African Buffalo Optimized Generative Mamdani Fuzzy Controller Based Deep Belief Network for Efficient Speed Control in Permanent Magnet Synchronous Motor

Permanent magnet synchronous motors (PMSMs) have gained attention owing to their advantages including high efficiency, brushless operation, rapid performance, safety, and superior dynamic capabilities. These characteristics make PMSMs ideal for applications that require precise speed control. However, PMSM performance and efficiency can be affected by external load disturbances and parameter deviations such as nonlinearity, time variance, and coupling. To address these challenges, researchers have developed the African buffalo-optimized generative Mamdani fuzzy controller-based deep belief network (ABOGMFC-DBN) model as a solution. The ABOGMFC-DBN model uses a two-step approach to optimize PMSM speed control and achieve higher current values. The first step involves developing a Deep Belief Network (DBN) model to determine the optimal number of hidden neurons in the hidden layer. This DBN model was fine-tuned using the African Buffalo Optimization (ABO) algorithm to optimize the weight parameters and refine the hidden neurons. A Mamdani fuzzy proportional-integral (PI) controller was implemented to eliminate the steady-state error. In the second step, the optimized DBN model adjusts the fuzzy controller input parameters and optimizes the rules and fuzzy membership functions. The ABOGMFC-DBN model was then applied to examine PMSM speed regulation. The performance of the model was evaluated using PMSM parameters, including rise time, settling time, peak value, peak time, and peak overshoot, and was compared with traditional and other heuristic controllers. This approach demonstrates the effectiveness of the ABOGMFC-DBN model in enhancing the PMSM performance and addressing external disturbances and parameter deviations.