MPC
Research Projects (2008-09)

Identifying Number

MPC-311

Project Title

Forecasting Bridge Deterioration Rates and Improvement Costs

University

North Dakota State University

Project Investigators

Subhro Mitra, Denver Tolliver and Kurt Johnson

Description of Project Abstract

Including box culverts, North Dakota has over 5,000 bridges located on state and county highways. Forty-two percent of these bridges are more than 40 years of age.

In a recent study, NDDOT estimated that bridge investment needs will exceed $525 million during the next 20 years on the state highway system alone. This estimate does not include all maintenance needs. An in-progress study indicates that roughly $400 million will be needed for county bridges during the next 20 years. Both of these studies use linear approximations for bridge deterioration rates. The studies indicate the magnitude of future investments and the need for improved deterioration models, as well as the need for improved bridge management information. The research described in this problem statement is especially important when increasing materials costs and limited bridge funding are considered.

The results of this project will be: (1) a set of deterioration functions for various types of bridges and elements that consider the effects of environment, traffic, and maintenance treatments; (2) a set of probability transition matrices that forecast the likelihood of specific bridge elements transitioning from one structural condition rating to another over time; (3) the comparative costs and effectiveness of various types of interim improvements to decks and other elements; and (4) an improved decision management tool integrated with Pontis. A related objective is the integration of various data used by NDDOT to enhance future bridge management. Finally, the results will be transferred to counties through summary procedures distributed through the North Dakota LTAP Center.

Project Objective

The immediate research objectives are to: (1) provide improved bridge forecasting models that consider the effects of environment, traffic, and maintenance treatments; (2) a set of probability transition matrices that can be used with Pontis to forecast the likelihood of bridges of certain types and elements transitioning from one structural condition rating to another over time; (3) the integration of all bridge data sources with Pontis to provide a comprehensive asset management system; and (4) an update of the recently-conducted bridge investment studies to include longer planning horizons and improved deterioration functions. The long-term objective for the SAFETEA-LU period is to achieve (through a sequence of research projects) a fully functional bridge life-cycle cost and asset management system.

Project Approach/Methods

The initial task will include the completion of a literature review and survey of other states in the region to determine their uses of Pontis, the bridge deterioration functions they use, and how these functions were estimated. Next, an assessment will be made of all available bridge data (including worksheets and notes). A plan will be devised to utilize and integrate these data sources. At a minimum, the project database will include: (1) detailed sub element condition ratings from 1996 onward; (2) NBI data from 1989 to the present; and (3) internal worksheets and notes about changes and improvements to bridges over time.

The outcome of the data development task will be a panel dataset that includes both cross-sectional and time-series bridge data (for 11 years). The data will be analyzed using autoregressive and other modeling techniques designed to exploit panel datasets. Multivariate statistical models will be estimated for specific bridge types and elements, as a function of time, traffic load, environment, maintenance treatments, and other variables. In addition, piecewise deterioration functions will be estimated for each type of bridge and element. Because only 11 years of data exist, deterioration functions will estimated within age categories to derive piecewise functions. Other forms of deterioration models will be investigated. Chemical exposure is a factor that must be considered, even though information is not routinely collected to model these effects.

Although statistical models will be estimated in this project, the models will be supplemented with expert engineering judgment and engineering performance models (e.g., Virtis). A panel of bridge engineering experts will be assembled to assess the “real world” fit of the deterioration functions and results.

As noted above, a particular issue is the effect of chemicals used for deicing. This factor may be analyzed in several ways. One way is a case-study method in which estimates of chemical applications to specific bridges are compiled over time from disparate records, interviews with transportation department personnel, and typical area-wide application levels. Because of the time-consuming nature of this task, several representative highway districts will be selected for case study. If possible, estimates of chemical effects will be included in the bridge deterioration models. Other methods of modeling this type of deterioration may also be considered.

MPC Critical Issues Addressed by the Research

Infrastructure Longevity (#16) and Integrated Asset Management Systems (#19)

Contributions/Potential Applications of Research

This project is the first in a series intended to provide enhanced bridge management tools to NDDOT and the region. Improved (and more specific) deterioration and cost models are prerequisites for future projects. The series will eventually result in a fully functional bridge life-cycle cost and asset management system. The proposed project will provide immediate benefits to NDDOT by enhancing the inventory capabilities of Pontis; integrating multiple data sources; and providing improved forecasting and needs assessments. Moreover, the project will include a survey of practice in the region and provide other state transportation departments with useful comparative information for assessing bridge deterioration rates and improvements. The statistical modeling approaches may enhance the capabilities of Pontis by providing improved probability transition matrices and a better understanding of the effects of individual factors through multivariate regression analysis.

Technology Transfer Activities

The report will provide a family of deterioration functions that can potentially be integrated with Pontis; as well as regression parameters and confidence bands that describe the effects of various factors (load, climate, etc.) on deterioration rates. The project report will include a survey of practice in the region and identify potential improvements in the utilization of Pontis. In addition, a technology transfer package will be developed for counties—which will include spreadsheet deterioration functions, as well as procedures that utilize the NBI. The results will be disseminated via the Transportation Learning Network and LTAP Centers in the region.

Time Duration

July1, 2008-December 31, 2009

Total Project Cost

$182,139

MPC Funds Requested

$120,000

TRB Keywords

Pontis, bridge deterioration, bridge management, asset management