The following work aims at demonstrating how a smart thermostatic (radiator) valve network can be used to reduce heating costs by controlling the heat generator (heat-pump, gas boiler, ...) in a more efficient way.

Currently, a large proportion of heat generators is controlled by the means of a heating curve, or a similar rule-based logic that mostly rely on outdoor temperature or single point indoor temperature measurements. These simple control laws are in general commissioned to minimize the number of complaints of “cold users”. This results in high forward temperatures, that are energetically non-optimal as they create increased losses in the piping network and also have a negative impact on the heat generator efficiency. 

In the proposed data driven approach, a controller was developed to ensure that the radiators receive fluid with the lowest temperature possible, while satisfying the heating needs. To achieve this goal, smart thermostatic valves are used to monitor the radiator activity. The monitored information is used by a real-time algorithm to adapt the hot water temperature to continuously ensure user comfort. The solution was deployed in a multi-apartment building located in Neuchâtel (Switzerland). 

The solution has been running with success during the 2020-2021 heating season. The results point out that an average saving of 15% is obtained with respect to the baseline (i.e. heating curve) controller under similar conditions, without any degradation of comfort (under heating in particular). The system will now be deployed on 6 houses in Denmark and remain active at least until 2023.