Search

[Display omitted] • Easy-to-use machine learning model for hourly renewable production and load profile modeling. • Novel variance-correction method to improve the reliability of the modeled profiles. • Probabilistic energy supply simulation based on 42 years of meteorological data. • Detailed evaluation and novel categorization of Dunkelflaute events. • Uncertainty analysis of the share of photovoltaic, wind and nuclear electricity. The increasing penetration of weather-dependent renewable energy generation calls for high-resolution modeling of the possible future energy mixes to support the energy strategy and policy decisions. Simulations relying on the data of only a few years, however, are not only unreliable but also unable to quantify the uncertainty resulting from the year-to-year variability of the weather conditions. This paper presents a new method based on artificial neural networks that map the relationship between the weather data from atmospheric reanalysis and the photovoltaic and wind power generation and the electric load. The regression models are trained based on the data of the last 3 to 6 years, and then they are used to generate synthetic hourly renewable power production and load profiles for 42 years as an ensemble representation of possible outcomes in the future. The modeled profiles are post-processed by a novel variance-correction method that ensures the statistical similarity of the modeled and real data and thus the reliability of the simulation based on these profiles. The probabilistic modeling enabled by the proposed approach is demonstrated in two practical applications for the Hungarian electricity system. First, the so-called Dunkelflaute (dark doldrum) events, are analyzed and categorized. The results reveal that Dunkelflaute events most frequently happen on summer nights, and their typical duration is less than 12 h, even though events ranging through multiple days are also possible. Second, the renewable energy supply is modeled for different photovoltaic and wind turbine installed capacities. Based on our calculations, the share of the annual power consumption that weather-dependent renewable generation can directly cover is up to 60% in Hungary, even with very high installed capacities and overproduction, and higher carbon-free electricity share targets can only be achieved with an energy mix containing nuclear power and renewable sources. The proposed method can easily be extended to other countries and used in more detailed electricity market simulations in the future. [ABSTRACT FROM AUTHOR]

• Irradiance separation and transposition modeling are the two most critical steps. • 13% MAE, 12% RMSE and 23–33% skill score improvement is possible by model selection. • Detailed physical model chains, inlc. reflection and shading, have the lowest MAE. • Simple models have the lowest RMSE, but high bias and underdispersed forecast. • Wind speed has only marginal effect on physical photovoltaic power forecasting. Forecasting the power production of grid-connected photovoltaic (PV) power plants is essential for both the profitability and the prospects of the technology. Physically inspired modelling represents a common approach in calculating the expected power output from numerical weather prediction data. The model selection has a high effect on physical PV power forecasting accuracy, as the difference between the most and least accurate model chains is 13% in mean absolute error (MAE), 12% in root mean square error (RMSE), and 23–33% in skill scores for a PV plant on average. The power forecast performance analysis performed and verified for one-year 15-min resolution production data of 16 PV plants in Hungary for day-ahead and intraday time horizons on all possible combinations of nine direct and diffuse irradiance separation, ten tilted irradiance transposition, three reflection loss, five cell temperature, four PV module performance, two shading loss, and three inverter models. The two most critical calculation steps are identified as irradiance separation and transposition modelling, while the inverter models are the least important. Absolute and squared errors are two conflicting metrics, as the more detailed models result in the lowest MAE, while the simplest ones have the lowest RMSE. Wind speed forecasts have only a marginal effect on the PV power prediction. The results of this study contribute to a deeper understanding of the physical forecasting approach in the research community, while the main conclusions are also beneficial for PV plant owners in preparing their generation forecasts. [ABSTRACT FROM AUTHOR]