Indirect field oriented control and direct torque control comparison with/without artificial neural networks on asynchronous motors.

The flux, speed, and torque control performance of asynchronous motors are affected by parameter deviations and nonlinear variations of the asynchronous motor. In this study, Direct Torque Control (DTC) and Indirect Field Oriented Control (IFOC) structures are examined and asynchronous motor parameter deviations in both control structures are varied to desensitize with Artificial Neural Networks (ANN). In the literature, PI controllers are used in the IFOC structure. ANN is proposed for parameter desensitization, to the best of our knowledge no comparison and assessment has been made in the literature for these two methods. Comparisons are usually on the Direct Field Oriented Control (DFOC). This study proposes the parameter desensitization of IFOC and DTC with / without artificial neural networks and examines the effect on output performance. With the proposed control structure, it has been observed that the values of flux, torque and speed of asynchronous motor outputs capture the reference value at the desired performance and decrease the error values. With the proposed desensitization with ANN, IFOC performed over 50% better particularly in the time of overshoot and sitting than DTC. The proposed algorithms are implemented with Matlab / Simulink and the same reference values are used for each method. [ABSTRACT FROM AUTHOR]