The evolution of energy systems into the era of digitalisation and the green transition calls for intensive use of data-driven modeling especially models combining physics and statistics.
Grey-box models are exactly the combination of physics and statistics and can be used for characterization, forecasting and control of dynamical systems based on measured data. In other words they can form the core of digital twins of energy systems. The 4-days training course will give a hands-on introduction to non-linear time series modelling techniques. During the first day, some basics of non-linear statistical modelling will be given, e.g. kernel and splines methods, and through practical examples of model selection and validation the techniques will be exercised. Second and third day will offer grey-box modelling and forecasting, again through hands-on practical examples. The fourth day will be on flexibility functions and modelling techniques for flexible load integration in systems. Finally, the participants can choose to work on a small project, either brought by themselves or from a set of predefined challenges and datasets.
The statistical modelling software R will be used and run on the participants own laptops. The daily agenda will be a combination of lectures and exercises, such that the concepts are introduced hands on in small steps. For the exercises scripts and data will be given, together with step-by-step instructions.
For more information and registration process click here.