摘要
为提高长输管道泄漏检测的准确率,提出基于GA_Elman神经网络的管道泄漏检测方法,该方法通过使用遗传算法(GA)对Elman神经网络的权值和阈值进行优化,不但克服了Elman神经网络易陷入局部极值的缺陷,而且提高了Elman神经网络的预测精度。实验证明该方法可用于管道泄漏检测,其效果优于BP神经网络与Elman神经网络检测模型,预测精度96.9%。
For purpose of improving leakage detection accuracy of long-distance pipelines,the GA_Elman neural network-based leakage detection method for pipelines was proposed,which employs genetic algorithm(GA)to optimize Elman neural network’s weight number and threshold value;and it overcomes the Elman neural network’s defect of being susceptible to falling into local extremum,improves the forecast accuracy of the Elman neural network.The experimental results show that,this method can be used to predict leakage detection of pipelines and its prediction effect outperforms the detection model of both BP neural network and Elman neural network owing to a prediction accuracy of 96.9%.
作者
张勇
韦焱文
王明吉
周兴达
刘洁
杨文武
WEI Yan-wen;WANG Ming-ji;ZHOU Xing-da;LIU Jie;YANG Wen-wu(School of Physics and Electronic Engineering,Northeast Petroleum University)
出处
《化工自动化及仪表》
CAS
2022年第2期182-186,245,共6页
Control and Instruments in Chemical Industry
基金
教育部重点实验室开放基金项目(MECOF2019B02)。
