Performance Enhancement of Embedded System Based Multilevel Inverter Using Genetic Algorithm

This paper presents an optimal solution for eliminating pre specified order of harmonics from a stepped waveform of a multilevel inverter topology with equal dc sources. The main challenge of solving the associated non linear equation which are transcendental in nature and therefore have multiple solutions is the convergence of the relevant algorithms and therefore an initial point selected considerably close to the exact solution is required. The paper describes an efficient genetic algorithm that reduces significantly the computational burden resulting in fast convergence. An objective function describing a measure of effectiveness of eliminating selected order of harmonics while controlling the fundamental component is derived. The performance of cascaded multilevel inverter is compared based on computation of switching angle using Genetic Algorithm as well as conventional Newton Raphson approach. A significant improvement in harmonic profile is achieved in the GA based approach. A nine level cascaded multi level inverter is simulated in MATLAB Simulink and a proto type model has been fabricated to validate the simulation results.