论文部分内容阅读
针对神经网络智能诊断与专家系统中知识难于理解和诊断解释能力差等问题,研究了一种新的基于功能性观点的神经网络规则提取方法,介绍了方法流程及关键算法.并用UCI(加利福尼亚大学埃尔文分校)机器学习数据对方法进行了分析和验证.最后,将方法应用于实际航空发动机磨损故障诊断中,采集了某型航空发动机实测油样光谱数据237个样本,利用神经网络规则提取方法提取了发动机磨损故障诊断知识规则,并对其进行了解释,结果表明了方法的正确有效性.
Aiming at the problem of difficult to understand knowledge and poor ability to explain and interpret the knowledge in neural network intelligent diagnosis and expert system, this paper studies a new neural network rule extraction method based on functional viewpoint and introduces the method flow and the key algorithm.Using UCI (University of California, Elvin) machine learning data were analyzed and validated.Finally, the method was applied to the actual aeroengine wear fault diagnosis, and 237 samples of aero-engine measured oil spectrum data were collected and extracted by the rule of neural network The method extracts the knowledge rules of engine wear fault diagnosis and explains it, and the results show that the method is correct and effective.