论文部分内容阅读
针对采煤机故障征兆和故障的非线性对应关系,采用广义回归神经网络作为故障诊断的智能分类器。输入层为采煤机的故障特征参数,中间层为径向基神经元,感知待诊断故障向量与训练样本的相似度,输出层为故障模式分类。分析了广义回归神经网络的优越性和结构特征,建立了不同光滑因子和训练样本数目的采煤机故障诊断模型,并在MATLAB进行了仿真。
Aiming at the nonlinear relationship between fault signs and faults of shearer, a generalized regression neural network is adopted as an intelligent classifier for fault diagnosis. The input layer is the fault characteristic parameter of shearer, the middle layer is the radial basis neuron, the similarity between the fault vector to be diagnosed and the training samples is sensed, and the output layer is the fault pattern classification. The advantages and structural characteristics of generalized regression neural network are analyzed. The fault diagnosis models of shearer with different smoothing factors and training samples are established and simulated in MATLAB.