| In the process of natural gas pipeline operation,residual welding slag impurities during pipeline construction and solid particles such as sand and abrasive particles produced during pipeline transportation will cause erosion wear to key components such as valves and elbows.At present,the research on erosion and wear of pipeline system at home and abroad is often limited to simple components such as straight elbow,tee,etc.,and mainly focuses on liquid-solid erosion,The causes of damage,failure model analysis,quantitative analysis of influence factors and service life after damage are not clear.To this end,this thesis carried out the following research.(1)The causes of erosion wear of ball valve in natural gas pipeline are studied.The important characteristics of erosion wear,the theory and model of solid particle erosion wear,and the experimental research methods of erosion wear are summarized.The mathematical model and calculation method of erosion wear influenced by multi factors and multi parameters are put forward,and the problems of erosion wear caused by small particle erosion on curved surface of materials are also discussed,also put forward the research ideas and technical routes of this article.(2)A gas-solid two-phase flow erosion wear experimental device was designed and developed.Multi factor coupling gas-solid two-phase flow erosion wear experiments were carried out with plane and spherical surfaces as geometric characteristics.The deformation and cutting mechanism of plane and spherical erosion wear were analyzed and obtained.It was revealed that no micro ploughing feature was found on spherical surface erosion wear.In the spherical erosion experiment,it is found that the erosion rate is the highest when the particle size reaches 400μm,the erosion rate is the highest when the erosion angle is 30 ° and the erosion resistance of high strength steel is the highest.The influence weights of six kinds of characteristic quantities are proposed.(3)Three kinds of CFD refined numerical simulation of gas-solid erosion were carried out,and the erosion of the complete ball valve core in the pipe body,the jet erosion of the plane specimen and the jet erosion of the ball specimen were simulated respectively.The principle and process of gas-solid erosion were explored,and the velocity cloud,pressure cloud and particle trajectory of different characteristic parameters in the process of erosion were obtained.Referring to Oka erosion prediction model,the erosion rate and surface erosion morphology of the model under different working conditions were analyzed,and the numerical simulation was compared with the theoretical model,the correctness of the simulation boundary conditions is verified.(4)Multiple linear regression and machine learning were used to predict the erosion wear of ball valve.According to the experimental and simulation results,the erosion rate and core wall thickness prediction models were established.In multiple linear regression,analysis of variance and F-test were used to determine the significance of influencing variables;In the machine learning method,the SVM algorithm is adopted,and the particle swarm optimization(PSO),genetic algorithm(GA)and grid search algorithm(GS)are selected to optimize the SVM model respectively.The prediction model of erosion rate and valve core sealing surface thinning based on the improved SVM is optimized,which has better regression fitting effect.At the same time,it is applied to the actual case analysis and research to achieve a more accurate comprehensive prediction and evaluation of the residual life of the ball valve. |
PDF Full Text Request