Enhance the accuracy of control algorithm for multilevel inverter based on artificial neural network

Abstract

<p>In converters or multilevel inverters it is very important to ensure that the output of the<br />multilevel inverters waveforms in term of the voltage or current of the waveforms is<br />smooth and without distortion. The artificial neural network (ANN) technique to<br />obtaining proper switching angles sequences for a uniform step asymmetrical modified<br />multilevel inverter by eliminating specified higher-order harmonics while maintaining<br />the required fundamental voltage and current waveform. However, through this paper a<br />modified CHB-MLI are proposed using artificial intelligence optimization technique<br />based on modulation Selective Harmonic Elimination (SHE-PWM). A most powerful<br />modulation technique that used to minimize a harmonic contants during the outout<br />waveform of multilevel inverter is a SHE-PWM method. The proposed a five-level<br />Modified Cascaded H-Bridge Multilevel Inverter (M-CHBMI) with ANN controller to<br />improve the output voltage and current performance and achieve a lower Total<br />Harmonic Distortion (THD). The main aims of this paper cover design, modeling,<br />prediction for real-time generation of optimal switching angles in a single-phase<br />topology of modified five level CHB-MLI. due to the heavy cost of computation to<br />solving transcendental nonlinear equations with specified number, a real-time<br />application of Selective Harmonic Elimination-Pulse Width Modulation (SHE-PWM)<br />technique is limited. SHE equations known as a transcendental nonlinear equation that<br />contain trigonometric functions. The prototype of a 5-level inverter in Digital Signal<br />Processing (DSP) TMS320F2812 reveals that the proposed method is highly efficient<br />for harmonic reduction in modified multilevel inverter.</p>