Parameter Estimation in High Performance Sensor Less Vector Controlled Drives
M. Sasikumar, S. Chenthur Pandian · 2007
An Artificial Intelligence (AI) based Estimator is robust to parameter variations and noise and it avoids the use of mathematical models. Such a system is not restricted by the many assumptions used in the conventional methods and is capable of mapping any degree of non linearity. It can also yield the results more quickly. By the application of minimum configuration it is possible to obtain cost effective simple solutions using FPGA. The conventional methods are direct synthesis from state equations, Model Reference Adaptive System (MRAS) and Flux Observers. All these techniques use complex mathematical model of the motor which includes many assumptions. The estimation is not robust to parameter variations. The time taken for computation is also long. In this paper to develop AI based Estimators to Estimate the Speed, Torque and Flux of an Induction Motor for DTC drives.