Two years ago, I wrote on this very blog about phase-change materials (PCM) and heat pumps: how to efficiently generate heat, store it, and use it later. In the THUMBS UP project, we’ve continued to build on that idea. Once thermal storage is already integrated into the building, the next question is how to use it and how to verify that the control system is delivering the results we’re looking for.
From Thermal Storage to Building Energy Management
One of the cases we have studied is the HSB Living Lab in Gothenburg, a multifamily building connected to a district heating network. The system combines heat pumps, water tanks, and FractLES thermal storage based on PCM. The pilot design provides a total of 101 kWh of nominal storage capacity, with one portion dedicated to domestic hot water and the other to heating.
Before implementing a strategy in an actual building, we can test it in a dynamic model. At CARTIF, we use TRNSYS to simulate the building’s behavior and study what happens when the way the storage is charged and discharged changes. The BEMS (Building Energy Management System) is then responsible for coordinating those decisions within the facility, using the available signals and adhering to its operating limits.
In the study, we compared a baseline system without PCM and two different strategies for controlling the same system. In both cases, the storage system is used for both domestic hot water and heating. The results show, however, a different response during operation, especially when the PCM is charged for the domestic hot water (DHW) system. We wanted to know whether a similar improvement in the power indicator of the rate actually meant that the two strategies were performing equally well.
Power Peaks: When the Daily Average Doesn’t Tell the Whole Story
To evaluate storage, we used, among other indicators, one related to the District Heating rate in Gothenburg. It is calculated based on the average power output on days of peak demand. Using this criterion, the two strategies appeared to be virtually equivalent: both reduced the indicator by about 4% compared to the scenario without storage.
But that 4% refers to a daily average. The maximum power the building draws over a short interval is a different metric. One tells us how the day plays out as a whole; the other lets us see what’s happening within that day. That’s why we review the same simulations using data collected every 15 minutes.
The difference was clear. In the simulation without storage, the maximum power observed in the 15-minute interval was about 22 kW. With the first strategy, during some PCM charging periods, demand came close to 79 kW. With the second strategy, the maximum power remained around 22 kW, practically at the same level as the building without PCM.
The explanation lies in the time scale. A short-term increase in power barely affects the 24-hour average, even though it is very noticeable when we look at the data every 15 minutes. Furthermore, both strategies completed an equivalent number of DHW storage charges. The second strategy achieved those charges with a much more moderate power response.
Energy Consumption and Costs: What Happens When You Analyze a Full Year
When the analysis was expanded to cover the entire year, the two strategies once again diverged. The first strategy reduced the average daily power indicator but increased the energy imported from the heating network by 2.6%. The second strategy maintained virtually the same improvement in that indicator and reduced the annual energy imported by about 4%.
The effect was also reflected in the cost. In the scenario used in the study—based on the rate published by Göteborg Energi—the estimated savings were about 0.5% with the first strategy and close to 3% with the second. This comparison is based on a common rate benchmark, not the building’s actual bill.
The daily indicator was fulfilling its function, but it could not, on its own, describe how the facility was performing. When looking at the same system at different time scales, the assessment of control changes substantially.
What Can We Take Away from This Analysis?
This case helps explain why the BEMS is a key component when integrating energy storage into a building. A load decision can change the demand at that moment, affect subsequent consumption, or result in more or less energy being available a few hours later. The control system must coordinate these interactions within the limits of temperature, charge status, and equipment operation.
That is why, at THUMBS UP, we evaluate the control system using more than one time scale. The indicators associated with the tariff are compared with short-term behavior and with the energy balance for the entire period. We also maintain traceability of the operation: which signals were used, which version of the control was active, and which setpoints were applied. Without that information, it is difficult to explain later why the system responded in a certain way.
This concept is useful beyond thermal storage. Something similar can happen with heat pumps, batteries, photovoltaic self-consumption, or flexible industrial processes: improving the metric we’re tracking doesn’t guarantee that the entire system is behaving as we expect.
At CARTIF, we work precisely on this process. We use models to test strategies before implementing them at the facility, develop the necessary controls to apply them, and analyze operational data to verify the results. The goal is to ensure that the flexibility provided by storage can be measured, explained, and improved using traceable information.
Thermal storage makes it possible to shift energy over time. The control system determines how to do this. The case of Gothenburg offers a very practical lesson: to determine whether a strategy works, you have to measure the results at the appropriate scale.
More about heat storage…
