Solar power has rapidly become an increasingly important energy source in many countries over recent years; however, the intermittent nature of photovoltaic (PV) power generation has a significant impact on existing power systems. To reduce this uncertainty and maintain system security, precise solar power forecasting methods are required. This study summarizes and compares various PV power forecasting approaches, including time-series statistical methods, physical methods, ensemble methods, and machine and deep learning methods, the last of which there is a particular focus. In addition, various optimization algorithms for model parameters are summarized, the crucial factors that influence PV power forecasts are investigated, and input selection for PV power generation forecasting models are discussed. Probabilistic forecasting is expected to play a key role in the PV power forecasting required to meet the challenges faced by modern grid systems, and so this study provides a comparative analysis of existing deterministic and probabilistic forecasting models. Additionally, the importance of data processing techniques that enhance forecasting performance are highlighted. In comparison with the extant literature, this paper addresses more of the issues concerning the application of deep and machine learning to PV power forecasting. Based on the survey results, a complete and comprehensive solar power forecasting process must include data processing and feature extraction capabilities, a powerful deep learning structure for training, and a method to evaluate the uncertainty in its predictions.

This study proposed a model for deterministic and probabilistic wind power generation forecasting and its corresponding procedures. The main contents include numerical weather prediction (NWP) systems, data preprocessing techniques, and forecasting models that use artificial intelligence methods. NWP wind speeds generated by the Central Weather Bureau (CWB) of Taiwan based on three atmospheric models, namely deterministic weather research and forecasting (WRFD), radar weather research and forecasting (RWRF), and WRF-based ensemble prediction system (WEPS), were used as model inputs. In terms of data preprocessing, the NWP wind speeds were corrected based on the height of the wind turbines, and principal components analysis (PCA) and empirical mode decomposition (EMD) were also evaluated for their feature extraction performance. Artificial neural network (ANN), long short-term memory (LSTM), gated recurrent unit (GRU), and eXtreme gradient boosting (XGBoost) were the forecast models used to predict wind power generation. The forecast results demonstrated that the use of the XGBoost model in conjunction with both PCA and EMD data preprocessing achieved the most accurate forecasting. The average forecasting errors (i.e., normalized root mean square error (NRMSE) and mean absolute percentage error (MAPE)) were 5.43% and 3.30% for one-hour-ahead and 9.78% and 6.83% for one-day-ahead, respectively. The empirical data collected at a wind farm in Taiwan verified the accuracy of the proposed method. Thus, the importance of model selection, NWP, and data preprocessing is ultimately self-evident.

With the ongoing energy transition and the increasing installation capacity of wind power generation, recent advancements in research have demonstrated that accurate numerical weather prediction (NWP) can improve the quality of wind power forecasting. While most studies have primarily focused on the effects of bias-corrected NWPs on wind speeds, few studies have explored the relationship between bias-corrected NWPs and wind power forecasts. Therefore, this study aimed to improve wind power forecasts by applying bias correction technologies to NWP-derived wind speeds. Specifically, this study established a judicious post-processing strategy to rectify NWP outputs. Three distinct NWP models—namely, the Radar Weather Research and Forecasting (RWRF), Deterministic Weather Research and Forecasting (WRFD), and a 20-member WRF ensemble prediction system (WEPS)—provided by the Taiwan Central Weather Bureau (CWB), were systematically subjected to bias correction using decaying average and probability matched mean methodologies to improve wind-speed forecasts. These bias correction techniques are particularly pertinent when dealing with limited NWP datasets. Subsequent to the bias correction of NWP products, hourly and day-ahead wind power forecasts were performed using machine learning (Long short-term memory (LSTM), Gated Recurrent Unit (GRU), and eXtreme Gradient Boosting (XGBoost)) and deep learning (Transformer) models. The results indicated that forecasting accuracy was improved after the bias correction of NWP-derived wind speeds. Comparative analyses with experiments conducted without bias correction revealed a reduction in root mean square error (RMSE) for wind speed in RWRF, WRFD, and WEPS during validation months by 46.13%, 5.84%, and 3.82%, respectively, indicating enhanced wind power prediction accuracy. Therefore, the necessity of NWP correction in a typical forecasting process was verified.

In recent years, the penetration of photovoltaic (PV) power generation in Taiwan has increased significantly. However, most photovoltaic facilities, especially for small-scale sites, do not include relevant monitoring and real-time measurement devices. The invisible power generation from these PV sites would cause a huge challenge on power system scheduling. Therefore, appropriate methods to estimate invisible PV power generation are needed. The main purpose of this paper is to propose an improved fuzzy model for estimating the PV power generation, which includes the clustering processing for PV sites, selection of representative PV sites, and the improvement of the conventional fuzzy model. First, this research uses the K-nearest neighbor (KNN) algorithm to fill in some of the missing data; then, two clustering algorithms are applied to cluster all the photovoltaic sites. Next, the relationship between the power generation of a single PV site and the total generation of all sites at the same cluster is further analyzed to select the representative PV sites. Finally, an improved fuzzy model is implemented to estimate the PV power generation. This research used actual data that were measured from PV sites in Taiwan for the estimation, verification, and comparison study. The numerical results demonstrate that the proposed method can obtain an average estimation error about 7% by using limit measurements from PV sites, highlighting the high efficiency and practicability of the proposed method.