Data-driven modelling of operational district energy networks

There has been a resurgence of interest in district energy networks due to the cost, energy and emission reductions they can deliver. The prohibitively large computational cost of estimating thermal energy demand based on district- and urban-scale energy simulations is a major barrier to the wide-scale use of modeling in design and operation of district energy networks and in feasibility studies of technologies such as thermal energy storage. In this paper, a simple, computationally efficient modelling approach is proposed where operational data from the district energy network is used to construct temporal load profiles and thereby eliminate the need for building energy simulations. The proposed model is validated against data from a newly developed natural-gas powered district energy network in British Columbia, Canada. The utility of this data-driven approach is demonstrated through case studies on the feasibility and effectiveness of hourly thermal energy storage. It is shown that hourly thermal energy storage in a water tank can reduce the daily peak loads on the boilers by as much as 20%. Furthermore, using thermal energy storage, the highly fluctuating demand can be met with a constant-power supply, which would facilitate the use of biomass as an alternative energy source.