Boosted by a strong solar power market, the electricity grid is exposed to risk under an increasing share of fluctuant solar power. To increase the stability of the electricity grid, an accurate solar power forecast is needed to evaluate such fluctuations. In terms of forecast, solar irradiance is the key factor of solar power generation, which is affected by atmospheric conditions, including surface meteorological variables and column integrated variables. These variables involve multiple numerical time-series and images. However, few studies have focused on the processing method of multiple data types in an inter-hour direct normal irradiance (DNI) forecast. In this study, a framework for predicting the DNI for a 10-min time horizon was developed, which included the nondimensionalization of multiple data types and time-series, development of a forecast model, and transformation of the outputs. Several atmospheric variables were considered in the forecast framework, including the historical DNI, wind speed and direction, relative humidity time-series, and ground-based cloud images. Experiments were conducted to evaluate the performance of the forecast framework. The experimental results demonstrate that the proposed method performs well with a normalized mean bias error of 0.41% and a normalized root mean square error (nRMSE) of 20.53%, and outperforms the persistent model with an improvement of 34% in the nRMSE.

In recent years, the rapid development of photovoltaic (PV) power generation has led to an increased focus on accurate forecasting of PV power output. Interval forecasting, which provides uncertainty measurement information for forecasting results, has become a hot research topic in this field. However, the accuracy of single models or traditional multi-model forecasting methods is often insufficient for meeting the forecasting requirements. In addition, interval forecasting places higher demands on the learning and generalization capabilities of forecasting models. To address these issues, this paper proposes an intelligent multi-model forecasting method based on output features clustering and convolutional neural network–long short term memory (CNN–LSTM) for PV power interval forecasting. First, multiple feature indices are constructed to represent the differences in PV output features for different months. The intelligent clustering method is then employed to achieve the multi-model clustering for forecasting model. Finally, CNN–LSTM is utilized to implement the PV power interval forecasting. The combination of CNN and LSTM effectively improves the modeling accuracy of the intelligent forecasting model. Based on actual data from PV power stations, the method described in this paper narrows down the forecasting interval compared with the single model, reducing prediction interval normalized average width by more than 4%.

The operation of the power grid will be affected inevitably with the increasing scale of wind farm due to the inherent randomness and uncertainty, so the accurate wind speed forecasting is critical for the stability of the grid operation. Typically, the traditional forecasting method does not take into account the frequency characteristics of wind speed, which cannot reflect the nature of the wind speed signal changes result from the low generality ability of the model structure. AdaBoost neural network in combination with the multi-resolution and multi-scale decomposition of wind speed is proposed to design the model structure in order to improve the forecasting accuracy and generality ability. The experimental evaluation using the data from a real wind farm in Jiangsu province is given to demonstrate the proposed strategy can improve the robust and accuracy of the forecasted variable.