Numerical Simulation Of Erosion Characteristics Of Oil And Gas Pipelines Containing Defects And Research On Residual Strength Prediction Technology

Pipeline transportation is the most widespread way of transporting oil and gas,any leakage will cause serious consequences.In the process of oil and gas transportation,the fluid passing through at high speed mixed with impurity particles causes erosion and wear on the inner wall of the pipelines,which is the main reason for the thinning of the pipelines wall and even the leakage of perforation.In this thesis,numerical methods are used to explore the erosion and wear laws of pipelines.The neural network model is used to accurately predict the erosion rate,residual life and residual strength of pipelines with defects.The main research contents are as follows:First of all,based on the erosion wear mechanism,the thesis conducts research on the erosion wear law of defective elbows,establishes a calculation model for the erosion wear of defective elbows,and obtains the impact of mass flow rate,inlet flow velocity,and particle size on the erosion of elbows with defects.The influence of erosion and wear.Secondly,in order to analyze the influencing factors of the residual strength of the defective pipelines,this thesis combined the failure judgment criteria to establish a calculation model for the residual strength of the defective elbow.What’s more calculated the open experimental data through the numerical simulation method to verify the correctness of the calculation model.Using this model,the effects of point-shaped defects and strip-shaped defects with different parameters on the remaining strength of the elbow were studied.It was found that the depth of the defect had the greatest influence on the residual strength of the pipelines.On the basis of the above research,this thesis predicts the erosion rate,residual life and residual strength of defective elbows.Extreme learning machine prediction model based on genetic algorithm optimization was established,and the validity of the prediction model was verified according to the public experimental data.The factors affecting the prediction accuracy of the model were analyzed.Using the numerical simulation data in this thesis to carry out prediction work,the results show that the improved model has a prediction accuracy of 93.679%,93.547%,and98.499% for the erosion rate,residual life,and residual strength of defective pipelines,which is better than traditional extreme learning machine prediction.model.