System identification with cross validation technique for modeling inverter of photovoltaic system

In this paper, modeling of one type of grid connected single phase inverter commercially available in Thailand is carried out. An inverter of a grid-connected photovoltaic system has been tested and its model determined. The inverter operates in six steady state conditions and its modeling is done using nonlinear system identification approach with Hammerstein Wiener Model. The models with no cross validation and cross validation data of six steady state conditions have been experimented. The results comparison on modeling by using cross validation and no cross validation data has been analyzed by model order, model accuracies and model error. The average percentage accuracy of system identification with cross validation data and no cross validation is 81.89 and 63.83 respectively.