Workshop on Optimal Control of Thermal Systems in Buildings using Modelica location:University of Freiburg date:23 - 24 March 2015
Model Predictive Control (MPC) is a well-known control strategy. However practical implementation in building applications happens only rarely. Reasons for this are that the required tools are far more complex than rule based control and that only few good demonstration cases of its potential exist. Our goal is to overcome these difficulties through the development of a methodology for implementing MPC in office buildings, applied to a state-of-the-art cases study: the Solarwind office building in Luxemburg.My presentation will first focus on the case study and the systems it contains. A qualitative description of the three major parts: building envelope, HVAC and control systems, will be given. Then the underlying mathematical description of individual component models will be discussed to provide insight in the control problem type, based on the model equations. Secondly an MPC control strategy will be proposed. The application of direct optimization methods would be too slow due to the size of the problem. Therefore the model will need to be simplified. This includes the use of 1) Model Order Reduction for the building envelope, 2) fitting performance curves to non-linear (individual or groups of) HVAC components and 3) handling pressure drops. The goal is to automate this process as much as possible using the Modelica language, coupled to python.