论文部分内容阅读
对于复杂故障,溯因推理网络(abductive reasoning network,ARN)与相同结构的神经网络相比,具有优越的诊断准确性和更简单的结构。对于非线性多故障的化工系统故障,采用溯因推理网络对故障进行分类,能有效实现实时故障诊断,提高诊断精度。通过T-E过程(Tennessee-Eastman)仿真,结果表明所提出的方法优于小波神经网络方法。
For complex faults, abductive reasoning network (ARN) has superior diagnostic accuracy and simpler structure than neural networks of the same structure. For non-linear and multi-fault chemical system faults, the abnormity reasoning network is used to classify the faults, which can effectively realize the real-time fault diagnosis and improve the diagnostic accuracy. Through the T-E process (Tennessee-Eastman) simulation, the results show that the proposed method is superior to the wavelet neural network method.