Analyzing the Performance of AC Microgrids in Stand-Alone Operation with Artificial Neural Network Controllers

The emergency of renewable resources-based power generation depends on the development of voltage source inverters to interface wind or solar power plants. The accurate functioning of these inverters, in turn, depends on the robust control systems capable of keep the inverters operational point in stable conditions even after non-controllable electrical contingencies. In this sense, literature has proposed the application of artificial intelligence algorithms to guarantee control robustness, however, it still lacks a detailed analysis for the application of Artificial Neural Networks (ANN) to replace linear controllers in microgrid (MG) environment. Given this theoretical frame, the main contribution of this work is to analyze ANN performance for the control of stand-alone inverters that work without the presence of the main grid. The proposed ANN-based control scheme will be is tested for stand-alone MG under different load conditions. The results obtained demonstrate the control strategy's effectiveness in managing abrupt load transitions and simulation results also exhibit that voltage or current THD is 0.71 which makes it suitable for VSIs based MGs.