基于神经网络的故障模糊诊断研究

Research on fault fuzzy diagnosis based on neural network

文章用模糊算子代替标准的传递函数,利用征兆、规则、故障间的逻辑关系进行故障诊断,构造了四层模糊结构,给出了诊断的知识流程、机理分析和诊断步骤,并推导出一种能加快网络收敛速度的故障诊断及改进算法.在对某J8-wp13发动机滑油典型故障样本和非故障样本的仿真中,诊断结果良好.

<Abstrcat> The paper uses the fuzzy operators to replace the standard transmission function for fault diagnosis according to the logic relations among the signs, the rules and the faults, constitutes a four-layered fuzzy neural network, gives the knowledge flow, the principe analysis and the process for fault diagnosis, and deduces a fuzzy diagnosis algorithm and a modified algorithm which quicken the convergence speed of the network. The J8-wp13 Engine Lubricating Oil fault diagnosis of the typical and nontypical samples are simulated, the diagnosis results are good.