Application Of Big Data Method In Riks Analysis Of Pressure Piping

In this paper,in order to solve the problem that the quantitative RBI method has a wide variety of data and a long period of analysis in the risk analysis,through the study of pressure pipeline in a petrochemical atmospheric and vacuum distillation unit,put forward a big data method to analyze the risk of pressure pipeline,introduces the risk based inspection and calculate the risk of all the pressure pipes.It constructs the prediction model of corrosion rate,failure probability and failure consequence of pressure pipeline and determines the variable data that affects the importance of the model,and verifies the model prediction results with the risk results calculated by the quantitative RBI technology.The main work of this article is as follows:1.Put forward a big data analysis model of pressure pipeline,which includes six parts:data generation,data understanding,data processing,data analysis,model evaluation and release model.Introduce the method of data processing and data analysis.Through research and analysis,it is found that neural network algorithm,C5.0 algorithm and CHAID algorithm can be used to construct the corrosion rate,failure possibility and failure consequence model of pressure pipeline.2.The process of using the quantitative RBI method to calculate the risk of pressure pipes is explained.The factors of the failure possibility and consequences are elaborated in detail,and the corrosion rate of the calculation of the results of the risk analysis is introduced.The risk analysis of all pressure pipelines in a petrochemical atmospheric and vacuum unit is analyzed with the RBI module of PCMS software.The risk matrix diagram is obtained,and the data of the pressure pipeline risk analysis by big data method are generated.3.In view of the importance of the corrosion rate in the risk analysis of pressure pipes,it is different from the expert advice and the measured corrosion rate.A new method to determine the high risk pressure pipeline through the big data to predict the corrosion rate is proposed.The principal component analysis method is used to extract 4 factors that affecting the corrosion rate of pressure pipes.The importance of each factor and the representative variables are analyzed,and the equation of 4 factors is given.The BP neural network algorithm is used to construct a big data model for predicting the corrosion rate of pressure pipes,and the importance of each factor in the model is discussed.And compared with the high corrosion rate pipeline of validation set,the good prediction results are obtained.4,Understand the data generated by quantitative RBI analysis of pressure piping of atmospheric and vacuum distillation unit,and determine input variables and target variables.The standard transformation of pressure pipeline data is carried out,then use the method of correlation analysis to simplify the data,and select the variables which have the greatest relationship with the failure possibility and the consequence of the pressure pipe.Using C5.0 algorithm and CHAID algorithm to construct the failure probability and the consequence model of pressure piping,and obtain the variables and the importance of variables in the model construction.By comparing the risk calculated by the big data method and the quantitative RBI method,the results show that the model has good accuracy and high reliability.