Deterioration Models for Highway Bridges

Sobanjo, John O.; ,

Deterioration Models for Highway Bridges

by John O. Sobanjo, California Dept of Transportation, Los Angeles, United States,

Document Type: Proceeding Paper

Part of: Infrastructure: Planning and Management

Abstract:

Infrastructure management, especially in case of highway bridges, need a realistic and effective deterioration model in order to provide information on the present and future condition of the bridge. This paper reviews and evaluates efforts made to date to develop suitable models. A major shortcoming identified is the unrealistic assumptions of some existing models, not portraying the actual nature of bridge deterioration. Using available bridge data, statistical regression analyses, simple probability analyses, and expert opinions may be used to formulate preliminary models of bridge deterioration process suggests stochastic modeling as a more realistic approach. In a stochastic model, these uncertainties can be adequately quantified or at least to some extent, taken into account in a bridge related decision making process. Existing bridge deterioration model (stochastic) based on the simple Markov chain model is evaluated and its shortcomings presented in terms of first-order assumptions of dependence among the states or bridge condition over time, and also the sojourn times (time spent by a bridge in a particular state.) A higher-order Markov chain model or a lumpable Markov chain model is discussed as a more realistic alternative. Also presented, is the feasibility of stochastic modeling of bridge deterioration through semi-Markov and simulation models.

Return to search