Ai based pv power generation in grid tied systems with power quality enhancement

Abstract

Due to challenging climatic conditions and energy crisis, renewable energy generation comprising solar generation has exponentially increased in recent times. The high penetration level of photovoltaic (PV) production increased in smart grids (SG). Solar power is intermittent and dynamic since the solar source at the ground level is highly based on different characteristics such as atmospheric aerosol levels, cloud cover variability, and other atmosphere variables. The intrinsic changeability of large-scale solar generation poses major issues related to energy management in SGs. Since the penetration of solar PV in the SG gets increased, the solar power forecasting becomes difficult to guarantee the economical functioning of the SG. Traditionally, the solar power generation prediction approaches can be categorized into four ways namely statistical method, artificial intelligence (AI) model, physical model, and hybrid model. Practically, various prediction models are preferably based on distinct scales of prediction horizons for satisfying the needs of the decision-making process. newlineSolar PV systems integrated with a grid in the past few years using a phase-locked loop (PLL), string inverter, micro-inverter, etc. Many problems occur which are: a PLL-based clock driver costs two to five times as much as a gate-based clock driver. In PLL, VCO must run due to complex circuits and speeds. String inverter has single point failure and expandability. If the string inverter fails, then the whole system gets shut down and the rating of the inverter could not be altered once installed. The issue related to solar energy generation is the fluctuations that exist in the produced direct current because of the displacement of the sun and deviance in the number of solar rays from one position to another newline