African Buffalo Optimized Generative Mamdani Fuzzy Controller based Deep
Belief Network for Efficient Speed Control in PMSM
Abstract
Permanent Magnet Synchronous Motors (PMSM) are employed for
highly efficient motor drive. PMSM are efficient, brushless, fast, safe,
and have high dynamic performance. Many researchers pursued their areas
of interest in PMSM in order to improve their performance through speed
control. However, the PMSM’s efficiency was not reduced, and speed
control was not carried out in an efficient manner. This problem is
addressed by the African Buffalo Optimized Generative Mamdani Fuzzy
Controller-based Deep Belief Network (ABOGMFC-DBN) model. Specifically,
the ABOGMFC-DBN mode l is to handle the PMSM speed in order to attain a
higher current value. The ABOGMFC-DBN model performs two processes: the
multivariate African Buffalo hidden neuron and its weight optimization
process, and the generative Mamdani fuzzy controller-based deep belief
network process. The first procedure optimises the amount of hidden
neurons in the deep belief network and its weight parameters. The PMSM
speed is handled by the Mamdani fuzzy controller in the second step,
which uses four layers. The mean square error (MSE) is then calculated
in order to get the minimal rated current value using a Gaussian
activation function. Finally, the PMSM’s performance improves. Using the
PMSM parameter, the performance of the ABOGMFC-DBN model is evaluated on
the basis of rising time, settling time, peak value, peak time, and peak
overshoot. With a higher output current value compared to traditional
techniques, the simulation findings of the ABOGMFC-DBN model enhance the
PMSM’s performance.
