基于灰色支持向量回归的起伏管线临界携液流速预测模型OA北大核心CSTPCD

The undulating geographical morphology of Yan'an gas field makes it easy for gas production pipelines to accumulate liquid at low points and uphill sections,which affects the normal production of the gas field.Based on the extended dual fluid phase separation model and PR gas property equation,a sample dataset of critical liquid-carrying velocity was established using the minimum pressure gradient method combined with uniform design.The weights of sample factors were calculated using grey correlation,and a regression prediction model for critical liquid-carrying velocity was established by introducing them into the support vector regression algorithm.The hyperparameters in the model were optimized using grid optimization algorithm and K-CV(cross validation).The accuracy of the model was verified through on-site operational data from Yan'an gas field.The results indicate that the diameter size of gas pipelines has the greatest impact on the critical liquid-carrying velocity,followed by the uphill inclination angle,water content,and operating pres-sure,while other factors have a relatively small impact.Compared with genetic algorithm and particle swarm optimization algorithm,seg-mented grid optimization algorithm has stronger stability in optimizing the hyperparameters of the model.This prediction model has high accuracy in predicting the critical liquid-carrying velocity of the Yan'an gas field production pipelines,and can provide theoretical sup-port for preventing liquid accumulation in gas transmission pipelines.