基于改进SSA-BPNN的管道腐蚀剩余强度预测研究  

Prediction of Residual Strength of Pipeline Corrosion Based on Improved SSA-BPNN

作　　者：焦庆雨 吴紫翔 罗昊 JIAO Qing-yu;WU Zi-xiang;LUO Hao(PetroChina Daqing Petrochemical Company Motorized Equipment Division,Daqing 163000;College of Mechanical and Electrical Engineering,Beijing University of chemical Technology,Beijing 100029;School of Materials Science and Engineering,Beijing University of chemical Technology,Beijing 100029)

机构地区：[1]中国石油大庆石化公司机动设备处,黑龙江大庆163000 [2]北京化工大学机电工程学院,北京100029 [3]北京化工大学材料科学与工程学院,北京100029

出　　处：《制造业自动化》2025年第2期140-147,共8页Manufacturing Automation

摘　　要：针对原麻雀搜索算法(SSA)搜索精度不高,预测石油管道腐蚀剩余强度的SSA-BP模型精度较低的问题,提出了一种混合策略麻雀算法(BOSSA)。采用改进的Circle混沌映射初始化种群,提高种群初始化的多样性;结合鲸鱼算法的位置迭代公式替换发现者位置变化公式,提高了算法的全局搜索能力;将正余弦算法引入加入者公式,以提升算法的局部搜索能力;引入非线性递减策略控制种群中警戒者的数量,加快了种群的收敛速度;选取部分麻雀进行交叉变异,提升种群的多样性。利用四个测试函数进行测试,结果表明改进SSA算法(BOSSA)与其他算法相比有更好的寻优能力和迭代速度。将改进后的算法优化神经网络,以某石油管道为例对管道的剩余强度进行预测。结果表明:BOSSA-BPNN模型的预测结果的平均误差仅为2.21%,显著高于其他模型,可以为管道的检测和维修提供可靠的理论和技术支持。A hybrid strategy sparrow algorithm(BOSSA)is proposed to address the low search accuracy of the original sparrow search algorithm(SSA)and the low accuracy of the SSA-BP model for predicting multi factor corrosion of oil pipelines.An improved Circle chaotic map is used to initialize the population and increase the diversity of population initialization;The discoverer's position change formula is replaced by the position iteration formula that is combined with the whale algorithm to improve the overall search ability of the algorithm;The sine cosine algorithm is brought into the joiners’formula to improve the algorithm's local search ability;A nonlinear decreasing strategy is introduced to control the number of vigilantes in the population to accelerate the convergence speed of the population,while some sparrows for cross mutation are selected to enhance population diversity.With four test functions being used for testing,the results show that the improved SSA algorithm(BOSSA)has better optimization ability and iteration speed compared to other algorithms.Taken a certain oil pipeline as an example,the neural network is optimized with the improved algorithm to predict the remaining strength of pipelines.The results show that the average error of the prediction results of the BOSSA-BPNN model is merely 2.21%,which is significantly higher than that of other models.It can provide reliable theoretical and technical support for the pipeline inspection and maintenance.

关 键 词：麻雀搜索算法 Circle混沌映射 正余弦算法 交叉变异 管道剩余预测强度 

分 类 号：TG174[金属学及工艺—金属表面处理]