Research On Improving The Performance Of Low-speed In Direct Torque Control

The technology of direct torque control (DTC) has got wide attention with its advantages such as new thoughts, simple control scheme and excellent performance. In this thesis the author makes a further research about the characteristic of DTC and some existing problems at low speed operation, proposes some schemes to improve the performance of the DTC system. This paper introduces a kind of modified integrator strategy to replace the pure integrator of voltage model in DTC, it can compensate for magnitude and phase angle of flux, restrain the cumulate errors of integrator and dc drift, enhancing the performance of system at low speed operation. Aim at the DTC's limitation of serious pulsating torque, a method of subdivision voltage vector is proposed. It controlled the torque by quantitative analyzing the variety of torque at field oriented control, which reduces the ripple of torque availably. The simulation result of DTC system based on MATLAB indicates that the above two modified scheme are valid. This paper also is concerned with speed sensorless. By analyzing and comparing several common algorithm of speed identification using in DTC system, a speed identification system based on and Model Reference Adaptive System is designed, which combines with the modified controlling strategy in this paper. A new PI adaptive method of speed identification based on MRAS is put forward and studied. From the simulation result, it is verified that the method has some feasibility. At last, study is also made on the implementing scheme of the digital speed sensorless DTC system. The chip of TMS320LF2407A DSP for special motor controlling is used as control core of system, the hardware structure and software flow are designed.