Search

Photovoltaic (PV) power generation is directly effected by global horizontal irradiance (GHI) and has also large variations in spatial and/or temporal scales. For a safety control of an energy management system (EMS), a day-ahead forecast or several hour forecast of solar irradiance by a numerical weather prediction model (NWP) becomes important for a control of reserve capacity (thermal power generation, etc.). In particular, a large forecast error of PV power and/or GHI forecasts has to be prevented in the EMS. The Japan Meteorological Agency (JMA) developed two NWPs with different horizontal resolutions. First one is a mesoscale model with horizontal grid spacing of 5 km and second one is a local forecast model with that of 2 km. The two NWPs have been used as an operational model in JMA. In this study, GHI forecasts obtained from the two models are validated and conducted a case study for large forecast error (outlier events) case of GHI.

Photovoltaics (PV), which is one of variable renewable energies, have been installed largely. In the environment of largely installed PV, the output fluctuates of PV influences on the operation of power gird network adversely. Therefore, PV energy generation forecasting is valid for power grid operation and management and needs the reduction of large forecast error as well as average error. Due to a day‐ahead forecast of the total PV energy generation in the power grid area, we developed the forecasting method by using meteorological forecast data of many points in the target area. In order to forecast the regional PV energy generation, the proposed method uses auto‐encoder and convolutional neural network (CNN) for extracting requirements and valid data, and forecast with neural network (NN) by means of extracted information. In this paper, we improved the forecasting error by using multiple meteorological elements as input data. As a result, we indicated that using both of global horizontal irradiance and low cloud cover is required for the reginal PV energy generation forecasting, and large forecasting error can be reduced by adding either dew point depression, or cumulative precipitation as input data. © 2024 Institute of Electrical Engineer of Japan and Wiley Periodicals LLC. [ABSTRACT FROM AUTHOR]

Photovoltaics (PV) output prediction, which is indispensable for power system operation, can affects demand and supply adjustment adversely when large prediction error occurs. Thus, the reduction of large error as well as average error is required in PV power prediction. In 2019, the operation of the Meso‐scale Ensemble Prediction System (MEPS) of numerical weather prediction started from the JapanMeteorological Agency, and the amount of forecasting information would be potentially useful for the improvement of PV power prediction. However, very few studies on inputting multiple meteorological elements of the MEPS have been reported. In this paper, we newly develop the prediction model for an area day‐ahead PV power output composed of Just‐In‐Time Modeling (JIT Modeling) with multiple elements of theMEPS. The developed method achieves precise forecasts with low computational load by both selecting meteorological elements valid for improving prediction accuracy and adequately devising the structure of JIT Modeling. Some numerical examples demonstrating the effectiveness of the developed method are also presented. In particular, the proposed method reduces large error significantly. [ABSTRACT FROM AUTHOR]

The objective of this study is to propose and evaluate a set of modifications to enhance a machine-learning-based method for forecasting day-ahead solar irradiation. To assess the proposed modifications, they were implemented in an initial forecast method, and their effectiveness was analyzed using two years of data on a national scale in Japan. In addition, the accuracy of the modified method was compared with one of the forecast methods for solar irradiation used by the Japan Meteorological Agency (JMA), namely, the mesoscale model (MSM). Such forecasts were made publicly available only recently, which makes this study one of the first ones to compare them with machine-learning-based forecasts. The annual root-mean-square error (RMSE) of local forecasts of the JMA-MSM varied from 0.1 to 0.14 kW h m−2; the regional equivalent varied from 0.062 to 0.091 kW h m−2. In comparison with these results, the modified model achieved an average RMSE reduction of 7.5% on the local scale and 16% on the regional scale. The modified model also had a skill score that was 23% higher than that of the JMA model. Furthermore, the performance of the JMA model had strong spatial and seasonal dependencies, which were reduced in the machine-learning-based forecasts. The results show that the proposed modifications are effective in reducing large forecasts errors, but they cannot compensate for situations in which the input data used to make the forecasts are highly inaccurate. [ABSTRACT FROM AUTHOR]