基于模糊神经网络的铝电解槽工作状态诊断

Diagnosis of Working Conditions of Aluminum Reduction Cells Based on Fuzzy Neural Network

针对铝电解槽电压波动信号的频谱特点和工作状态的层次性,将小波理论和模糊逻辑理论引入高阶BP神经网络中,提出了一种基于小波高阶模糊神经网络的铝电解槽工作状态诊断模型,此模型运用串联方式将小波包分析、模糊逻辑和神经网络融合在一起。给出了小波高阶模糊神经网络的结构,确定了各模糊子集上的隶属度函数。并在320KA预焙铝电解槽上进行实验和仿真,理论和实验证明了该诊断方法可行性。

The wavelet and fuzzy theory were introduced into high-order BP Neural Network according to the frequency spectrum characteristics of voltage vibration signal and arrangement of working conditions. Diagnosis model of working conditions of aluminum reduction cells were developed based on the wavelet high-order fuzzy neural network which integrated the three promising technologies： wavelet transform, artificial neural network s （ANNs） and fuzzy logic. The structure of FNN was given and the membership function was developed according to the actual situation. All simulated reasons of working Conditions were emulated on 320KA prebaked aluminum reduction cells. The feasibility of this novel method is proved by the simulation results.