Building modeling: on selection of the model complexity for predictive control

Abstract Model Predictive Control has become a wide-spread solution in many industrial applications and is gaining ground in the field of energy management of the buildings. A model with good prediction properties is an ultimate condition for good performance of the predictive controller. In this paper grey box modeling and model predictive control relevant identification are used for construction of candidate models of a building. We introduce a two-stage iterative procedure for model selection: in the first stage a minimum set of disturbance inputs is formed such that the resulting model is the best with respect to defined criterion; then the second stage comprises addition of states to obtain a final minimum parameter set maximizing the model quality. The procedure stops when it makes no sense to select more complex model as it brings no more quality improvements. Finally a case study is provided where all the above mentioned approaches are investigated.