Power Quality Improvement Using AI in Power System Application

Abstract

In this research study, an Artificial Intelligence (AI)-based controller for a single-phase Voltage Source Inverter (VSI) is developed in order to improve power quality. Artificial Neural Network (ANN) technology is used in the proposed controller. The ANN model is trained using Sliding Mode Control (SMC), which is recognized for its robustness and performance. Raw data is collected from SMC to train the ANN model under a variety of load circumstances on extreme load conditions, which includes both linear and non-linear loads. The data acquired is then utilized to build the ANN model, which is essential for generating switching signals for Pulse Width Modulation (PWM) switches in VSI. The robustness feature of ANN helps the controller obtain optimal transient responses in a relatively short period, hence improving the controller’s performance and robustness. Importantly, this method does not depend on mathematical models providing a modelfree option that also helps to improve Total Harmonic Distortion (%THD); which is a critical measure of power quality. Simulation results indicate that even under extreme load circumstances the %THD remains substantially unaffected highlighting the ANN-based controller’s durability. This novel strategy has the potential to greatly improve power quality in VSI while also providing flexibility to varying load circumstances. To assess the performance of the proposed controller, a comprehensive comparison was conducted with both conventional and AI-based controllers. The results clearly indicate that the proposed controller stands out as the optimal approach for optimizing power quality in VSI. Additionally, the evaluation suggests that the proposed controller has no significant impact on power quality even with increased load demands. This underscores its stability and effectiveness in maintaining power quality under varying load conditions.