走滑断层作用下X80管道设计应变规律和神经网络预测

Design strain laws and neural network prediction of X80 pipeline under effect of strike-slip fault

走滑断层是埋地管道常见的地质灾害威胁,断层作用下管道会发生较大的拉压应变而失效。为得到X80管道的设计应变,基于有限元方法建立了走滑断层作用下管道的应变响应数值计算模型,模型使用壳单元模拟管道,非线性弹簧单元模拟土壤约束,采用西二线实际工程的管道应变影响参数范围,计算了管道的设计应变;为预测管道的设计应变值,基于以上参数化分析得到的4 817组设计应变结果,采用人工神经网络建立了管道设计应变预测模型。结果表明:该神经网络模型预测结果的最大相对误差小于10%,预测准确性良好,且该方法具有较高的计算效率,可以为断层作用下埋地管道的应变设计与评估提供参考。

The strike-slip fault is a common geological disaster threat to the buried pipeline,and the pipeline can be destructed due to the large tensile and compressive strain caused by the effect of fault.Based on the finite element method,a numerical calculation model on the strain response of pipeline under the effect of strike-slip fault was established to calculate the design strain of X80 pipeline.In the model,the pipeline was simulated by the shell elements,and the soil constraint was simulated by the nonlinear spring elements.The design strain of pipeline was calculated by using the range of influence factors for the pipeline strain in the practical engineering of the Second West to East Gas Pipeline Project.To predict the design strain of pipeline,a prediction model of design strain for pipeline was established by using the artificial neural network based on 4 817 groups of design strain results obtained by the parametric analysis.The results showed that the maximum relative error of the prediction results by the neural network model were less than 10%,which means a good prediction accuracy.Otherwise,this method has a high calculation efficiency,and it can provide the reference for the strain design and evaluation of the buried pipeline under the effect of fault.