digilib@itb.ac.id +62 812 2508 8800

2018_EJRNL_PP_Ikhlaq_Hussain.pdf
Terbatas Irwan Sofiyan
» ITB

In this paper, an adaptable dual-mode single-stage solar PV (Photovoltaic) tied to a three-phase voltage-weak distribution network is proposed using multilayer perceptron (MLP) neural network control structure. The system operates in dual modes: (1) It performs as a PV supplying active power with Distributed Static Compensator (DSTATCOM) capability thereby improving power quality simultaneously and (2) it works as a DSTATCOM alone during unavailability of the SPV generation. The system is adaptable as it continues to operate satisfactorily for voltage-weak distribution grid which has frequent voltage fluctuations and flickers. A single-stage converter based system is utilized here, which incorporates an incremental conductance (INC) based approach for MPPT (Maximum Power Point Tracking) and MLP neural network structure for the extraction of fundamental components of load currents. The main advantages of MLP are its simplicity, generic nature and ease of implementation. The system and the control algorithm are validated by using MATLAB based simulation results and test results are achieved on a laboratory prototype.