A firefly optimized fast extreme learning machine based maximum power point tracking for stability analysis of microgrid with two stage photovoltaic generation system

Abstract

This paper proposes a new Firefly optimized Ridge Extreme Learning Machine (RELM-FF) algorithm based maximum power point tracking (MPPT) investigation for stability study of photovoltaic (PV) interactive microgrid dynamics. The proposed MPPT study is effective by managing minimum uncertainty limits, during solar uncertainty conditions (i.e., inconsistent irradiation, partial shading, etc.), compared with conventional techniques. A two stage converter dynamics is considered for the proposed grid interactive PV system. To withstand stability during worst grid operating scenario (both PV side and grid side uncertainty), a robust nonlinear Lyapunov based finite time sliding mode (LFSM) control is proposed for PV-VSC (voltage source converter) system. As the proposed multiple distributed energy resources based microgrid is considered as weak (low short circuit ratio), and PV is considered as an active power source, the Q-V based state variables are used for the dynamic control loop. A small signal stability analysis with the two stage PV converters is presented using the proposed RELM-FF algorithm and LFSM control. The performance evaluation of the microgrid operating with the PV array is demonstrated through various case scenarios, in MATLAB/Script environment. A hardware validation with TMS320C DSK board and embedded MATLAB coder is performed to establish effectiveness of the proposed MPPT algorithm, in real time. The superiority of the proposed MPPT over the conventional techniques is validated by comparing the stability improvement performance in grid interactive PV-VSC microgrid.