基于不同模型的煤气管道剩余强度评价方法对比

Comparison of residual strength evaluation methods of gas pipeline based on different models

以不同钢级煤气管道为研究对象,借助Matlab软件,通过建立BP、GA-BP和带动量项BP人工神经网络数学模型,对其进行剩余强度预测,并与真实剩余强度进行对比分析;通过引入相对误差和压力比等参数,考察其准确性和适用性。结果表明,从运行时间和最大均方误差角度出发,GA-BP模型是最优的方法;从相对误差和压力比出发,GA-BP模型准确度最高,适用性最好。所得结论对煤气管道的正常运行有一定的指导意义。

With the help of Matlab software, the residual strength of gas pipeline with different steel grades is predicted by establishing BP,GA-BP and driving quantity TERM-BP artificial neural network mathematical model, and compared with the real residual strength.The relative error and pressure ratio are introduced to test the accuracy and applicability.The results show that GA-BP model is the best method in terms of running time and maximum mean square error.From the relative error and pressure ratio, GA-BP model has the highest accuracy and the best applicability.The conclusion has a certain guiding significance for the normal operation of gas pipeline.