Markov Chain Research In Double Random PWM Technique

With the rapid development of modern power electronic and digital controltechnology, study of inverter has become one of the most important issues. whicharoused more and more attention at home and abroad. Inverters are widely used in theAC/DC power conversion, the frequency control of AC motor speed, the power qualitycomprehensive treatment fields. PWM technique plays an important role in inverter. Forthe conventional PWM technique, the high amplitude of harmonic components appear inintegral multiple times of the switching frequency, which generate the annoyingelectromagnetic interference (EMI). To solve this problem, random PWM technique isproposed in this paper. The remarkable feature of adopting random PWM in inverter isthat its output harmonic spectra are dispersed and continuously distributed in the widerange. Then, the acoustic electromagnetic noise and mechanical vibration, and thepeak–peak ripple in current can be reduced.In this paper, we begin with a brief introduction to Markov chain theory, then studyrandom numbers generation from Markov chain. Conventional random numbersinevitably appear continuous multiple random numbers are greater than or less than theexpected. To deal with this problem, we adopting Markov chains in random numbergeneration to make the random numbers distribution more uniform. Based onMatlab/Simulink software platform, we build the normal random algorithm model, twostate Markov chain algorithm simulation model and three state Markov chain simulationmodel. The simulation results show that three states Markov chain algorithm generaterandom number more uniform, is better than the other models.Based on voltage space vector pulse width modulation (SVPWM), adopting threekinds of random numbers generation models in the random switching frequencySVPWM techniques and the double random SVPWM techniques. We build six kinds ofrandom SVPWM technique of the frequency control system, such as double randomSVPWM technique and random switching frequency SVPWM technique of two stateMarkov chain. The simulation results show that double random SVPWM of three statesMarkov chain is better than the other models, which make high amplitude of harmonicspectra dispersed and continuously distributed in the wide range, is an effectivealgorithm to reduce the electromagnetic noise and the mechanical vibration.