TY - JOUR
T1 - Energy-Saving Dynamic Bias Current Control of Active Magnetic Bearing Positioning System Using Adaptive Differential Evolution
AU - Chen, Syuan Yi
AU - Song, Min Han
N1 - Funding Information:
This work was supported by the Ministry of Science and Technology in Taiwan under Grant MOST 103-2218-E-003-001.
Funding Information:
Manuscript received May 12, 2016; revised January 15, 2017; accepted March 23, 2017. Date of publication April 25, 2017; date of current version April 15, 2019. This work was supported by the Ministry of Science and Technology in Taiwan under Grant MOST 103-2218-E-003-001. This paper was recommended by Associate Editor Y.-J. Liu. (Corresponding author: Syuan-Yi Chen.) The authors are with the Department of Electrical Engineering, National Taiwan Normal University, Taipei 106, Taiwan (e-mail: chensy@ntnu.edu.tw).
Publisher Copyright:
© 2013 IEEE.
PY - 2019/5
Y1 - 2019/5
N2 - This paper proposes an evolutionary algorithm-based energy-saving control strategy to control a highly nonlinear and time-varying active magnetic bearing (AMB) positioning system that considers energy efficiency and control performance simultaneously. Most AMB control schemes use a bias current with a superimposed control current to improve the linearity and dynamic performance of the system. However, the bias current causes power losses even if no electromagnetic force is required. As such, a recurrent wavelet fuzzy neural network with adaptive differential evolution (RWFNN-ADE)-based dynamic bias current control strategy is proposed in this paper so as to minimize the energy consumed by an AMB without altering its positioning performance and robustness. To begin with, this paper analyzes the operation principle of the AMB positioning system with a differential driving mode. Subsequently, the proposed RWFNN-ADE control scheme, in which the control current and bias current are controlled by the RWFNN and ADE, respectively, is introduced in detail. Finally, the experimental results demonstrate the high-accuracy control and significant energy-saving performances of the proposed RWFNN-ADE-controlled AMB positioning system. In the tests corresponding to operation periods of 10 and 50 s, the energy improvements compared to the baseline values were 20.24% and 17.65%, respectively, in nominal cases, and 18.89% and 18.68%, respectively, in parameter variation cases, for the proposed control strategy.
AB - This paper proposes an evolutionary algorithm-based energy-saving control strategy to control a highly nonlinear and time-varying active magnetic bearing (AMB) positioning system that considers energy efficiency and control performance simultaneously. Most AMB control schemes use a bias current with a superimposed control current to improve the linearity and dynamic performance of the system. However, the bias current causes power losses even if no electromagnetic force is required. As such, a recurrent wavelet fuzzy neural network with adaptive differential evolution (RWFNN-ADE)-based dynamic bias current control strategy is proposed in this paper so as to minimize the energy consumed by an AMB without altering its positioning performance and robustness. To begin with, this paper analyzes the operation principle of the AMB positioning system with a differential driving mode. Subsequently, the proposed RWFNN-ADE control scheme, in which the control current and bias current are controlled by the RWFNN and ADE, respectively, is introduced in detail. Finally, the experimental results demonstrate the high-accuracy control and significant energy-saving performances of the proposed RWFNN-ADE-controlled AMB positioning system. In the tests corresponding to operation periods of 10 and 50 s, the energy improvements compared to the baseline values were 20.24% and 17.65%, respectively, in nominal cases, and 18.89% and 18.68%, respectively, in parameter variation cases, for the proposed control strategy.
KW - Active magnetic bearing (AMB)
KW - adaptive differential evolution (ADE)
KW - differential driving mode (DDM)
KW - energy saving
KW - recurrent wavelet fuzzy neural network (RWFNN)
UR - http://www.scopus.com/inward/record.url?scp=85018943218&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85018943218&partnerID=8YFLogxK
U2 - 10.1109/TSMC.2017.2691304
DO - 10.1109/TSMC.2017.2691304
M3 - Article
AN - SCOPUS:85018943218
SN - 2168-2216
VL - 49
SP - 942
EP - 953
JO - IEEE Transactions on Systems, Man, and Cybernetics: Systems
JF - IEEE Transactions on Systems, Man, and Cybernetics: Systems
IS - 5
M1 - 7911357
ER -