Research On Risk Assessment Of Submarine Pipelines Based On Third Party Damage

As the world’s onshore oil and gas exploitation has entered a period of exhaustion,and the safety,economy,automation of submarine pipelines are high,more and more countries are focusing on offshore oil and gas.However,once the submarine pipeline fails,it will pose a huge threat to the safety of the pipeline and the marine ecological environment.According to the analysis of the PARLOC database,corrosion and third-party damage are the two key factors causing the failure of submarine pipelines.Since third-party damage is more uncertain and difficult to control,it has become the primary cause that affects the normal operation of submarine pipelines.Therefore,this article studies the safety accidents of submarine pipelines caused by third-party damage in offshore areas.The main research contents are as follows:(1)Introduce the Gaussian Mixture Model(GMM)and Probabilistic Neural Network Model(PNN)into the article,and construct the GMM-PNN model to evaluate the risk of oil and gas pipelines.Use the GMM model to improve the shortcomings of the traditional PNN,and stochastic gradient descent method is used to train the parameters of Gaussian function.Finally,an example is introduced and compared with the traditional probabilistic neural network model to verify the effectiveness of the proposed model.(2)Establish a dynamic model for third-party damage risk analysis.Use the method of fault tree combined with event sequence diagram to find dangerous factors,develop fault scenarios,and then transform them into Bayesian network models,and use Boolean operations and historical data to derive node parameters,and finally perform probability update and probability adaptation.Based on reasoning,the Bayesian network model can realize the deficiency that the GMM-PNN model cannot get the failure probability.(3)First,establish a submarine pipeline anchorage failure probability model,which includes calculating the probability of the ship passing through the submarine pipeline area,the probability of the ship anchoring and the probability of the ship anchor hitting the submarine pipeline.A probabilistic calculation model for limit condition of submarine pipeline is established.Analyze the damage degree of pipeline and calculate the conditional failure probability,The limit state of the pipeline is used to analyze the influence of the anchor quality,the size and material of the pipeline,and the distance from the anchor drop point of the pipeline area on the possibility of failure of the submarine pipeline.Finally,the protection measures for the failure risk of the submarine pipeline are proposed.The research results show that the GMM-PNN model proposed in this paper has high accuracy and can accurately grade the third-party damage risk of submarine pipelines.The results obtained by the Bayesian network model are consistent with the actual results.The model constructed in this paper has certain reference significance for the third-party damage risk assessment of submarine pipelines.