一种改进的SOM神经网络在污水处理故障诊断中的应用  

An Application of Improved SOM Neural Network in Fault Diagnosis of Wastewater Treatment

作　　者：岳宇飞 罗健旭[1] 

机构地区：[1]华东理工大学自动化研究所,上海200237

出　　处：《华东理工大学学报（自然科学版）》2017年第3期389-396,共8页Journal of East China University of Science and Technology

基　　金：国家自然科学基金(61304071)

摘　　要：自组织映射(SOM)神经网络初始权值的选取对神经网络的性能有重要的影响。采用改进的帝国竞争算法(IICA)优化局部权重失真指数(LWDI)寻优SOM神经网络的初始权值;利用改进后的SOM神经网络(IICA-SOM)对污水处理过程数据进行聚类和故障诊断。实验结果表明,与传统的SOM算法相比,IICA-SOM算法取得了更好的聚类效果,且故障诊断的误诊率更低。The selecting of the initial weights of the self-organizing map （SOM） neural network has an important impact on the performance of the network. In this paper, by means of an improved imperialists competitive algorithm （IICA） to minimize the locally weight distortion index, the initial weights of the SOM neural network can be obtained optimally. Besides, the improved SOM neural network （IICA-SOM） is applied in the clustering and fault diagnosis of the wastewater treatment process data. The experiment result demonstrates th a t, compared with the traditional SOM neural network, IICA-SOM can attain better performance in clustering and lower misdiagnosis rate in the fault diagnosis.

关 键 词：故障诊断 聚类 自组织映射 帝国竞争算法 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]