Intelligent Modeling And Fault-tolerant Control Of Switched Reluctance Motor

Switched reluctance motor (SRM) is a new type of variable-speed drives, which wasdeveloped in1980s. SRM is gaining increasing attention in automobile and aviation industries,household appliances and servo system due to its simple construction, low cost, wide speed rangeand good fault-tolerant capability. Modeling, sensorless control and fault-tolerant control of SRMare studied in this paper.A magnetic characteristic measurement scheme with DSP TMS320F2812is presented in thispaper. Through filtering and integrating real-time phase voltage and current signals, the fluxlinkage-phase current-rotor position characteristic curves are obtained.The flux linkage characteristic of SRM is nonlinear due to its switch and reluctance, whichmakes it very difficult to derive an analytic mathematical model for SRM. Based on the fluxlinkage characteristic curves obtained from experiments, an offline Radio Basis Function (RBF)network modeling method is presented to approximate the nonlinear flux linkage characteristic. Toimprove the accuracy of RBF model, an online adjusting method for output weights is developed.Simulation results demonstrate that both offline model and online model can describe the SRMflux linkage characteristic and the online model is more accurate.To implement the simulation model in real-time control, several simplification andoptimization strategies are proposed to reduce the computing time of RBF network and keep theaccuracy of RBF network. IQmath is used to deduce the floating-point calculation; the relationshipbetween output weights and the spread factor of RBF network is studied to guarantee the evendistribution of output weights; the mapping relationship of Gaussian function between input andoutput is developed to improve the operation speed of RBF network. The experimental resultsshow that both offline model and online model can describe the SRM characteristic correctly onexperimental platform, besides, online model is more accurate.The accurate and real-time rotor position information, which is provided by a mechanicalrotor position sensor, is very important for high performance operating of SRM, but the existenceof sensor reduces the reliability of the system, especially under adverse environments. To improvethe reliability of SRM, RBF network is used in this paper to map the nonlinear function of rotorposition with respect to flux linkage and phase current. Integrating the estimate angle in eachphase, the right rotor position is shown in simulation results. The results demonstrate that RBFnetwork model can estimate the rotor position of SRM accurately.When running under adverse environments, the SRM output performance is influenced by variable faults. Fault-tolerant control of SRM under phase open-circuit fault is studied in thispaper. Based on mapping relationship of desired current-rotor position, a new current trackingcontroller is proposed. The controller enlarges torque width and reduces peak current of the othertwo healthy phases to minimize torque loss and reduce torque ripple. Furthermore, a new CTC isdeveloped based on mapping relationship of desired current-rotor position-desired speed, in whichdesired current is the output of RBF network. These measures can minimize the current and torqueripple under both fixed and variable speed conditions. Simulation results verify that the proposedfault-tolerant control scheme can work effectively.