Research On Low Speed Performance Of Direct Torque Control

Direct torque control (DTC) was first proposed by German scholar M.Depenbrok and Japanese scholar I.Takahashi for induction motor in the 1980s. Currently, it's develops very rapidly because of the simple control structure, good torque dynamic response and good robustness. However, the torque ripples are large at a low speed, and this poor low-speed performance makes hard to achieve the method in reality.This paper analyzes the reasons of the torque ripples at a low speed base on introduce the operation principle for the induction motor, besides, propose a new scheme fuzzy direct torque control according to these reasons the inaccurate flux observe and low switching frequency. And in the new scheme a new flux observer is designed—modified adaptive integrator without PI controller to resolve the inaccurate problem. Traditional adaptive integrator within a PI regulator and the nonlinear saturation link, and it's difficult to debug, this paper makes appropriate modifications to settle that problem—remove the PI regulator. On the other hand, the new scheme use fuzzy controller instead of the hysteresis controller to select the optimal space voltage vector, improve the switching frequency through to divide the torque error into different intervals and give different control voltages for each of them. In the fuzzy DTC system, a stator flux angle mapping technique which mapping the entire 360°range to the first sector has been defined to reduce the size of the rules base on the symmetric characteristics of the circular flux, so that the fuzzy reasoning speed increases greatly.Finally, the paper use Simulink toolbox, SimPowersystem toolbox and Fuzzy Logic Toolbox of Matlab software to build up the model of classic DTC system and fuzzy DTC system and simulation. The simulation results show that modified adaptive integrator is not having the issue of DC offset and the initial phase, the new strategy can significantly reduce the torque ripple and stabilize the speed, and the response of the fuzzy DTC is faster obviously. All this shows that the work in the paper for improving low-speed performance of DTC system is correct and reasonable.