Author:

Abstract:

Optimal climate control for building systems is facilitated by linear, low-order models of the building and of its heat, ventilation and air conditioning systems (HVAC), and by a quadratic or at least convex cost function for fast optimizations towards a global optimal solution. While the building envelope behavior is relatively linear (see other abstract: Optimal linearization of complex building envelope), HVAC systems are typically highly non-linear. This work presents the main difficulties in setting up a convex cost function and linear controller models for model-based controllers of complex building systems. We also propose ways to tackle non-linearities caused by radiation, latent heat, temperature dependent coefficient of performance, hydraulic losses, multiplication of control inputs, etc. by means of hierarchical MPC with sub-controllers, linearization, feedback linearization and convex envelope. The techniques are illustrated on a detailed office building modelled in Modelica©.