Research On Residual Life Prediction Method Of Corrosion Defects Pipeline

Pipeline transportation plays an increasingly important role in nowadays'production. However, with the continuous use of the pipeline, there will be inevitable corrosion caused by pipeline leakage, which eventually becomes an important safety hazard of pipeline transportation, may cause significant economic losses. To prevent the occurrence of pipeline leakage, it is necessary to use device to detect pipeline in order to know the status of pipeline's corrosion and defects, and thus to predict the remaining life and health status of the pipeline. Predict the remaining life of the pipeline is an issue to understand the pipeline's entire life circle and to assist the development of the pipeline operation's long-term planning. According to the predicted results, it is able to find out the balance between safety and effectiveness of the pipeline, which can make the maintenance and repair of the pipeline planned and targeted.In this thesis, it proposed a new method to predict the pipeline life based on multi-source data fusion. This thesis mainly reflects the following aspects:(1) Studied the basic theory of gray theory method and neural network method, and applied improved prediction of corrosion defects to these methods, proposed an improved GM(1,1,?.) model and a BP neural network based on genetic algorithm. Meanwhile, simulated the injection pipeline test data got from Shengli Oilfield test area, compared the basic method with the improved method, and analyzed the applicability and accuracy of prediction results by using gray network prediction method and neural network prediction method. Finally, by combining the two methods, proposed a defect prediction method based on multi-source data fusion, which will make up the deficiencies of a single model, play the advantages of the combination model, and improve the accuracy of prediction of corrosion defects.(2) Combined the pipeline residual strength evaluation method from home and abroad, after obtained the change of defect size by using prediction method of developmental corrosion and defects, got pressure curve of defect pipeline, and thus to predict the life of the pipeline. And used the results of life prediction to analyze pipeline safety, cycle and feasibility of decrease pressure. Finally, proved the validity of the prediction method by using simulation experiments.(3) Designed the hardware and software platform of the residual life prediction system, including data acquisition program design, software processes and functional data, data files and interface design, and make it in the LabVIEW.In this thesis, by using theory research and simulation test, the method of predicting the developmental of pipeline corrosion based on multi-source data fusion can be proved validly to predict the development of defects. After introduced residual strength evaluation method and combined that with results from the method of developmental corrosion prediction, it can be obtained the pressure change of pipeline, the life of pipeline and other reference information that can be used to assist pipeline's run and maintenance, which has some practical significance. With further enhancing the ability of theory and practice, this method will play a greater role in the safe operation and maintenance of the pipeline and the prediction of the pipeline life.