PREDICTIVE CONTROL FOR INTEGRATED ROOM AUTOMATION

In order to operate buildings more energy and cost effective, predictive integrated room automation can be used instead of conventional – possibly integrated – room automation. Thereby the predictive integrated room automation controllers operate the buildings’ passive thermal storages based on predicted future disturbances (e.g. weather forecast) by making use of low cost energy sources. A specific predictive integrated room automation application is considered here: The room temperature can – technically – be controlled by heating, cooling with chiller, free cooling and blind positioning. To satisfy the thermal comfort demand, the room temperature is controlled within a defined comfort range. This is achieved by a model predictive control strategy which makes use of the passive thermal storage of the building: To reduce the energy costs the thermal capacity of the building can be loaded or unloaded with low cost energy (free cooling, solar gains influenced by blinds) as long as the room temperature remains in the comfort range. The predictive controller periodically calculates an optimal future profile of the manipulated variables while constraints on the manipulated variables and predicted disturbances are taken into account. The optimization problem is solved numerically by applying linear programming (LP) algorithms. A performance bound is determined by simulations. Furthermore conventional (nonpredictive) control strategies are compared and assessed using the performance bound as a benchmark. These analyses show that predictive control is promising to be a substantial improvement compared to non-predictive control regarding cost and energy efficiency.