摘要
为解决输油管道易腐蚀,且腐蚀程度难以测量的问题,提出使用改进的粒子群算法(PSO)优化误差反向传播神经网络(BPNN)对输油管道内腐蚀速率进行预测。改进的PSO算法提升了自身搜索到全局最优的能力,可为BPNN提供最优初始权值和阈值,从而有效避免BPNN易陷入局部最优的问题发生。以某条输油管线为例,分别运用标准的BPNN模型、PSO-BPNN以及改进的PSO-BPNN对该管线内腐蚀速率进行预测。结果表明:基于改进的PSO-BPNN的预测结果平均相对误差为5.57%,预测精度较BPNN和PSO-BPNN有明显提升。使用改进的PSO-BPNN预测输油管道的腐蚀速率可为管道的检测维修提供可靠的理论和技术支撑。
In order to solve the problems that the oil pipeline is easy to occur the corrosion and the corrosion degree is difficult to measure,it was proposed to predict the internal corrosion rate of oil pipeline by using the improved particle swarm optimization(PSO)to optimize the back propagation neural network(BPNN).The improved PSO algorithm promoted its ability to search for global optimum,which could provide the optimal initial weights and thresholds for BPNN,thus effectively avoid the problem that BPNN is prone to fall into local optimum.Taking a certain oil pipeline as an example,the standard BPNN model,PSO-BPNN and improved PSO-BPNN were used respectively to predict the internal corrosion rate of this pipeline.The results showed that the average relative error of prediction results based on the improved PSO-BPNN was 5.57%,and the prediction accuracy was significantly improved compared with those of BPNN and PSO-BPNN.So predicting the corrosion rate of oil pipelines by using the improved PSO-BPNN can provide the reliable theoretical and technical support for the inspection and maintenance of pipelines.
作者
凌晓
徐鲁帅
梁瑞
郭凯
崔本廷
岳守体
LING Xiao;XU Lushuai;LIANG Rui;GUO Kai;CUI Benting;YUE Shouti(College of Petroleum and Chemical Engineering,Lanzhou University of Technology,Lanzhou Gansu 730050,China;Taiyuan Satellite Launch Center,Taiyuan Shanxi 030027,China)
出处
《中国安全生产科学技术》
CAS
CSCD
北大核心
2019年第10期63-68,共6页
Journal of Safety Science and Technology
基金
甘肃省重点研发计划-工业类(1604GKCA022)
关键词
输油管道
粒子群算法
BP神经网络
腐蚀速率
oil pipeline
particle swarm optimization(PSO)
BP neural network
corrosion rate
