Modeling And Algorithm Of Risk-based Maintenance And Life-cycle Cost In Bridge

With the continuous development of China's transportation,the size of the bridge construction gets a real big increase.However,what should be taken seriously is that the lack of management and maintenance of bridge,which lead to a large number of bridge performance degradation.And this will even endanger the safety of the use of structures over time.Now there is a problem that every deteriorated states of the bridge have a certain maintenance strategie to be chosen,and each of the maintenance policy is bound to have different effects on the performance of the structure.Therefore,any single maintenance strategy should not be considered in isolation,but should be integrated within the entire life cycle of the structure.On the other hand,any kind of maintenance strategy will lead to a different cost,so a poor decision will result in huge financial losses.Meanwhile,the way we select to maintenance is based on the structural condition assessment.The detection information determined by the state of structure is often incomplete and there is some uncertainty in it,which gives some risk to the maintenance decision.In view of this,the Soto steel truss bridge is considered in this paper for using the reliability theory to evaluate the reliability of it.And then we use the Markov Decision Process to construct the model of its maintenance process in the whole life of the bridge.Two factors are mainly considered in the deterioration of the structure which are fatigue and corrosion.We use the fuzzy evaluation and inverse matrix method to derive the transition matrix of this structure states.Finally,the enumeration method,linear algorithms and policy improvement algorithm are analyzed and compared to found that although the results got by the enumeration method is accurately and the linear algorithm can make the model build speed,but they also have a huge amount of computing and the defects subjective judgement in calculation process.The policy improvement algorithm has been simplified and it is suitable for the search of bridge maintenance strategies in practical engineering.