Steepest Descent Laplacian Regression Based Neural Network Approach for Optimal Operation of Grid Supportive Solar PV Generation
Bhim Singh, Nishant Kumar, Bijaya Ketan Panigrahi
Abstract
For an optimal operation of a three-phase single-stage grid-tied solar photovoltaic (PV) system, the steepest descent Laplacian regression (SDLR) based adaptive control technique is proposed in this brief. In this topology, the local loads are considered on CPI (Common Point of Interface). Therefore, the objectives of SDLR based control technique, are harmonics mitigation and power quality improvement of the grid currents. Moreover, an additional feature of DSTATCOM (Distribution Static Compensator) is also included in the control scheme. During the night, or when solar PV power generation is zero, then the voltage source converter and capacitor of DC-link are operated as DSTATCOM, which provides reactive power support to the grid. An effectiveness of SDLR based control is validated through experimentation. Here during testing, different types of adverse conditions are considered, such as solar insolation variation, unbalanced load condition, unbalances in grid voltages, etc.