Our PhD student's work published in IEEE Transactions on Smart Grid

Miha Grabner, PhD student of Prof. Dr. Vitomir Struc, published his thesis entitled "A Global Modelling Framework for Load Forecasting in Distribution Networks" in IEEE Transactions on Smart Grid (SCI IF: 10.275).

This work was done in collaboration with Prof. Dr. Boštjan Blažič from the Faculty of Electrical Engineering, University of Ljubljana, Qingsong Wen from DAMO Academy, Alibaba Group (U.S) Inc. and Yi Wang from the University of Hong Kong.

Summary

With the increasing number of smart meter deployments, scalable and efficient load forecasting techniques are essential to ensure situational awareness for sustainability in distribution networks. Distribution networks include a large amount of different loads at different levels of aggregation, such as individual consumers, low voltage lines and substations. The development of individual (or so-called local) forecast models for each load separately is not practical, (i) because local models 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 data for all historical loads to make predictions, and (iii) are not able to adapt to changes in the behaviour of the loads without re-training. To address these issues, we propose a global modelling 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 nature of the framework significantly reduces the computational burden typically required in training multiple local forecasting models, efficiently exploits inter-series information shared between different loads, and facilitates forecasts even when historical load data is missing or load behaviour changes over time.

Link to the paper (open access).
Open source (coming soon).