Admixed recurrent Gegenbauer polynomials neural network with mended particle swarm optimization control system for synchronous reluctance motor driving continuously variable transmission system. (February 2020)

• Record Type: Journal Article
• Authors: Lin, Chih-Hong; Chang, Kuo-Tsai
• Is Part Of: Proceedings of the Institution of Mechanical Engineers. Volume 234:Number 2(2020)
• Page Start: 183
• Page End: 198
• Publication Date: 2020-02
• Subjects: Continuously variable transmission -- Gegenbauer polynomials neural network -- particle swarm optimization -- synchronous reluctance motor -- Lyapunov stability; Mechanical engineering -- Periodicals; Automatic control -- Periodicals; Systems engineering -- Periodicals; 621.3
• DOI: 10.1177/0959651819850654
• Languages: English
• ISSNs: 0959-6518

Abstract

To cut down influence of nonlinear time-varying uncertainty action in a synchronous reluctance motor driving continuously variable transmission system, an admixed recurrent Gegenbauer polynomials neural network with mended particle swarm optimization control system is posed for improving control performance. The admixed recurrent Gegenbauer polynomials neural network with mended particle swarm optimization control system involves an observer control, a recurrent Gegenbauer polynomial neural network control and a remunerated control. The weights of recurrent Gegenbauer polynomials neural network controller are regulated by using the adaptive law and the gradient descent technology. The remunerated control with a reckoned law is derived and computed by means of the Lyapunov stability theorem so as to pledge stability of the control system. Likewise, to speedup convergence of weights in the recurrent Gegenbauer polynomial neural network, the mended particle swarm optimization algorithm is used for regulating two kinds of learning rates. At last, three kinds of experimental results are demonstrated to confirm the usefulness of the put forward control system with comparative studies.