Title :
Parameter adaption for field-oriented AC-drives using neural network
Author :
Hofmann, Wilfried ; Liang, Qinglong
Author_Institution :
Dept. of Electr. Machines & Drives, Tech. Univ. Chemnitz, Germany
Abstract :
Modern high dynamic drive systems controlled by field-orientation require precise modeling of the induction motor. Because of the differences between the real motor and the model and the parameter changes during runtime, it is necessary to analyze the model error. Using a neural network allows the method of field-orientated control to be employed adaptively, taking into account parameter changes of the motor during runtime. The neural network is trained through classified patterns by error of outputs between induction motor and the dynamic computational model. The new configuration has been verified by simulation results
Keywords :
adaptive control; control system synthesis; error analysis; induction motor drives; neural nets; neurocontrollers; observers; pattern classification; AC-drives; dynamic computational model; dynamic neural network; error pattern classification; field-orientated control; flux observer; induction motor; model error; output errors; vector control; Backpropagation; Computational modeling; Computer networks; Electric variables control; Electronic mail; Induction motors; Neural networks; Rotors; Runtime; Steady-state;
Conference_Titel :
Neural Networks, 1995. Proceedings., IEEE International Conference on
Conference_Location :
Perth, WA
Print_ISBN :
0-7803-2768-3
DOI :
10.1109/ICNN.1995.487725
