Research On Iron Loss Of Switched Reluctance Starter/Generator For Hybrid Electric Vehicle

Switched reluctance starter/generator has been widely used in electric drive because of its simple structure and working principle,low cost,good fault tolerance and high reliability.Due to the serious pollution of fuel vehicles and the immature development of pure electric vehicles,hybrid vehicles are gradually being widely used.Switched reluctance starter/generator is the core of the integrated starter generator system for hybrid electric vehicles.Due to the nonlinear and non-sinusoidal characteristics of the flux density waveform in each part of the machine caused by the special doubly salient structure,the iron loss calculation of the machine is complex,and the iron loss is directly related to the efficiency.Therefore,the iron loss research of switched reluctance starter / generator is of great significance.Considering that the analytical method of magnetic circuit cannot distinguish the direction of transient magnetic flux density waveforms and the slow calculation speed,poor universality of transient finite element method,this thesis proposes a look-up table method based on the static magnetic flux density waveforms to solve the transient magnetic flux density waveforms.This method applies static finite element analysis to solve the static magnetic flux density waveforms under the condition of separate conduction of different phases,then the nonlinear model of the machine is established in MATLAB / Simulink.According to the current and rotor position of the nonlinear model,the static magnetic density waveforms of representative parts are looked up and the transient magnetic flux density waveforms of representative parts are obtained.Then the calculation formula of Bertotti’s three-term separation iron loss with variable coefficients is improved,and a method using Fourier-fitting is proposed to fit the hysteresis loss coefficient,Steinmetz coefficient and stray loss coefficient,respectively.Considering the influence of minor hysteresis loop on hysteresis loss and skin effect on eddy current loss,the accuracy of iron loss model is further improved.According to the different characteristics of periodic variation in starting stage and periodic stability in power generation stage,the iron loss models in time domain and frequency domain are established respectively,which are suitable for the starting stage and power generation stage of SRS/G respectively.Then,the integrated simulation model of SRS/G is established in MATLAB /Simulink,and the transient magnetic density solving modules of each part of the machine are set up.The time-domain iron loss module and the frequency-domain iron loss module are established to solve the iron loss in the starting stage and the power generation stage,respectively.According to the solution results,the influences of load,conducting angle and rotational speed on the iron loss of SRS/G are analyzed.Afterwards,several groups of different turn-on angle,conducting angle and efficiency are obtained by simulation.The relationship between turn-on angle,conducting angle and efficiency is predicted by wavelet neural network modeling,and the turn-on angle and turn-off angle of maximum efficiency are obtained by genetic algorithm.Finally,a prototype experimental platform is built.The iron loss of the prototype is measured by the elimination method,and the efficiency of the power generation stage is also measured.By comparing the measured value with the simulation value,it is found that the error of the experimental results is small,which verifies the correctness of the loss model and the maximum efficiency prediction model of the turn-on and turnoff angle.There are 57 figures,14 tables and 91 references in the thesis.