针对抓斗纠偏系统复杂性、不确定性、模糊性的特点，提出基于故障树的模糊神经网络作为抓斗纠偏系统故障诊断的方法。该方法利用故障树知识提取抓斗纠偏系统故障诊断的输入变量和输出变量，引入模糊逻辑的概念，采用模糊隶属函数来描述故障的程度，利用Levenberg-Marquardt优化算法对神经网络进行训练，系统推理速度快、容错能力强，并通过实例分析验证了抓斗纠偏系统模糊神经网络故障诊断的有效性。%For the complexity,uncertainty,ambiguity grab correction system,this paper proposed fuzzy neural network based fault tree as a method for grab correction system fault diagnosis.The method extracted fault diagnosis input and output varia-bles of grab correction system by using fault tree knowledge.It introduced the concept of fuzzy logic,used fuzzy membership functions to describe the degree of failures.It used Levenberg-Marquardt algorithm to train the neural network system,got a better performance in inference speed and fault-tolerant.The experiments verify the effectiveness of fault diagnosis grab correc-tion system based on fuzzy neural network.
展开▼
