Regional forecasts of photovoltaic power generation according to different data availability scenarios: a study of four methods

Abstract

The development of methods to forecast photovoltaic (PV) power generation regionally is of utmost importance to support the spread of such power systems in current power grids. The objective of this study is to propose and to evaluate methods to forecast regional PV power 1 day ahead of time and to compare their performances. Four forecast methods were regarded, of which two are new ones proposed in this study. Together, they characterize a set of forecast methods that can be applied in different scenarios regarding availability of data and infrastructure to make the forecasts. The forecast methods were based on the use of support vector regression and weather prediction data. Evaluations were performed for 1 year of hourly forecasts using data of 273 PV systems installed in two adjacent regions in Japan, Kanto, and Chubu. The results show the importance of selecting the proper forecast method regarding the region characteristics. For Chubu, the region with a variety of weather conditions, the forecast methods based on single systems' forecasts and the one based on stratified sampling provided the best results. In this case, the best annual normalized root mean square error (RMSE) and mean absolute error (MAE) were 0.25 and 0.15 kWh/kWhavg, respectively. For Kanto, with homogeneous weather conditions, the four methods performed similarly. In this case, the lowest annual forecast errors were 0.33 kWh/kWhavg for the normalized RMSE and 0.202 kWh/kWhavg for the normalized MAE. Copyright © 2014 John Wiley & Sons, Ltd.

Bayesian parameterisation of a regional photovoltaic model – Application to forecasting

“…The regional PV power estimate is then obtained by a weighted sum of the simulated power values, the weights corresponding to the frequency of occurrence of the considered configurations. The unknown weights can whether be evaluated on the basis of authors' experience Schubert (2012);Fonseca Junior et al (2015) or on the basis of a statistical analysis of PV system metadata Saint-Drenan et al 2017; Killinger et al (2018). A drawback of this approach is that possible differences between the linear coefficients chosen for the regional forecast and those corresponding to the regional estimates may penalize the forecast accuracy.…”

Photovoltaics (PV) System Energy Forecast on the Basis of the Local Weather Forecast: Problems, Uncertainties and Solutions

“…To our knowledge, there is no model on the market that relies only on the local weather forecast without the knowledge of the forecasted irradiance value. The existing short-term forecasting models start with the forecasted irradiance value [16][17][18][19]. According to [18], forecasting global solar irradiance is the same problem as forecasting the PV power output.…”

Outlier Events of Solar Forecasts for Regional Power Grid in Japan Using JMA Mesoscale Model

“…The OCCTO establishes a Japan wide of grid interconnection aiming at increasing power system security with a growing share of variable renewable energy. Recently, day-ahead forecasts of regionally integrated PV power and GHI have been performed using numerical weather prediction (NWP) models (e.g., Lorenz et al [16]; Fernandez-Jimenez et al [17]; Fonseca Jr. et al [18,19]; Jimenez et al [20]; Haupt and Kosovic [21]). However, NWP-based GHI forecasts are often associated with large errors (or outlier events).…”

“…In the U.S., seven different NWPs are blended and seamless GHI predictions are made using a solar power forecasting system called "SunCAST" (Haupt and Kosovic [21]). Oozeki et al [22] and Fonseca Jr. et al [19] estimated the errors in regional PV power generation forecasts for central Japan based Recently, day-ahead forecasts of regionally integrated PV power and GHI have been performed using numerical weather prediction (NWP) models (e.g., Lorenz et al [16]; Fernandez-Jimenez et al [17]; Fonseca Jr. et al [18,19]; Jimenez et al [20]; Haupt and Kosovic [21]). However, NWP-based GHI forecasts are often associated with large errors (or outlier events).…”

“…In the U.S., seven different NWPs are blended and seamless GHI predictions are made using a solar power forecasting system called "SunCAST" (Haupt and Kosovic [21]). Oozeki et al [22] and Fonseca Jr. et al [19] estimated the errors in regional PV power generation forecasts for central Japan based on support vector regression by using NWP. Recently, the value of day-ahead forecast was evaluated in the fields of operational electricity generation costs and economy (e.g., Martinez-Anido et al, [23]; Antonanzas et al, [24]) Martinez-Anido et al, [23] was analyzed by simulating the operation of the independent system operator under a range of scenarios with varying solar power penetrations and investigated the impact of solar power forecasting improvements.…”