Short Term Prediction of Solar Irradiance Fluctuation Using Image Processing with ResNet

PV power generation has a big challenge in dealing with solar irradiance fluctuation. Several prediction methods have been proposed based on whole sky image processing techniques. This paper presents a combination technique of image processing with a convolution neural network based on ResNet for raising trigger events in advance before the sun cover happens 1 to 2 minutes in advance. The first step is captured sky image, After that image preprocessing. Next step we use Hough transform to finding the Sun position and optical flow to cloud movement tracking. Finally, use a convolution neural network to generate trigger events in advance before the sun cover happens. The results for predict stage show low error of 27.50%.