Simplified Modelling Approach for District Heating Networks

Abstract

Several ways of simulating district heating (DH) networks exist, with differing goals and uses. In recent years, methods to reduce the spatial and calculation complexity of dynamic simulations have arisen, aiming for a good compromise between accuracy and simplicity. Two DH case studies with a discretized and non-discretized network topology and two temperature regimes (high temperature (HT) and low temperature (LT)) were analyzed in order to determine if steady state modelling is suitable for simulating the network’s thermal behaviour in comparison to dynamic approaches. One case study belongs to the ENaQ project and the other one is a study conducted by the Technical University of Munich (TUM). Four models were developed using TESPy as a simplified steady state simulation approach and were compared against their dynamic equivalent versions in Carnot Toolbox and Dymola. Results have shown that in terms of annual heat losses, TESPy overestimated the ENaQ’s HT and LT systems by 29.3% and 27.4% respectively, in comparison with Carnot Toolbox. Overestimations of 6.5% and 3.5% respectively were obtained for the HT and LT systems of the TUM case study using Dymola. Moreover, an annual average outlet temperature underestimation of 6.5% and a overestimation of 0.7% was determined for the ENaQ’s HT and LT systems respectively. Similar behaviour was also observed in the TUM case study. An hourly analysis of the simulations showed that the constant overestimation calculated by the steady state modelling approach is not necessarily a function of the network’s temperature regime, but a combination of the mass flow rates and the heat loss calculation methods considered in each simulation tool. The analysis has also shown that mass flow conditions play an important role at the time of calculating heat losses and network temperatures, with a higher effect om heat losses. Therefore, special attention must be taken when low mass flow rates characterize DH networks. Additionally, it has been seen that a combination of higher mass flow rates and low temperature regimes leads to a better coefficient of determination (above 0.94 for heat losses and 0.99 for temperatures) between dynamic and steady state simulations as found in the ENaQ’s LTDH system. Finally, given the fact that the factors differing between both simulation approaches were identified, it is expected that accurate improvements could be added in future analysis in order to acquire more precise estimation and analysis in both HT and LTDH systems.