基于改进角分类神经网络的冷水机组故障诊断  

Fault Diagnosis of Water Chiller Based on Improved FDCC Neural Network

作　　者：罗方芳[1] 陶求华[2] LUO Fang-fang, TAO Qiu-hua（1.Computer Engineering College, Jimei University, Xiamen 361021,China;2.Mechanical Engineering College, Jimei University, Xiamen 361021 ,China）

机构地区：[1]集美大学计算机工程学院,福建厦门361021 [2]集美大学机械工程学院,福建厦门361021

出　　处：《电脑知识与技术》2008年第5期717-719,共3页Computer Knowledge and Technology

摘　　要：针对冷水机组的故障诊断问题及其特点，提出了一种基于改进角分类神经网络故障诊断模型FDCC（Fault Diagnosis Comer Classification）。该模型克服了角分类神经网络（CC4）输出结果为二进制的局限，根据故障模式所落入的k最近邻的样本泛化空间来进行故障诊断并输出结果向量，其各分量为各故障原因可能出现的概率。According to the fault diagnosis of water chiller and its features, a new kind of FDCC neural network model is proposed to detect and diagnose common water chiller faults. It can over the binary system output disadvantage of CC4 by use of this model. According to the k-nearest neighbor samples＇generalization space, FDCC neural network model diagnosis fault and give output Vector, whose components are the fault Probabilities.

关 键 词：角分类 神经网络 冷水机组 故障诊断 

分 类 号：TP391[自动化与计算机技术—计算机应用技术;自动化与计算机技术—计算机科学与技术]