Nonlinear time series forecast model for grid-connected photovoltaic station

Based on nonlinear characteristics of photovoltaic cells, a PV power generation forecast system is presented in this paper. A control model of grid-connected PV plant based on nonlinear neural network is proposed as well, which can effectively improve the dynamic adjustment ability of photovoltaic power station, realize the function multiplexing, improve the power quality, reduce the system loss, and save the equipment investment. Based on the nonlinearity characteristics of photovoltaic cell and grid scheduling, we develop photovoltaic cell output prediction model, photovoltaic plant output prediction model, and power control model based on the nonlinear neural network, simulate and verifies these models.