Small and medium size bridge maintenance sequence analysisby optimization technique

摘要

Bridge in traffic transportation sys system play an important role to solve the obstaclernof topography and build up the elevated interchange. Due to the increasing of traffic flow, overrnweighted heavy vehicle and natural disaster from typhoon or earthquake bridge damage arerngetting more seriously threaten the transportation network. But the maintenance budget from thernpublic authority is limited. How to build up the priority of maintenance sequence of therngoverned bridges to make the best cost benefit efficiency is one of the main problems for bridgernauthority. Local government usually lack of budget and expert but face to a lot of small &rnmedium size bridge to keep the daily life and traffic going well between the small town andrnvillages. This paper is intended to find a suitable weight value for priority analysis ofrnmaintenance the small & medium size bridge in a specified area.rnEm At first, we use D. E. R&U Criterion to classify the bridge damage in Degree (D) and at whatrnExtend (E). Then by considering the security and relevancy ? the inspector will evaluate thernUrgency (U) to make the sequence of maintenance for all the bridge in the considered region.rnThis method is officially adapted by the Ministry of Transportation and Communication in Taiwan.rnWe have further developed the model by using Analytic Hierarchy Process (AHP) methodrnthrough the interview with experts of more than 10 years experience in bridge maintenance,rnquestionnaires to determine the appropriate weight of evaluation factors. Comparing thernweight and priority orders with the Back-Propagation Neural Network model (BPN), AHPrnmodel and D.E. R&U we found BPN model as similar as the weight of evaluation factor of D.rnE. R&U method. The weight of evaluation factors with AHP model is higher than the other twornmodels. We consider that because of the debris flow have frequently damage the bridge foundationrnthus the experts have giving more concentrated on the infrastructure of the bridge.rnThe case study is taken basically on the bridge inspection of Nanto County in Central Taiwanrnwhere most of the region is mountain. We have the case of totally 2230 bridges to analyze. Butrnonly 317 bridges inspection data are trained by BPN network.rnTo sum up, the effect of priority analysis that the BPN model is better than AHP and D. E. R&Urnmodels.