基于ISOA-RBPNN的埋地管道剩余强度预测*

为提高腐蚀管道剩余强度的预测精度，提出引入弹性梯度下降法改进BP神经网络，并融合改进海鸥优化算法(ISOA)，构建腐蚀管道剩余强度预测模型。关于改进BP神经网络模型的参数寻优，首先采用Cat混沌映射初始化改进海鸥优化算法(SOA)初始种群的分布，提升寻优能力，优化SOA的搜索方向和攻击形式，增强其全局搜索能力并提高收敛速度，然后用ISOA对弹性BP神经网络（RBPNN）模型中的权值和阈值进行寻优，最后构建ISOA-RBPNN预测模型。以管道爆破数据为例，利用MATLAB进行仿真模拟，并与PSO-BPNN模型和IFA-BPNN模型预测结果进行对比分析。研究结果表明:ISOA-RBPNN模型的各项评价指标均优于其他2个模型，预测结果较实际值误差更小，在预测腐蚀管道剩余强度领域具有更好的性能，可为后续研究腐蚀管道剩余寿命和制定维修策略提供参考依据。

Residual strength prediction of buried pipeline based on ISOA-RBPNN

In order to improve the prediction accuracy of the residual strength of corroded pipelines,the elastic gradient descent method was introduced to improve the BP neural network,and the improved seagull optimization algorithm (ISOA) was integrated to construct the residual strength prediction model of corroded pipelines.Regarding the parameter optimization of the improved BP neural network model,the Cat chaotic map was used to initialize and optimize the distribution of the initial population of the seagull optimization algorithm (SOA) for improving the optimization ability,and the search direction and attack form of SOA were optimized to enhance its global search ability and improve the convergence speed.Then ISOA was used to optimize the weights and thresholds in the resilient BP neural network (RBPNN) model,and finally the ISOA-RBPNN prediction model was established.Taking the pipeline blasting data as an example,MATLAB was used for simulation,and the prediction results of the model were compared and analyzed with those of PSO-BPNN model and IFA-BPNN model.The results showed that the evaluation indicators of the ISOA-RBPNN model were all better than the other two models,the error between the prediction results the actual value was smaller,and it had better performance in predicting the residual strength of corroded pipeline.It can provide reference for the follow-up study on the remaining life of corroded pipelines and the formulation of maintenance strategies.