Probabilistic modeling of future electricity systems with high renewable energy penetration using machine learning

Enhancing Photovoltaic Farm Capacity Estimation: A Comprehensive Analysis with a Novel Approach

“…Considering the variable generation of the PV unit in each hour, it is necessary to rewrite the capacity outage probability table (COPT) for each hour. By considering these changes and adding a constant load to the entire load profile, the new LOLE value of the system is calculated using Equation (2): [9]…”

“…Considering the variable generation of the PV unit in each hour, it is necessary to rewrite the capacity outage probability table (COPT) for each hour. By considering these changes and adding a constant load to the entire load profile, the new LOLE value of the system is calculated using Equation (2): [ 9 ] …”

“…[ 8 ] Additionally, probabilistic forecasting has emerged as a valuable approach for PV power forecasting, offering a range of possible outcomes instead of a single‐point prediction. [ 9 ] Combined with machine‐learning, metaheuristic optimization techniques can enhance the accuracy of PV power forecasting models and capacity value estimates by fine‐tuning model parameters to minimize prediction errors. [ 10 ] The research [ 11 ] aims to provide policymakers and energy experts with a streamlined approach to designing solar farms, addressing the imperative for renewable energy in urban areas.…”

Investigating the role of nuclear power and battery storage in Hungary's energy transition using hourly resolution electricity market simulations

“…To determine the hourly profile of consumption, the data for 2019 for neighboring countries were downloaded from the website of the European Network of Transmission System Operators for Electricity (ENTSO-E) [ 35 ], and the Hungarian data from the website of the Hungarian Electricity Transmission System Operator (TSO), called MAVIR Zrt [ 36 ]. The methodology used for data extraction was described in our previous article [ 37 ].…”

“…A more accurate estimate could only be made if we had detailed projections of weather data for 2030 and the impact of economic growth and electrification on hourly consumption, but these data were not available at the time of writing this paper and this research was not intended to produce such a dataset. In a previous paper [ 37 ], we developed a neural network-based methodology that can generate hourly-resolution consumption and renewable energy production data series based on 42 years of hourly-resolution meteorological data. These synthetic datasets could be a suitable input for such electricity market research.…”