基于DCPSO的模糊神经网络的管道泄漏检测方法  

Study on pipeline leak detection methods of fuzzy neural network based on DCPSO optimization

作　　者：李炜[1] 张美玲[1] 李庆卿 

机构地区：[1]兰州理工大学电气工程与信息工程学院,兰州730050 [2]甘肃第一安装工程有限公司,兰州730060

出　　处：《工业仪表与自动化装置》2010年第6期3-7,共5页Industrial Instrumentation & Automation

基　　金：教育部"春晖计划"(Z2005-1-62001);兰州理工大学特色学术梯队基金项目(0950)

摘　　要：提出了一种基于发散-收敛PSO(DCPSO)优化模糊神经网络的管道泄漏检测与估计方法。该方法采用广义概率积、广义概率和模糊算子代替普通神经网络中的传递函数,并用DCPSO算法对该模糊神经网络的权值进行优化。通过实际管道泄漏数据对网络进行仿真研究,结果表明文中所述方法在管道泄漏的检测与估计中,不仅比BP算法具有更快的收敛速度,其结果也更优,进而也昭示出该方法在管道泄漏检测与估计中的可用性。This paper puts forward a kind of new method applied in pipeline leak detection and estimation.That is fuzzy neural network based on divergent-convergent PSO（DCPSO） optimization algorithm.This method used the generalized probability and the probability and generalized fuzzy operator to substitute for the transfer function of neural network,and used DCPSO optimization algorithm to optimize the fuzzy neural network weights.Through the actual pipeline leak of network data simulation,indicated that this method avoided the defects of slow in network training and easy in the local superior of the BP neural network in the pipeline leak detection and estimation,and verified the validity of this method.

关 键 词：管道 泄漏检测与估计 广义概率积 广义概率和 模糊神经网络 DCPSO优化算法 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]