Webinar: Smart energy systems / AI in the green transformation

Abstract

The recent report from “Klimapartnerskabet for Energi og Forsyning” shows that the use of Smart Technology e.g. smart charging of EV’s, intelligent control of heat pumps etc. can reduce the cost in grid investment towards 2030 by 30 billion DKK. Hence, there is a clear economic incitement to develop a smart energy system, but it comes with a risk regarding the resilience of the energy system.

When the end user devices become at critical part of the control of the energy system,  we introduce a new cyber risk to the system. How do we ensure that the end user devices meet a certain cyber-standard and are continuously updated to that standard in their entire lifetime.

Speaker

Jørgen S. Christensen is the CTO of Danish Energy Association and he has worked with R&D within the Danish power sector the past 25 year. The Danish Energy Association is a commercial and professional organization for Danish energy companies. The Danish Energy Association takes care of its member companies’ interests and thus works to improve conditions and competition among these companies in order to ensure development, growth and well-being in Denmark.  

Abstract

What infrastructure do we need to meet the demands of a green society? Utilities, who are the backbone of the electrification of heating and transport, are experts at spending millions on digging holes and putting hardware in the ground, but there are increasingly more affordable alternatives. These include using data-driven decision making and coordinated activation of flexibility (demand response).

Utiligize makes asset management software for utilities that predicts local electric vehicle, heat pump and photovoltaic developments, and the right investment decisions to support them. In this talk, we’ll show how machine learning is used on orthophoto data to future electric vehicle charging behaviour, and present the latest in electric vehicle trends.

Speaker

Emil Mahler Larsen has a PhD from DTU and is a former journalist at Personal Computer World and statistician at Danish Energy. He co-founded Utiligize in 2018 together with another colleague from Danish Energy to build great asset management software for utilities, a market that is expected to double over the next 10 years. Day to day, he coordinates with customers and software developers, and maintains a cloud platform that integrates with utility GIS, accounting and smart meter databases.

Abstract

Developments in AI and machine learning techniques, deep neural networks in particular, currently happen at astonishing rates, and numerous classical signal analysis tasks have been revolutionized by deep learning over the last decade. At DIKU, we aim to use these new techniques for analysis of data from a digitalized power grid for various tasks: Load forecasting at different grid levels; state, stress of failure prediction of hardware units; or automated control to ensure stable grid operation. The overall aim is enable efficient operation of the power grid with high penetration of green energy sources and high loads from electrified heating and transport sectors, thus keeping the cost of grid upgrades required by the green transition to a minimum. In the talk, we will discuss DIKU activities in this area, relevant state-of-the-art machine learning techniques, and how we envision AI and machine learning can benefit the green energy transition.

Speaker

Stefan Sommer received his M.Sc. in mathematics in 2008, and his PhD in computer science in 2012 from the University of Copenhagen (UCPH). He is currently Professor at the Department of Computer Science, UCPH. His research interests include shape analysis; statistics and machine learning on data with complex, geometric structure (geometric statistics); foundational aspects of reinforcement learning, and applications of machine learning techniques for analysis and operation of electric power grids. Stefan is heading the machine learning and data science education at UCPH.

The webinar will end with an open discussion.