Research On Torque Ripple Suppression In Electric Spindle Based On Multi-level Inverter By DTC

Electric spindle is the key components of NC machine tool, the dynamic performance of the drive control determines the quality of workpiece, the attention is an important index of high speed CNC machine tool design and manufacture. Direct torque control is one of the control method of high speed electric spindle, it is a research hot spot at present flow control technology. By using the conventional direct torque controlmodulation of the two level inverter,there is only one voltage vector in a control cycle,so that torque ripple increases.Therefore, it is necessary to study the electromagnetic torque ripple suppression techniques in order to improve the static and dynamic performance of BLDCM and improve its application level.First of all, the theory of multi-level inverter is reviewed.Described the principle of direct torque control based on three-level inverter and the principle of direct torque control based on five-level inverter. The topology and principle of three-level inverter and five-level inverter are introduced. The principle of neural network and fuzzy control are described, torque hysteresis comparator of a three-level multilevel inverter based on neural networks is analyzed. Based on principle of the fuzzy control, fuzzy controller is used instead of hysteresis controller in the direct torque control of five-level multilevel.In order to reduce the torque ripple in the direct torque control. The simulation model of high-speed motorized spindle is built, torque hysteresis comparator of a three-level multilevel inverter based on neural networks is designed and fuzzy controller of a five-level multilevel inverter is designed too. And then simulation studies is done through Matlab/Simulink. The simulation results show that compare with the conventional direct torque control,torque hysteresis comparator of a three-level multi-level inverter based on neural networks can reduce torque ripple by 45% in a constant torque load,the torque ripple is reduced by 40% if the torque load changes.Compare with the conventional direct torque control,fuzzy controller of a five-level multilevel inverter can reduce torque ripple by 50% in a constant torque load,it can reduce torque ripple by 45% if the torque load changes.The research show that torque hysteresis comparator of a three-level multi-level inverter based on neural networks and fuzzy controller of a five-level multilevel inverter can reduce the torque ripple, that these methods make the system to be more stable and faster,the torque response and torque controller running speed to be more faster.Meanwhile,electric spindle’s performances and speed modulation are excellent.