The Improvement Of Speed Identification Method For Electric Traction System Based On Sliding Mode Observer

The electric traction system is the prime mover of high-speed EMUs and high-power electric locomotives,and its control performance determines the running performance of the whole vehicle.Accurate speed information is the precondition for ensuring the safe and reliable operation of the train.However,the speed sensor is influenced by the temperature and the installation location and other factors.The failure rate is high,which becomes a bottleneck restricting the reliability of the traction system.It also increases volume and weight,which is not conducive to the realization of energy saving,efficiency and weight reduction of the train.Therefore,research on speed sensorless technology has very important practical significance.This paper uses an asynchronous traction motor as the control object,and expounds its mathematical model and coordinate transformation principle,and lays a theoretical foundation for the full text.In the current speed identification strategy,the basic principles of the open-loop speed identification strategy and the MRAS are introduced in detail.The paper pointed out that both systems are vulnerable to parameter changes.It is difficult to ensure the accuracy of the speed identification at low switching frequency.In order to improve the robustness of the system,a speed identification scheme based on SMO was adopted.In addition,an improved fuzzy SMO speed identification strategy is proposed for the problem of system control performance degradation caused by chattering.By introducing the square term of the speed error,the switching gain of the traditional observer is improved,and the sliding mode buffeting is forced to decay by time.An adaptive PI structure based on fuzzy inference is adopted,and the corresponding controller parameters are designed fuzzy rules table.In order to solve the problem of speed identification accuracy at low switching frequency.The paper combines the improved fuzzy SMO and the MPC.The switch state constraints and delay compensation algorithms are added,and the weighted value method is used to design a variety of average switching frequencies with the value function of the goal.The paper uses Matlab/Simulink software and analyzes the control performance of each system in high and low speed areas,and emphasizes the comparison of the speed identification accuracy and system parameter robustness of each scheme under different torques and speeds.The results show that the improved fuzzy control strategy based on MPC can effectively reduce the chattering phenomenon of traditional observers,significantly improve the speed identification accuracy in the full speed range,and achieve the effect of suppressing the output torque ripple,and the system has strong parameters robustness.