Study on the Extent of the Impact of Data Set Type on the Performance of ANFIS for Controlling the Speed of DC Motor

This paper introduces an adaptive neuro-fuzzy inference system (ANFIS) for tracking SEDC motor speed in order to optimize the parameters of the transient speed response by finding out the perfect training data provider for the ANFIS. The controller was adjusted using PI, PD and PIPD to generate data sets to configure the ANFIS rules. The performance of the ANFIS controllers using these the different data sets was investigated. The efficiencies of the three controllers were compared to each other, where the PI, PD, and PIPD configurations were replaced by ANFIS to enhance the dynamic action of the controller. The performance of the proposed configurations was tested under different operating situations. Matlab’s Simulink toolbox was used to implement the designed controllers. The resultant responses proved that the ANFIS based on the PIPD dataset performed better than the ANFIS based on the PI and PD data sets. Moreover, the suggested controller showed a rapid dynamic response and delivered better performance under various operating conditions.