A fully automated transferable predictive approach was developed to predict photovoltaic (PV) power output for a forecasting horizon of 24 h. The prediction of PV power output was made with the help of a long short-term memory machine learning algorithm. The main challenge of the approach was using (1) publicly available weather reports without solar irradiance values and (2) measured PV power without any technical information about the PV system. Using this input data, the developed model can predict the power output of the investigated PV systems with adequate accuracy. The lowest seasonal mean absolute scaled error of the prediction was reached by maximum size of the training set. Transferability of the developed approach was proven by making predictions of the PV power for warm and cold periods and for two different PV systems located in Oldenburg and Munich, Germany. The PV power prediction made with publicly available weather data was compared to the predictions made with fee-based solar irradiance data. The usage of the solar irradiance data led to more accurate predictions even with a much smaller training set. Although the model with publicly available weather data needed greater training sets, it could still make adequate predictions.

Forecasting generation and demand forms the foundation of power system planning, operation, and a multitude of decision‐making processes. However, traditional deterministic forecasts lack crucial information about uncertainty. With the increasing decentralization of power systems, understanding, and quantifying uncertainty are vital for maintaining resilience. This paper introduces the uncertainty binning method (UBM), a novel approach that extends deterministic models to provide comprehensive probabilistic forecasting and thereby support informed decision‐making in energy management. The UBM offers advantages such as simplicity, low data requirements, minimal feature engineering, computational efficiency, adaptability, and ease of implementation. It addresses the demand for reliable and cost‐effective energy management system (EMS) solutions in distributed integrated local energy systems, particularly in commercial facilities. To validate its practical applicability, a case study was conducted on an integrated energy system at a logistics facility in northern Germany, focusing on the probabilistic forecasting of electricity demand, heat demand, and PV generation. The results demonstrate the UBM’s high reliability across sectors. However, low sharpness was observed in probabilistic PV generation forecasts, attributed to the low accuracy obtained by the deterministic model. Notably, the accuracy of the deterministic model significantly influences the accuracy of the UBM. Additionally, this paper addresses various challenges in popular evaluation scores for probabilistic forecasting with implementing new ones, namely a graphical calibration score, quantile calibration score (QCS), and percentage quantile calibration score (PQCS). The findings presented in this work contribute significantly to enhancing decision‐making capabilities within distributed integrated local energy systems.

Load-forecasting problems have already been widely addressed with different approaches, granularities and objectives. Recent studies focused not only on deep learning methods but also on forecasting loads on single building level. This study aims to research problems and possibilities arising by using different load-forecasting techniques to manage loads. For that behavior of two neural networks, Long Short-Term Memory and Feed-Forward Neural Network as well as two statistical methods, standardized load profiles and personalized standardized load profiles are analyzed and assessed by using a sliding-window forecast approach. The results show that personalized standardized load profiles (MAE: 3.99) can perform similar to deep learning methods (for example, LSTM MAE: 4.47). However, because of the simplistic approach, load profiles are not able to adapt to new patterns. As a case study for evaluating the support of load-forecasting for applications in energy management systems, the integration of charging stations into an existing building is simulated by using load-forecasts to schedule the charging procedures. It is shown that forecast- based controlled charging can have a significant impact by lowering overload peaks exceeding the house connection point power limit (controlled charging 20.24 kW; uncontrolled charging: 65.15 kW) while slightly increasing average charging duration. It is concluded that integration of high flexible loads can be supported by using forecast-based energy management systems with regards to their limitations.