Neural-Fuzzy Speed Controller Design for Induction Motors

• Main Authors: Jiong-Neng Liao, 廖炯能
• Other Authors: Chang Yeong-Hwa
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/53150858419783304925

碩士 === 長庚大學 === 電機工程研究所 === 95 === This thesis, based on the advantage of fuzzy inference and learning ability of neural network, mainly investigates the design of neural-fuzzy speed controller for induction motors, where the subjects of parameter variations, load disturbance, and mode uncertainties can be conquered with. The back-propagation algorithm is used for the parameter learning in which the result of convergence can be improved with variable learning rate. To verify the proposed result, Matlab is used for the simulation evaluation. In the stage of experiment implementation, a DSP/FPGA co-design is adopted. From the simulation and experimental results, the proposed neural-fuzzy controller can provide better training performance subject to possible parameter variations and load disturbance.