US-...E+ %.E- US-...E+ %.E- US-...E+ %.E- US-...E+ %.E- NJ-...E+ %.E- NJ-...E+ %.E- NJ-...E+ %.E- NJ-...E+ %.E- NJ-...E+ %.E- NJ-...E+ %.E- NJ-...E+ %.E-

20 Lou, Nassif, Su and Truban CONCLUSIONS This paper utilized condition ratings from NBI database to derive the deterioration models of bridge decks over time. Additionally WIM data from stations spread over New Jersey highway network were used to extract truck loading statistics on bridge decks. Correlation between truck load statistics and predicted service life on available major routes was quantified. Lastly, BLCCA was performed for cost premium of overweight trucks. The results of this study lead to the following conclusions:. Based on the average expected service life of bridge decks on three types of major highways, the bridge decks on interstate highways deteriorated the most while those on NJ state highways deteriorated the least.. Compared to US numbered highways and NJ state highways, interstate highways have highest ADTT and axles per day. This is the reason that decks on interstate highways have shorter service lives. Although interstate highways have higher truck weight than the other two highway types, the equivalent wheel loads are comparable for all three highway types since the trucks on interstate highways usually have more axles. More axles helps distribute the GVW. The highest equivalent wheel load is found to be on US- rather than on interstate highways.. NJ state highways have the highest proportion of overweight trucks. Most overweight trucks on interstate highways are Class with five axles. For the other two highway types, the majority of overweight trucks are Class with four axles.. Service life prediction functions for decks were proposed based on equivalent wheel loads, helping to quantify service life reduction due to increased wheel load.. Overweight trucks caused more marginal damage on NJ state highways due to the presence of a larger proportion of overweight trucks, larger wheel loads from overweight trucks, and fewer axles per truck. FUTURE RESEARCH The de-icing chemical treatment practices, maintenance practices, and environmental conditions are considered to have effects on the rate of deck deterioration. Meanwhile, different the wearing surface types, membrane types, and deck protection types would also matter in predicted service life. A more sophisticated deck deterioration model to include these factors. Since the condition rating data from NBI is discrete and ordinal, a more rational ordinal regression analysis is desired for this type of data. ACKNOWLEDGEMENTS This study is sponsored by NJDOT. The help and support of the NJDOT staff, Paul Thomas, and Andrew Ludasi are greatly appreciated. The help and support from graduate student Miguel Beltran is also appreciated.

21 Lou, Nassif, Su and Truban REFERENCES. Azad, A. K., Baluch M.H., Al-Mandil M.Y., and Al-Suwaiyan. Static and Fatigue Tests of Simulated Bridge Decks. Experimental Assessment of Performance of Bridges, Proceedings of ASCE Convention, Boston, Mass., Oct., pp. -.. Okada, K., Okamura H., and Sonada K. Fatigue Failure Mechanism of Reinforced Concrete Deck Slabs. In Transportation Research Record, TRB, National Research Council, Washington, D.C.,, pp.-.. Kato, T., and Goto Y. Effect of Water Infiltration of Penetration Deterioration of Bridge Deck Slabs. In Transportation Research Record, National Research Council, Washington, D.C.,, pp. -.. AASHTO. The Manual for Bridge Evaluation... AASHTO. LRFD Bridge Design Specifications... Batchelor, B., Hewitt B.E., and Csagoly, P. An Investigation of the Fatigue Strength of Deck Slabs of Composite Steel/Concrete Bridges. In Transportation Research Record, TRB, National Research Council, Washington, D.C., Vol.,, p.-.. Fang, I. K., Tsui C. K. T., Burns N. H., and Klinger R. E. Fatigue Behavior of Cast-in-Place and Precast Panel Bridge Decks with Isotropic Reinforcement. PCI Journal, Vol., No., May-June, pp.-.. Petrou, M., Perdikaris P. C., and Wang A. Fatigue Behavior of Non-composite Reinforced Concrete Bridge Deck Models. In Transportation Research Record, TRB, National Research Council, Washington, D.C.,, pp.-.. Perdikaris, P.C., Petrou M.F., and Wang A. Fatigue Strength and Stiffness of Reinforced Concrete Bridge Decks, Final Report FHWA/OH-/, Ohio Department of Transportation, March.. Youn, S., and Chang S. Behavior of Composite Bridge Decks Subjected to Static and Fatigue Loading. ACI Structural Journal,, Vol., No., pp.-.. Weyers, R. E., Fitch, M. G., Larsen, E. P., Al-Qadi, I. L., Chamberlin, W. P., and Hoffman, P. C. Concrete Bridge Protection and Rehabilitation: Chemical and Physical Techniques. Service Life Estimates (No. SHRP-S-),.. Liu, Y., and Weyers, R. E. Modeling the time-to-corrosion cracking in chloride contaminated reinforced concrete structures. ACI Materials Journal,, ().. Stewart, M. G., and Rosowsky, D. V. Time-dependent reliability of deteriorating reinforced concrete bridge decks. Structural Safety,, (), -.. Vu, K. A. T., & Stewart, M. G. Structural reliability of concrete bridges including improved chloride-induced corrosion models. Structural safety,, (), -.. Hawk, H. Bridge life-cycle cost analysis (No. ). Transportation Research Board,.

22 Lou, Nassif, Su and Truban. NJDOT. Weight-in Motion (WIM) sites. Truck Weight Monitoring Program,. Accessed Nov.,.. Saito, M., and Sinha, K. C. The Development of Optimal Strategies for Maintenance, Rehabilitation, and Replacement of Highway Bridges: Volume -Priority Ranking Method: Executive Summary and Final Report,.. Bolukbasi, M., Mohammadi, J., and Arditi, D. Estimating the future condition of highway bridge components using national bridge inventory data. Practice Periodical on Structural Design and Construction, (), -.. Hatami, A., and Morcous, G. Developing deterioration models for Nebraska bridges (No. M),.. NJDOT Structural Evaluation. Bridge re-evaluation survey report Structure No. -, Cycle No.,.