Research on PV array output characteristics based on shadow image recognition

The randomly varying power output of PV generation has been a big problem in the operation of grid connected PV stations. Shadow is one of the main cause of output power variation, but it is often costly to detect the arbitrary shaped PV shadow by a large number of sensing elements with high accuracy. In this paper, the output characteristics of PV array is obtained by combining shadow image recognition and output curve simulation. Based on the analysis of partial shadow influence on the PV module's output, a fast PV array simulation method specific to every PV cell is proposed. Considering the features of PV array images, a shadow recognition algorithm is designed, which contains technologies of grid lines smoothing and local threshold segmentation. Simulation results on actually measured PV array images indicate that the proposed algorithm is feasible in PV shadow recognition, and the output characteristic of PV array under partial shadow can be acquired quickly and accurately by the new simplified simulation technology. This research achievement has the potential of lowering the requirement of maximum power point tracking (MPPT) and improving the forecast accuracy of PV generation.