The design of future energy systems which can cope with fluctuating supply and flexible demand is an important societal concern. An essential aspect is the consumption of energy, particularly of complex systems such as factories or IT infrastructures. Important points are the flexibilization of energy consumption, so that the share of locally generated 'green' energy increases, robustness of energy provisioning, or the efficient design of new energy systems serving these purposes. To accomplish this, a core prerequisite is a structured collection, storage and analysis of energy status data. Energy status data describes the provisioning of energy, its storage, transmission and consumption, be it the outcomes of measurements, be it metadata such as the extent of fatigue of batteries, be it other relevant data such as electricity rates.
This graduate school targets at the handling of such data. To this end, an interdisciplinary approach (computer science, engineering, economics, law) is indispensable. It reveals new scientific challenges we will confront Ph.D. students with as part of their education. For instance, we have observed that different planning and control purposes require data of different temporal resolution and at different aggregation levels. This varying granularity leads to the question how to find outliers in such data at the right level of abstraction. Other graduates benefit from new approaches that detect such outliers. They can now work more efficiently, e.g., can identify shortcomings of existing models of energy systems systematically. An example of such a model would be one describing the behavior of Li-Ion batteries. The infrastructure for energy research of the KIT Helmholtz sector such as the EnergyLab 2.0 will be subject/object of our graduate school to a significant extent; the persons responsible for these facilities are part of the group of applicants of this graduate school.
An other distinctive feature of the research agenda graduates have to deal with as part of their education with us is the comprehensive treatment of the life cycle of energy status data, which consists of the phases 'collection', 'analysis' and 'deployment'. It yields a significant added value, compared to stand alone Ph.D. work that otherwise would have to cover that entire life cycle by itself: For instance, Ph.D. topics falling into an early phase of the life cycle might tailor specific methods of collecting energy status data if it is known how it will be used. Topics from the phase 'deployment' in turn, which want to design better energy systems in a data-driven fashion, can work with data of exactly the right quality.