Effstock 2009 location:Stockholm, Sweden date:14-17 June 2009
Controlling borehole dynamic response and maximizing efficiency requires an accurate dynamic model to develop an efficient control strategy. Therefore, it is tested to what extent a simple linear model is able to predict the dynamic response of the mean fluid temperature in a single borehole to a given heat input. Three model types are proposed, which have been identified and evaluated through detailed TRNSYS simulations by using model Type 557. The first model type proposed is a state-space model derived from a one-dimensional resistance-capacitance (RC) network. The second and third models are rational transfer functions in the Laplace variable s and square(s), respectively. The results reveal that RC-models best describe the time scale covered by the identification data. Square(s) rational models have superior predicting and extrapolation properties, due to the fact that diffusion phenomena, which are long term effects, mathematically lead to transfer functions in square(s).