DC–AC switching converter modelings of a PV grid-connected system under islanding phenomena

• Published in: Renewable energy Vol. 34; no. 12; pp. 2536 - 2544
• Main Authors: Chayawatto, Nattapong, Kirtikara, Krissanapong, Monyakul, Veerapol, Jivacate, Chiya, Chenvidhya, Dhirayut
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 01.12.2009; Elsevier
• Subjects: Applied sciences; Convertors; DC–AC full-bridge switching converter; Disturbances. Regulation. Protection; Electrical engineering. Electrical power engineering; Electrical machines; Electrical power engineering; Exact sciences and technology; Islanding; Power electronics, power supplies; Power networks and lines; PSpice; PV grid-connected; Regulation and control; State-space averaging; DC–AC full-bridge switching converter; PV grid-connected; Islanding; State-space averaging; PSpice; Averaging method; MATLAB software; State space; DC-AC full-bridge switching converter; Non linear equation; Switching conditions; Loss of mains; Resistive load; Mathematical model; Bridge convertor; Linearization; Inverter; Dynamic response; Large signal behavior; Experimental study; Step response; DC-AC power convertors; Current control; Voltage source converter; Analytical method; Comparative study; Dynamic load
• ISSN: 0960-1481; 1879-0682
• DOI: 10.1016/j.renene.2009.05.002

Nonlinear modeling of a dc–ac full-bridge switching converter PV grid-connected system under islanding phenomena is proposed. It is a model that can be easily derived by using simple analytical techniques. A state-space averaging technique (no linearization) and voltage source inverter with current control are performed as “large-signal modeling” that is used to analyze the dynamic response of load voltage under 3 different resistive loads: 125%, 100% and 25% of inverter output and RLC when the grid system is disconnected as well as a step change of load. The nonlinear equation from the proposed modeling is handled by MATLAB/SIMULINK. The results of the proposed model are compared with experiments and PSpice simulation which shows good agreement among them. Moreover, it is found that the proposed model consumes much less computation time than PSpice and does not encounter any convergence problem.