Assessing and improving the control of a small-scale fully renewables-based low-temperature thermal network in the historic city center of Bruges

Almshouses De Schipjes is a social housing neighborhood located in the historic city center of Bruges, Belgium. Between 2017 and 2020, the buildings were retrofitted to decrease the heat demand and a fully renewables-based low-temperature thermal network (supply temperature ≤ 50°C) was designed to meet the heat demand of the twelve households. The heat in the network is supplied by a centralized ground source heat pump (GSHP) connected to a borefield (8 boreholes) and thermal solar collectors (limited area to preserve the heritage character), each connected to its own central water storage tank to introduce some extra flexibility in the thermal network. The return water of the thermal network is preheated by the solar collectors in the first tank and then flows to the second tank which is heated by the GSHP to reach the supply temperature of the thermal network. The space heating in every house is provided by a combination of radiators in every room and a floor heating system on the ground floor. Domestic hot water (DHW) is produced by a booster heat pump which upgrades the low-temperature heat in the houses’ thermal network (30-40°C) to the required DHW temperature of 60°C. A simplified hydraulic scheme of the district heating network connected to one building is shown in Figure 1. The Modelica models are shown in Figure 2.
The objective of this case study is to improve the overall performance of the thermal system by assessing enhanced control strategies using dynamic simulations and applying these to the real-life thermal system. The main control rules of the reference rule-based controller (RBC) are:

• On/off GSHP control using a hysteresis curve around 50°C;
• Heating curve to determine the supply temperature in the building’s heating system;
• DHW production has priority over SH.

S. Feyaerts [1] proposed and assessed different rule-based control strategies to improve the system’s performance and concluded that the most promising adapted RBCs were:

• Preheating of DHW;
• Simultaneous space heating and DHW production;
• FH active during night;
• Heating curve for the DH network.

More details on this study can be found in the Master Thesis of S. Feyaerts (in Dutch) [1] and the paper of Boydens et al. [2].
Current research consists of applying a model predictive controller (MPC), developed using TACO [3], to the simulation model and comparing the results with the previously obtained simulation results of the RBCs.

Factsheet

Thermal zoning

Number of buildings	12
Number of thermal zones (per building)	1
Complexity of thermal zone model	High-order (white-box)
Coupling/Decoupling between district network and buildings	Coupling

Tools

Simulation tool	Dymola
Modelica libraries	Buildings/IDEAS
Additional packages/workflows/scripts	StROBe

Computational settings

Simulation time	One year
Computational time	1.5 hours
Solver and tolerance	Euler 15s
CPU speed	2.6 GHz