分离机故障诊断系统

Fault diagnosis system of separators

针对广州发电厂重油分离机的控制系统简单、故障判断困难的问题，分析了分离机振动特点，设计了一套基于ＬＶＱ神经网络技术的故障诊断系统。为了提高软件分析的可靠性，选择Ｃ．Ｊａｃｋｓｏｎ和Ｊ．Ｓ．Ｓｏｈｒｅ的研究结果作为分析知识库，并对振动状态进行了分类。在此基础上，论述了诊断系统软件结构，介绍了利用数值分析软件ＭＡＴＬＡＢ实现ＬＶＱ技术的编程方法。

The simplicity of the control system of the heavy oil separator in Guangzhou Power Plant makes it difficult to judge separator's failures, in view of which the vibration characteristics of the separator are analyzed and a fault diagnosis system based on LVQ neural network technique is designed. In order to improve the analytical reliability of the software, the research achievements of C. Jackson and J. S. Sohre are chosen as the analytical knowledge base and vibration statuses are classified. The software configuration of the diagnosis system is then illustrated and the programming method of using MATLAB numerical analysis software to realize LVQ technique is also included.