Single Neuron ANN Based Current Controlled Permanent Magnet Brushless DC motor Drives

摘要

Permanent Magnet Brushless DC (PMBLDC) motor has a huge scope of industrial application in modern automated industries. In this paper, an artificial neural network (ANN) based PMBLDC motor drives are proposed and the computational complexities of the neural network is avoided by using single neuron based neural network. Two types of neural network with different activation functions are considered in this paper. One is recurrent neural network (RNN) and another is feed forward back propagation (FFBP) based neural network. The performance of these proposed ANN based drives with respect to conventional adaptive PI controller is compared for different dynamic conditions. Considering starting characteristics there is no performance deviation among the proposed drives with respect to conventional drive system. But in case of dynamic speed changing condition, the transient current response of PMBLDC motor drive for proposed single neuron ANN based controller is moderated as compared to conventional drive.