Research Area:  Metaheuristic Computing

Abstract:

In this paper, a hybrid control topology is proposed for cascaded multilevel inverter (CMLI) with a grid-connected hybrid system involves wind and photovoltaic generation subsystem. The proposed hybrid control technique is the joint execution of Reptile Search Algorithm (RSA) and Gradient Boosting Decision Tree (GBDT) algorithm thus it is called RSA-GBDT approach. The purpose of proposed approach is to achieve load power demand and manage power regulation. Initially, the Cascade Multilevel Inverter (CMLI) is designed to obtain optimal control signal through proposed controller. The proposed control technique is utilized in two phases of the proposed system. In the first phase, RSA is used to evaluate the optimal gain parameters considering the range of source currents is also utilized to generate the optimal offline control signal data set. According to data set, in the second phase, the GBDT works and forecasts the most optimal control signals from the online CMLI. Finally, the resulting control signal is used to control CMLI-s Insulated Gate Bipolar Transistor (IGBT). The proposed RSA-GBDT technique develops the operation mode for solar and wind generation system discovers the switching states of converter. The proposed control system minimizes the variation of system parameters and external disturbances. Also, the load demand is achieved and the proposed RSA-GBDT control approach is executed on MATLAB/Simulink platform and performance is analyzed.

Keywords:  
Renewable energy
cascaded multilevel inverter
Reptile Search Algorithm
Gradient Boosting Decision Tree Algorithm

Author(s) Name:  K. Ranjith Kumar, M. Venkatesan, R. Saravanan

Publisher name:  Elsevier

DOI:  10.1016/j.solener.2022.07.021

Volume Information:  Volume 242, August 2022, Pages 323-334

Paper Link:   https://www.sciencedirect.com/science/article/abs/pii/S0038092X22005011