Study Of Improving Low-speed Performance Of Direct Torque Control System

This paper is to design a Direct Torque Control system, and improve its low-speed performance. This paper focuses on three areas:(1) building model and simulation in MATLAB/SIMULINK;(2) Design of experimental system platform and experiment about Direct Torque Control on the platform;(3) Put forward to a kind of artificial intelligence in DTC system. It is well known that the I-N model is used in low-speed of DTC system. Because the DTC system is a nonlinear system,and I-N mode strongly depends on parameters of motor . Robust of the system isn't so good .As result, the low speed of DTC system isn't good and the speed of motor is pulsant. This hinders the development of DTC system. U-I model is simple and good robust,meanwhile it is easily realized. Because integral of the error is large in low speed,the flux isn't accurate. In this paper U-I model is much researched and a new flux-fuzzy control method is proposed. Round flux,hexagon flux and fuzzy control are experimented in this system. The proposed theory is proved by results of experiments. At last the test prototype system is implement on a Digital signal Processor (TMS320F240). The real time control software is written in assembly language. Three methods have been experimented on a 180w motor without load.