Entity

Time filter

Source Type


Chen S.-Y.,National Taiwan Normal University | Hung Y.-C.,TECO Electrical and Machinery Co. | Hung Y.-H.,National Taiwan Normal University | Wu C.-H.,National formosa University
Computers and Electrical Engineering | Year: 2015

A new recurrent wavelet fuzzy neural network (RWFNN) with adaptive learning rates is proposed to control the rotor position on the axial direction of a thrust magnetic bearing (TMB) mechanism in this study. First, the dynamic analysis of the TMB with differential driving mode (DDM) is derived. Because the dynamic characteristics and system parameters of the TMB mechanism are high nonlinear and time-varying, the RWFNN, which integrates wavelet transforms with fuzzy rules, is proposed to achieve precise positioning control of the TMB. For the designed RWFNN controller, the online learning algorithm is derived using back-propagation method. Moreover, since the improper selection of learning rates for the RWFNN will deteriorate the control performance, an improved particle swarm optimization (IPSO) is adopted to adapt the learning rates of the RWFNN on-line. Numerical simulations show the validity of TMB system using the proposed RWFNN controller with IPSO under the occurrence of uncertainties. © 2015 Elsevier Ltd. Source


Hung Y.-C.,TECO Electrical and Machinery Co. | Lin F.-J.,National Central University
1st International Future Energy Electronics Conference, IFEEC 2013 | Year: 2013

A Takagi-Sugeno-Kang type fuzzy neural network with asymmetric membership function (TSKFNN-AMF) is proposed in this study for the fault tolerant control of six-phase permanent magnet synchronous motor (PMSM) drive system. First, the dynamics of six-phase PMSM drive system, the fault detection and operating decision method are briefly introduced. Then, to achieve the required control performance and to maintain the stability of six-phase PMSM drive system under faulty condition, the TSKFNN-AMF control, which combines the advantages of TSK type fuzzy logic system (FLS) and AMF, is developed. The network structure and online learning algorithm of the TSKFNN-AMF are described in detail. Moreover, to enhance the control performance of the proposed intelligent fault tolerant control, a 32-bit floating-point digital signal processor (DSP) TMS320F28335, is adopted for the implementation. Finally, some experimental results are illustrated to show the validity of the proposed intelligent fault tolerant control for the six-phase PMSM drive system. © 2013 IEEE. Source


Xu S.S.-D.,National Taiwan University of Science and Technology | Chen C.-C.,National Chiao Tung University | Wu Z.-L.,TECO Electrical and Machinery Co.
IEEE Transactions on Industrial Electronics | Year: 2015

This paper studies fault-tolerant control (FTC) designs based on nonsingular terminal sliding-mode control and nonsingular fast terminal sliding-mode control (NFTSMC). The proposed active FTC laws are shown to be able to achieve fault-tolerant objectives and maintain stabilization performance even when some of the actuators fail to operate. In comparison to existing slidingmode control (SMC) fault-tolerant designs, the proposed schemes not only can retain the advantages of traditional SMC, including fast response, easy implementation, and robustness to disturbances/uncertainties, but also make the system states reach the control objective point in a finite amount of time. Moreover, they also resolve the potential singularity phenomena in traditional terminal and faster terminal SMC designs; meanwhile, the proposed NFTSMC fault-tolerant scheme also possesses the benefit of faster state convergence speed of NFTSMC. Finally, the proposed analytical results are also applied to the attitude control of a spacecraft. Simulation results demonstrate the benefits of the proposed schemes. © 2015 IEEE. Source


Chen W.-H.,National Taipei University of Technology | Chang S.-K.,University of Pittsburgh | Hung W.-P.,TECO Electrical and Machinery Co.
Proceedings of the International Conference on Software Engineering and Knowledge Engineering, SEKE | Year: 2014

The largest energy consumption, particularly electricity use, for residential users and service industry is attributed to air-conditioning systems. Although novel design can yield energy-efficient air-conditioning systems, there is still room for reducing energy consumption by the control of their operation, especially for those already installed. This study presents a slow intelligence framework for energy-saving control of air-conditioners without affecting users' thermal comfort. The implementation of the proposed approach to a smart space, and the architecture used to integrate the SIS server into the gateway as an embedded system, are also described. Copyright © 2014 by Knowledge Systems Institute Graduate School. Source


Hung Y.-C.,TECO Electrical and Machinery Co. | Lin F.-J.,National Central University | Hwang J.-C.,National Taiwan University of Science and Technology | Chang J.-K.,National Central University | Ruan K.-C.,National Central University
IEEE Transactions on Power Electronics | Year: 2015

A wavelet fuzzy neural network using asymmetric membership function (WFNN-AMF) with improved differential evolution (IDE) algorithm is proposed in this study to control a six-phase permanent magnet synchronous motor (PMSM) for an electric power steering (EPS) system. First, the dynamics of a steer-by-wire EPS system and a six-phase PMSM drive system are described in detail. Moreover, the WFNN-AMF controller, which combines the advantages of wavelet decomposition, fuzzy logic system, and asymmetric membership function (AMF), is developed to achieve the required control performance of the EPS system for the improvement of stability of the vehicle and the comfort of the driver. Furthermore, the online learning algorithm of WFNN-AMF is derived using back-propagation method. However, degenerated or diverged responses will be resulted due to the inappropriate selection of small or large learning rates of the WFNN-AMF. Therefore, an IDE algorithm is proposed to online adapt the learning rates of WFNN-AMF. In addition, a 32-bit floating-point digital signal processor, TMS320F28335, is adopted for the implementation of the proposed intelligent controlled EPS system. Finally, the feasibility of the proposed WFNN-AMF controller with IDE for the EPS system is verified through experimental results. © 2014 IEEE. Source

Discover hidden collaborations