A stochastic dynamic method for the evaluation of the heating power input of a real building-plant system

A Prediction Error Method to set up a black-box model of a whole building-plant system based on experimental data series is proposed. This model is used to evaluate the dynamic heating power of the generator as a function of the external air temperature. In spite of the model having been built from of a limited monitored period, it is suitable for simulating the thermal performance of the system over longer periods. This allows the best power rating of the heat generator to be found; an over-dimensioned generator operates with a low load factor and a low efficiency. The proposed method can be applied very well for the substitution of heat-generators in existing plants.