Search

In this article, a novel reliability evaluation scheme to achieve quantitative reliability calculation for power converter in the whole operation region of switched reluctance motor system is proposed. First, the electrical model is built to obtain electrical stress and fault-tolerance capability in the Matlab/Simulink environment. Second, the thermal circuit model is performed to get the thermal stress of power converter quickly. Third, the dynamical Markov model with a novel failure criterion is applied to achieve quantitative reliability calculation by characterizing the fault-tolerance capability and thermal stress. The proposed failure criterion can simplify the model to decrease the number of states and avoid the divergence of results, especially the interconnected relationship between the three models is key emphasized. Next, the proposed reliability evaluation scheme is carried out to explore the reliability of power converter under different chopping modes. Finally, simulations in conjunction with experiment verify the validity of the proposed reliability evaluation scheme.

Purpose This paper aims to present a novel position sensorless control scheme with fault-tolerance ability for switched reluctance motor at low speed. Design/methodology/approach First, the detection pulses are injected in the freewheeling and idle intervals of each phase. Second, the aligned position of each phase can be detected by comparing the consecutive rise time of detection current. Third, the whole-region rotor position and real-time rotational speed can be updated four times for the improvement of detection accuracy. Finally, the fault-tolerant control strategy is performed to enhance the robustness and reliability of proposed sensorless scheme under faulty conditions. Findings Based on proposed sensorless control strategy, the estimated rotor position is in good agreement with the actual rotor position and the maximum rotor position error is 1.5°. Meanwhile, the proposed sensorless scheme is still effective when the motor with multiphase loss and the maximum rotor position error is 1.9°. Moreover, the accuracy of the rotor position estimation can be ensured even if the motor is in an accelerated state or decelerated state. Originality/value The proposed sensorless method does not require extensive memory, complicated computation and prior knowledge of the electromagnetic properties of the motor, which is easy to implement. Furthermore, it is suitable for different control strategies at low speed without negative torque generation.