State-Based Stochastic Performance Deterioration Modelling of Flood Defence Assets

Flood defence assets are aging, and their performance in terms offlood prevention deteriorates withtime. In order to develop effective flood riskmanagement strategies and flood alleviationprogrammes, accurate current condition state and future performance predictions are essential forthe safe operation of the assets. This paper presents a state-based probabilistic performancedeterioration model to predict future states ofaging flood defence assets, where condition gradesand deterioration rates are utilised. The stochastic process e.g. Markov chains is employed in thisstudy to investigate the probability distribution ofvarious condition states during the assets’ lifecycle. Condition grade data obtained from visualinspection of flood defence assets are adopted toestimate the transition probabilities by using nonlinear optimization technique. The estimatedtransition probabilities are then utilised to evaluate the future condition of the assets and examinedby using statistical hypothesistest. Finally, the effectiveness of the proposed method isdemonstrated by a case study, and the results showthat the proposed method can provide reliablepredictions for future performance ofthe aging flood defence assets.