Selected Research Projects

The arising smart metering infrastructure generates large amounts of data about energy consumption in the residential sector. This data contains extensive (hidden) evidence of household characteristics such as saving potential, suitability for self-supply and storage.

Based on ordinary 15-min smart meter data and further external information (e.g. geographical, socio-economical, weather, statistics), we develop and test methods to automatically derive individual household characteristics in this project. We focus on household characteristics in the field of energy efficiency (e.g. interest on renewables like eco-tarrifs for electricity) and the suitability of households for self-supply and storage. Furthermore, we will identfiy atypical consumption patterns, estimate base load, and aim to discover individual saving and load shifting potential.

Together with our implementation partner BEN Energy, we develop algorithms with smart meter data from the utility CKW, our industry partner, and test them in field. The realiziation of the project will follow Swiss and European data protection regulations.

We expect validated and highly scalable methods as a project result that discover energy saving potential in the residential area that helps utilities to improve their sales strategy for green and sustainable products in line with national energy strategies.

Contact: Andreas Weigert, Samuel Schöb

Duration: 01.06.2017 — 30.06.2019

Total Budget: 802'378 € (industry and public funding)

Project Partners:


Energy Data Analytics: Increasing Service Quality and Energy Efficiency in the Residential Sector

Utilities have a large customer base, yet their knowledge about individual households is small. This adversely affects both the development of innovative, household specific services and the utilities’ monetary KPIs. Our software solutions help utility companies to engage their customers in energy saving campaigns and support them in selling respective services.

In this project, we develop further and test in field experiments machine learning algorithms that infer household characteristics (apartment size, number of inhabitants and appliances, etc.) and predict the willingness to participate in efficiency or load shifting campaigns from load profiles, location information, and other existing customer data. Our tools provide customer insights at low cost and at scale, thereby solving an eminent business problem, improving the effectiveness of energy conservation campaigns, and ultimately increasing the customer value and adoption of related services.

Funded by the EU Eurostars Programme

Projektmitarbeiter an der Universität Bamberg: Konstantin Hopf

Project duration: 01.11.2015 - 30.10.2018

Project Partners:

Total Budget: 818'840 € (industry and public funding)