Dynamic Optimization Model Based On Hidden Semi Markov Research On The Prediction Of The Remaining Life Of The Submarine Pipeline

With the continuous development of China’s industrial technology,the exploitation of marine resources,as the main means of transportation,had become an important means to obtain natural resources.However,due to the complex submarine environment and the corrosion of pipelines,the medium leakage would cause serious environmental pollution,casualties and incalculable economic losses.In recent decades,the subsea pipeline had entered into the middle and later stage of service,with frequent safety accidents,and the regular inspection and maintenance of subsea pipeline had become an urgent work direction.It was of great social and economic significance to predict the remaining life of submarine pipeline for prolonging its service life,reducing maintenance cost and ensuring industrial production safety.In order to extend the service life of submarine pipeline,this paper aimed to predict the remaining life of submarine pipeline.The main research contents were divided into three parts: the first part was to optimize the hidden semi Markov model in time dimension.In order to conform to the actual degradation rule of submarine pipeline failure,the time-varying matrix was used to replace the original fixed state matrix,and the original matrix was replaced by the constant matrix,multiplier matrix and index matrix respectively,so as to solve the problem that the original model prediction was in a ladder shape and improve the accuracy of remaining life prediction.In the second part,the hidden semi Markov model was optimized in spatial dimension,and the combination model was formed by particle swarm optimization.By asynchronous time-varying method,particle swarm optimization algorithm was improved to make it dynamic.Then,the original parameters of hidden semi Markov model were improved to improve the spatial dynamic classification and optimization ability of the combined model and further improve the prediction accuracy.In the third part,based on the research of life prediction,combined with the technical and economic indicators of submarine pipeline maintenance,the maintenance strategy model was established.The corrosion degree of submarine pipeline was divided into five categories,and the state transition matrix based on different maintenance strategies was formed.The maintenance cost of submarine pipeline was obtained,and the single factor maintenance strategy was optimized.As the direction of submarine pipeline failure risk prediction,this paper mainly used the hidden semi Markov model to study,endowed it with time and space dynamics,and obtained some achievements.The hidden semi Markov model could transform the hidden sequence into the known sequence,and the subsequent research could use this feature to form the combined model combined with other models to continue the research.At the same time,how to make the maintenance index form a dual dynamic maintenance strategy could also be used as a follow-up research direction.