Bridge Corrosion Cost Model Implementation and Coating Maintenance Model Using Dynamic Programming

This second of two papers studying the minimization of the cost of coating maintenance for steel bridges deals with two main topics: the computer implementation of a life-cycle cost analysis (developed in the first paper) in a spreadsheet environment using equivalent annual costs to compare three single maintenance strategies—spot repair, overcoat, and recoat; and the development of a coating maintenance model using the dynamic programming approach that can optimize combinations of maintenance procedures for minimum cost. The flexibility of the life-cycle cost spreadsheet implementation allows the user to adjust key parameters in order to account for variability in costs and environmental conditions between different regions in British Columbia. The program can be easily adapted for different climatic conditions. In addition, an on-line “Help” feature is provided to reduce the time needed to operate the program. The computer application and demonstrative example are described here and the source code is available on request. The second evolved coating maintenance model using the dynamic programming approach presented here minimizes the sum of the costs resulting from a sequence of rehabilitation choices (a combination of maintenance strategies). It is a more flexible analysis technique than the equivalent annual cost approach and can determine the optimal sequence of rehabilitation activities over the service life of the structure. The procedures developed in this analysis can be easily converted into algorithms for possible computer applications. Based on results from a preliminary analysis of in-service bridges with good coating condition assessments using the life-cycle cost and dynamic programming approaches, spot repair is the most cost-effective rehabilitation method. Overcoating is the second most effective strategy and recoating is usually the most expensive solution. However, the difficulty in obtaining adequate deterioration and cost functions for the coating systems will significantly reduce the accuracy of these analysis techniques. Therefore, a uniform data-collection system should be implemented so that a database will be available for these models.