With the increasing numbers of smart meter installations, scalable and efficient load forecasting techniques are critically needed to ensure sustainable situation awareness within the distribution networks. Distribution networks include a large amount of different loads at various aggregation levels, such as individual consumers, low-voltage feeders, and transformer stations. It is impractical to develop individual (or so-called local) forecasting models for each load separately. Additionally, such local models also (i) (largely) ignore the strong dependencies between different loads that might be present due to their spatial proximity and the characteristics of the distribution network, (ii) require historical data for each load to be able to make forecasts, and (iii) are incapable of adjusting to changes in the load behavior without retraining. To address these issues, we propose a global modeling framework for load forecasting in distribution networks that, unlike its local competitors, relies on a single global model to generate forecasts for a large number of loads. The global nature of the framework, significantly reduces the computational burden typically required when training multiple local forecasting models, efficiently exploits the cross-series information shared among different loads, and facilitates forecasts even when historical data for a load is missing or the behavior of a load evolves over time. To further improve on the performance of the proposed framework, an unsupervised localization mechanism and optimal ensemble construction strategy are also proposed to localize/personalize the global forecasting model to different load characteristics. Our experimental results show that the proposed framework outperforms naive benchmarks by more than 25% (in terms of Mean Absolute Error) on real-world dataset while exhibiting highly desirable characteristics when compared to the local models that are predominantly used in the literature. All source code and data are made publicly available to enable reproducibility: https://github.com/mihagrabner/GlobalModelingFramework.

There has been an increasing trend of integrating photovoltaic power plants (PVs). One of the important challenges for distribution system operators is to evaluate the total installed power of a PV that a particular network can host (or PV hosting capacity) while keeping voltage and element constraints within required limits. The major drawback of the existing methods for calculating PV hosting capacity is that they use the same installed power of the PV systems for all simulated PVs, as these methods do not use external data sources about building roofs. As a consequence, this has a significant impact on the final accuracy of the results. This paper presents a probabilistic methodology for calculating the PV hosting capacity in low voltage (LV) networks. The main contribution of this paper is the improved modeling of PV generation using actual building roof data when calculating the PV hosting capacity, as every building is treated according to its actual solar potential. Monte Carlo simulations with incorporated stochastic consumption and PV generation models are utilized for load flow calculations of the actual LV network. The simulation results presented in this paper prove that the proposed methodology increases the accuracy of the final PV hosting capacity calculations.