TU Delft runs its campus heating system on autopilot with Gradyent’s real-time Digital Twin
Delft University of Technology (TU Delft), the oldest and largest public technical institution in the Netherlands, aims to establish a CO2-neutral, circular, and climate-adaptive campus by 2030.
To achieve this goal, TU Delft plans to use present energy sources as responsibly as possible while also making the best use of heat from a new incoming geothermal source. To this end, the institution is in the process of enhancing its heating system by renovating its substations and installing a big 26MW geothermal energy source.
Challenge
To make the best use of the multiple heat sources, TU Delft required modern heat system control, which automates heat generation, distribution, and supply for greater overall efficiency.
Solution
Following an invitation to tender, TU Delft signed a 10-year agreement with Gradyent to install a Digital Twin of its heating system, which would provide end-to-end autopilot capabilities, paving the way for faster decarbonisation and real-time optimisation of the entire system.
The Digital Twin creates a real-time digital copy of TU Delft's heating system, providing continuous visibility into production, all pipes, each substation, and individual user. It serves as the foundation for making accurate projections and optimisations for the TU Delft campus heating system in the following hours and days.
An integrated approach
The Digital Twin encompasses all subsystems, including heat sources, hydraulic network elements (pipes, distribution pumps, booster pumps, and valves), and substations. Modern district heating systems require an integrated optimisation due to the inherent interconnectedness of all aspects. Separate simulations may produce inferior or hydraulically impractical results in optimisation. The Digital Twin helps to alleviate this by accounting for the entire heating system.
User control and demand response
One part of the optimisation strategy of the Digital Twin is to use the flexibility of TSAs and buildings to redistribute demand. A trained and calibrated model will be added to the Digital Twin for each building separately and then will be linked in real-time to all distribution and generation components. This will enable demand forecasts with 15-minute intervals for the next 48 hours per building.
Enabling new energy sources
Gradyent's Digital Twin will also maximise sustainable heat production, reducing the use of expensive and gas-fired generators as much as possible. It can optimally distribute the total remaining heat demand between different generators by minimising the total required heat capacity and peak shaving. In this, the Digital Twin will prioritise generators with low CO2 emissions and heat cost price (geothermal energy).
Results
The Digital Twin will provide a 24-hour demand forecast, as well as ideal supply temperature setpoints for all manufacturing units. By modeling each structure in detail, the Digital Twin will allow TU Delft to harness building thermal inertia to maximise system flexibility and lower return temperatures as much as feasible, hence increasing geothermal efficiency.