Stochastic modelling of central heating systems

This thesis has been prepared at the Department for Mathematical Modelling (IMM, Institut for Matematisk Modellering) at the Technical University of Denmark (DTU, Danmarks Tekniske Universitet), and at Grundfos A/S in partial fulfillment of the requirements for the degree of Ph.D. in engineering and the degree Erhvervsforsker (a special Danish degree, equivalent to ``Industrial Ph.D.''). The thesis is mainly concerned with experimental design and system identification for individual components in water based central heating systems. The main contribution to this field is on the nonlinear dynamic modelling of the following components: pipe, radiator, thermostatic valve and house. Each model is based on prior knowledge, on physical properties and experimental conditions. The estimated parameters are directly the physical model parameters. The model performance of each model has been assessed based on a number of validation methods.