HARMONIC MINIMIZATION IN A VSI FED INDUCTION MOTOR DRIVE

Abstract

The programmed pulse width modulation (PPWM) technique was originally proposed by H. S. Patel and R. G. Hoft in 1973. PPWM inverter generates high quality output spectra by eliminating several lower order harmonics, which in turn results in minimum current ripples and reduced torque pulsations in an AC drive. This technique necessitates solving a system of nonlinear transcendental equations and multiple solutions are possible. So, the main .challenge of solving these equations is the convergence, and therefore, an initial guess which is considerably close to the exact solution is required. This report discusses a simple algorithm to generate the initial guess, from the symmetry of exact switching angle trajectories at zero modulation index. Two objective functions for eliminating selected harmonics and minimizing distortion, while controlling the fundamental, are formulated. The main contribution in this work is development of a software program for minimizing the formulated objective functions using nonlinear constrained optimization functions in MATLAB and the optimal switching angles are obtained for a wide range of modulation index and number of switching angles. Usually minimization subroutines of NAG or IMSL library were employed in earlier work, which are not as easily available as MATLAB. Two sets of solutions have been identified, one in 00-600 range and other in 00-900 range. It is found that, the developed algorithm was able to find the optimal angles for all modulation indexes, even higher than unity, in a very less time. Simulation model of the VSI fed induction motor drive is developed in MATLAB-SIMULINK environment. A firing pulse - generator with digital logic is designed for converting the optimal angles into firing pulses to the inverter and the waveforms of the voltage and current at the input of the induction motor are obtained by simulating the system. The harmonic spectra of the obtained waveforms are analyzed at various operating conditions.