Blend recurrent Gegenbauer orthogonal polynomials neural network control of a SynRM servo-drive CVT system using amended artificial bee colony optimization

Author(s):  
Chih-Hong Lin ◽  
Hong-Yen Hsu
Author(s):  
Chih-Hong Lin

In comparison control performance with more complex and nonlinear control methods, the classical linear controller is poor because of the nonlinear uncertainty action that the continuously variable transmission (CVT) system is operated by the synchronous reluctance motor (SynRM). Owing to good learning skill online, a blend amended recurrent Gegenbauer-functional-expansions neural network (NN) control system was developed to return to the nonlinear uncertainties behavior. The blend amended recurrent Gegenbauer-functional-expansions NN control system can fulfill overseer control, amended recurrent Gegenbauer-functional-expansions NN control with an adaptive dharma, and recompensed control with a reckoned dharma. In addition, according to the Lyapunov stability theorem, the adaptive dharma in the amended recurrent Gegenbauer-functional-expansions NN and the reckoned dharma of the recompensed controller are established. Furthermore, an altered artificial bee colony optimization (ABCO) yields two varied learning rates for two parameters to find two optimal values, which helped improve convergence. Finally, the experimental results with various comparisons are demonstrated to confirm that the proposed control system can result in better control performance.


2018 ◽  
Vol 280 ◽  
pp. 32-45 ◽  
Author(s):  
George E. Tsekouras ◽  
Vasilis Trygonis ◽  
Andreas Maniatopoulos ◽  
Anastasios Rigos ◽  
Antonios Chatzipavlis ◽  
...  

Author(s):  
Vijo M. Joy ◽  
S. Krishnakumar

<p>The objective of this work is to solve the power scheduling problems for efficient energy management by assigning the optimal values. Artificial neural networks are used widely in the field of energy management and load scheduling. The backpropagation technique is used for the feed-forward neural network training and the levenberg-marquardt algorithm is used to minimize the errors. The slow speed of convergence and getting stuck in local minima are some negatives of backpropagation in complex computation. To overcome these drawbacks an innovative meta-heuristic search algorithm called modified artificial bee colony optimization algorithm is used. A hybrid neural network is introduced in this work. The simulation result shows that the efficiency of the system is improved when hybrid optimization is used. With this method, the system achieves an optimal accuracy of 99.23%.</p>


Sign in / Sign up

Export Citation Format

Share Document