Energy informatics and forecasting

Hjörleifur G. Bergsteinsson: Global warming is progressing faster than expected and people are getting more aware of the threat of the rising global temperature.

Therefore, sophisticated strategies for cleaner energy and robust approaches for smooth adaption to renewable energy sources are needed and developing more accurate forecast methods for the energy demand and the supply will help to reduce the energy waste. Along with the global warming awareness, there is also a trend in more accurate information about the user energy usage as there is an increase of wireless connectivity objects as the standard home appliance. The homes have become smarter and there is more data available about how the home is operated and it is possible to see the user energy usage in more detail. Hence, there are now more opportunities for both the energy providers and users to gain more information on when and how the energy (e.g., heating, electricity and gas usage) is being used. Therefore, having more data about energy usage gives the energy provider the ability to perform demand forecasting with higher accuracy. The more detailed information about energy consumption can also be used to make the end-users more aware of their energy consumption and give them with the opportunity to optimize their consumption resulting in better energy, cost and emission efficiency.

The project will consider different mathematical and statistical methods to forecast both the demand (load) for energy and the energy production for the various resource (e.g., solar, wind and hydro-power). It is essential to combine the demand and supply forecast as the optimal case is to produce exactly as much the demand requires to increase the efficiency. However, as it is impossible to forecast with a high degree of reliability and accuracy due to miscellaneous uncertainties, the forecast needs to include uncertainty in its prediction. Exploring different forecasting methods will give a better insight into the dynamics of the past historical observations leading to a decrease in the uncertainty in forecasting. Along, with the exploring the possibility of reducing the uncertainty as more information of the user energy usage is available through the smart homes censors.

PhD project

By: Hjörleifur G. Bergsteinsson

Section: Dynamical systems

Principal supervisor: Henrik Madsen

Co-supervisor: Jan Kloppenborg Møller

Project title: Energy Informatics and Forecasting

Term: 15/08/2018 → 20/11/2022 

Contact

Hjörleifur G Bergsteinsson
PhD Student
DTU Compute

Contact

Henrik Madsen
Professor, Head of section
DTU Compute
+45 45 25 34 08

Contact

Jan Kloppenborg Møller
Associate Professor
DTU Compute
+45 45 25 34 18