摘要：As a large number of photovoltaic power stations are built and put into operation, the totalamount of photovoltaic power generation accounts for an increasing proportion of the total electricity. Theinability to accurately predict solar energy output has brought great uncertainty to the grid. Therefore,predicting the future power of photovoltaic fields is of great significance. According to different time scales,predictions are divided into long- term, medium-term and ultra- short-term predictions. The main difficulty ofultra-short-term forecasting lies in the power fluctuations caused by sudden and drastic changes inenvironmental factors. The shading of clouds is directly related to the irnadiance received on the surface ofthe photovoltaic panel, which has become the main factor affecting the fluctuation of photovoltaic powergeneration. Therefore, sky images captured by conventional cameras installed near solar panels can be usedto analyze cloud characteristics and improve the accuracy of ultra-short-term predictions. This paper useshistorical power information of photovoltaic power plants and cloud image data, combined with machinelearning methods, to provide ultra- shot-term predictions of the power generation of photovoltaic powerplants. First, the random forest method is used to use historical power generation data to establish a singletime series prediction model to predict ultra-short-term power generation. Compared with the continuousmodel, the root mean square (RMSE) error of prediction is reduced by 28.38%. Secondly, the Unet networkis used to segment the cloud image, and the cloud amount information is analyzed and input into the randomforest prediction model to obtain the bivariate prediction model. The experimental results prove that, basedon the cloud amount information contained in the cloud chart, the bivariate prediction model has an 1156%increase in prediction accuracy compared with the single time series prediction model, and an increase of36. 66% compared with the continuous model.
其他摘要：As a large number of photovoltaic power stations are built and put into operation, the total amount of photovoltaic power generation accounts for an increasing proportion of the total electricity. The inability to accurately predict solar energy output has brought great uncertainty to the grid. Therefore, predicting the future power of photovoltaic fields is of great significance. According to different time scales, predictions are divided into long-term, medium-term and ultra-short-term predictions. The main difficulty of ultra-short-term forecasting lies in the power fluctuations caused by sudden and drastic changes in environmental factors. The shading of clouds is directly related to the irradiance received on the surface of the photovoltaic panel, which has become the main factor affecting the fluctuation of photovoltaic power generation. Therefore, sky images captured by conventional cameras installed near solar panels can be used to analyze cloud characteristics and improve the accuracy of ultra-short-term predictions. This paper uses historical power information of photovoltaic power plants and cloud image data, combined with machine learning methods, to provide ultra-short-term predictions of the power generation of photovoltaic power plants. First, the random forest method is used to use historical power generation data to establish a single time series prediction model to predict ultra-short-term power generation. Compared with the continuous model, the root mean square (RMSE) error of prediction is reduced by 28.38%. Secondly, the Unet network is used to segment the cloud image, and the cloud amount information is analyzed and input into the random forest prediction model to obtain the bivariate prediction model. The experimental results prove that, based on the cloud amount information contained in the cloud chart, the bivariate prediction model has an 11.56% increase in prediction accuracy compared with the single time series prediction model, and an increase of 36.66% compared with the continuous model.
