- Exploit increasingly available low-cost sensor and mobile device data to learn energy consumer patterns and preferences and learn representative models of current and future energy-value (with respect to users, buildings, and equipment) for use in optimal management strategies;
- Develop the foundational models of an “information layer” to facilitate effective large-scale data collection and exploitation in renewable decentralized energy systems.
- Develop a framework to map hourly variations of renewable energy potential derived from solar and wind, as well as a novel computational tool to model and optimize grid connected hybrid renewable energy systems with vehicles to grid;
- Exploit how ubiquitous data can leverage new business models in a transition from a predominantly linear throughput industry to a closed-loop building industry, with high energy and resource savings through: (i) prolonged lifetimes of materials, (ii) direct reuse of materials/products, (iii) re-manufacturing of products, and (iv) recycling of materials;
- Use measurement data from a larger set of residential buildings and their cost-optimal retrofit measures to apply machine learning in order to infer retrofit potential and optimal set of measures with only sparse input data; upscale the approach from individual buildings to buildings stocks;
- Investigate the extent to which extensive consumer-level data can be exploited to better understand electric consumption and detect opportunities for demand side management (including energy efficiency and demand-side management).
The research activities within this work package are structured into three modules:
- Data and Learning for Technology Advancement
- Data and Learning for Assessment of Potentials
Empa, Laboratory for Urban Energy Systems
- Automatic Control Laboratory
- Chair of Architecture and Building Systems
- Group for Sustainability and Technology
EPFL, Solar Energy and Building Physics Laboratory
University of Geneva, Chair for Energy Efficiency