MPPT For Hybrid Energy System Using Machine Learning Techniques

Abstract

The share of renewable energy sources (RES) in the total energy production has been growing rapidly. With the improvement in photovoltaic panel design and manufacturing techniques the per unit cost of photovoltaic systems (PVS) generated electricity has dropped considerably. Wind Energy Conversion Systems (WECS) have also seen a steady drop in the per unit electricity generation cost. This along with the environmental benefits of RES has made them a promising technology for power generation soon. To maximize the energy yield we need to operate the PVS and WECS at the maximum power point (MPP). This ensures that at any given moment we extract the maximum possible power from the systems. The PVS and WECS can be combined together to reduce the uncertainty around the availability of any given natural resource (solar or wind) at the given time. This Hybrid System can produce more power per square meter of land and is more reliable than the separate PVS or WECS. In this paper we discuss a machine learning (ML) based approach for faster maximum power point tracking (MPPT) of a hybrid energy system (HES) and its comparison with the conventional Perturb and Observe(P&O) method.