Investigation Of Control Method Of Brushless DC Motor Based On Model Prediction

The brushless DC motor is widely used in industry and daily life, corresponding topology and control strategy are becoming matured. However, in practice there are still many problems need to be solved, such as torque ripple suppression, speed sensorless control, fault-tolerant control, etc. The lack of such technologies limits the usage of brushless DC motor in high-precision, high reliability and extreme environments. In order to suppress the brushless DC motor commutation torque ripple and improve system reliability, a model prediction-based control strategy and topology approach was proposed.The thesis introduced six-switch three-phase brushless DC motor in detail, such as the working principle, common driving methods, topological structure, phase change monitoring program and the current sampling methods. It divided the motor mathematical model into commutation and non-commutation phase, which were compared and analyzed afterwards. In this thesis, the main reasons of commutation torque ripple were analyzed from the point of the change of the current model during the commutation phase, and the limitation of the traditional controller in solving this problem was explained. Take this as the premise, this thesis selected the model predictive control method to design the current loop since it has a good adaptability of model changing. This control method realized the suppression of commutation torque ripple by switching the prediction model. On the basis of theoretical analysis, MATLAB was used to build simulation model and implement the algorithm. Several simulation experiments were designed to verify the correctness of the model prediction control algorithm for six-switch three-phase topology, and other schemes were compared to illustrate its effectiveness in suppressing torque ripple.Took the logical relations between endpoint detection voltage and switch signal into consider, this thesis designed a fault diagnosis scheme. It realized the detection and isolation of switch fault in real time. This thesis also used model predictive control method to complete the design of the brushless DC motor controller, for the sake of the stable operation of four switch inverter topology.Experimental results proved that compared with the traditional control scheme, model predictive control has better torque ripple suppression ability and excellent control performance in fault tolerant topology.