On The SHEPWM Technique Applied To Inverter Based On PSO

With the development of the power electronics and microelectronics, PWM inverter is applied to AC frequency control system, UPS, AC regulated power source, high and medium-frequency source and other power electronics installation. With the increasingly demanding of converter efficiency, waveform quality, electro magnetic compatibility, the inverter harmonics generated by inverter get more and more seriously to outside installation. Based on the analysis of the several PWM modulation methods utilized in inverter, selective harmonic elimination PWM which has the best effect on restraining harmonic is profoundly studied in the paper.According to the basic principles of PWM and Fourier theory, the mathematical model of output waveform of SPWM voltage source inverter is established. Through theoretical analysis and calculation, the distribution regularity of harmonic which inverter produces during variable frequency and variable voltage is obtained. The harmonic problem existing in PWM technique is found, and these above provide theory reference for valid application of SHEPWM technique. At the same time, the several calculation methods commonly used in SHEPWM equations are quantitative and systematically analyzed.In SHEPWM, equations which represent relation of fundamental component and preparatory eliminated harmonic components are a system of nonlinear transcendental equations that require the numerical iterative techniques for its solution. Numerical iterative techniques, such as Newton-Raphson method, Homotopy algorithm, Walsh function and multivariable polynomials are applied to solve the SHE problem. However, such techniques need a good initial guess that should be very close to the exact solution and computational complexity. The convergence in solution needs a good initial guess that should be very close to the exact solution.Another approach to deal with the SHE problem is based on modern stochastic search techniques such as genetic algorithm (GA) and ant colony optimization (ACO), which can avoid the stringent requirements of the initial value. However, GA is easy to trap the local optima of search space. Only the ant-colony-based schemes for continuous optimization can be used to solve SHEPWM equations, but also some necessary means are need to prevent the algorithm into a local optimal solution.Compared with ant colony algorithm, particle swarm optimization (PSO) is simpler and easier to implement. In addition, PSO is no continuity on the requirements of the problem. Based on PSO, the equation for solving SHEPWM is established. The inertia weight, acceleration coefficients are reasonable chosen to avoid the local optimal solution and ensure convergence in this paper.This PSO has been successfully applied to the SHE-PWM problem that involves five switching angles in quarter-cycle. Simulation results are provided for the inverter to validate the accuracy of computational results. Results show that all undesired harmonics up to17th order have been effectively minimized at the output voltage waveform of inverter. The THD and the switching frequency of output voltage decreased dramatically in the inverter. The simulation and experimental results are provided to validate the accuracy of the computational results.