Job Description An emerging technology that allows the most efficient control of such a system is based on Model Predictive Control. In order to have a model that represents the system behavior it is essential that this is effectively created for any system configuration and covers the control of the refrigerant and coolant system, which can be over 10 parameters to control. This leads to a complex model and a lot of data is needed to train such a model. Purpose: The objective of the project is to create a methodology to use the existing simulation toolchain as effective as possible to create and validate the predictive models for usage in MPC. Project targets: Activity 1: Requirement analysis Overview of existing MPC specifications (MPC I/O) Consider modularity of Thermal System (System I/O) Activity 2: Model training works Perform the predictive Model training by using different methods Validate the model accuracy and document methodology