Control of hybrid standalone power supply system using artificial neural network

Abstract

Hybrid stand-alone systems are extensively used to supply power in different industries for a wide range of applications. In order to guarantee a steady supply of power to loads despite of variations in load, wind speed, and solar irradiation, these systems need a battery storage system. In standalone power systems, maintaining power quality is essential, particularly in systems that rely on hybrid energy sources. The battery is connected to the network using a bidirectional DC-DC converter with a suitable control mechanism. In this paper, wind turbines and multiple PV are used in parallel and series combinations to ascertain the proper rating of power supply systems. This system uses long short-term memory (LSTM) based artificial neural network (ANN) controllers. The controller for battery has been explicitly designed to guarantee that electricity is distributed equally between the load and the overall generation. Such methods can improve power quality in different areas, such as variations on the supply side from renewable sources and demand-side timescales. The performance analysis using the MATLAB/Simulink platform, and realistic results are generated by implementing Hardware-in-the-Loop through OPAL-RT modules. The results are verified with various case studies to justify the importance of adopted procedure in detail.