Research On The Reliability Assessment And Improvement Methods Of Switched Reluctance Machine System

Owing to the simple structure,high controllability,and strong fault tolerance,switched reluctance machine(SRM)has received lots of attention from the applications with high-reliability requirements.However,as the necessary parts of the SRM system,the power converter and the detection part own higher failure rate,which greatly affects the safe and reliable operation of the SRM system.Moreover,the reliability assessment can quantitatively determine the application effect of the reliability improvement methods and then lay the foundation for maximum system-level reliability improvement.Hence,it is worthwhile to explore the reliability assessment and improvement methods for SRM system.Firstly,the quantitative reliability assessment is carried out for SRM system.For the component-level reliability assessment,a three-dimensional thermal circuit model is built for common asymmetric half-bridge power converter(AHBPC),and then the junction temperature prediction and failure rate calculation of components can be developed under different control strategies.For the system-level reliability assessment,the combination reliability assessment model of k-out-of-n:G model and Markov model is proposed to conduct the static and dynamic reliability assessment.To quantitatively enhance the reliability of SRM system,the impact of control parameters,control strategies and redundancy strategies on system-level reliability are performed,which is helpful to select the control parameters,control strategies,and optimal redundancy levels from the reliability view.The results indicate that the proposed system-level reliability assessment method can achieve the fast and accurate reliability assessment.Meanwhile,the proposed method has good universality.Then,the quantitative reliability improvement of SRM system is explored from the perspective of the fast fault diagnosis and selection of remediation strategies.By analyzing the change of current paths before and after the power switch faults in AHBPC,the relationship can be obtained between the drive signal and neutral-point voltage of corresponding bridge arm.And the inconsistency of the relationship before and after faults is proposed as the diagnostic feature.To avoid the sampling process,the low-cost diagnostic circuits are designed to achieve the microsecond-level fault diagnosis.By building the system-level reliability assessment model of SRM system with different remediation strategies,the optimal remediation strategy can be determined,which can increase the reliability by more than 28%.The experimental results indicate that the proposed method can detect the power switches faults under different control strategies,multi-level faults and multiple topologies.Meanwhile,the effectiveness of reliability-oriented remediation strategies selection can be verified.Next,the quantitative reliability enhancement of SRM system is explored from the perspective of power converter topology.In order to lower the cost of SRM system,a novel integrated power converter is proposed.And the basic working models are presented,which validates the good control performance of proposed power converter.Considering that the power switches are the weaken parts in AHBPC,a novel control strategy is proposed to lower the electro-thermal stress of power switches on the basis of series-conduction working modes.To smooth the speed and torque ripple of SRM system after the faults of power switches,the fault-tolerance strategies are proposed for both the lower and upper power switches.And the comparison of reliability between AHBPC and proposed power converter is also carried out by setting the loose and tight failure criterion,which increases the credibility of reliability assessment.The simulation and experimental results indicate that the proposed converter would not enhance the thermal stress and lower the fault-tolerance capability,so the static reliability and dynamic reliability of SRM system can be enhanced.Finally,the reliability of SRM system is analyzed with different phase current sensing methods.By summing up the disadvantages and possible reliability problems of the existing phase current sensing methods,a two-sensor method is proposed to obtain multi-phase current information.And the suitable decoupling strategy is proposed in the whole phase current period,which can shorten the pulse injection regions and obtain the complete phase current information.By building the system-level reliability model of SRM system and combining the power loss and thermal stress distribution in different phase current sensing methods,the static reliability and dynamic reliability can be calculated.The simulation and experimental results show that the reliability of SRM system would not be enhanced by only reducing the number of current sensors.Moreover,the proposed method can enhance the reliability of SRM system without lowering the control performance.