Better insights with hourly data

We have already used hourly data implicitly in some other articles on hybrid design and active and passive cooling, but we have never really discussed the benefits compared to monthly load simulations. In this article, we discuss the insights you can gain when working with hourly load data.

What is a dynamic simulation?

For any type of geothermal simulation, you need to have a load profile of either your building or your borefield directly. In our previous article on estimating the building load (which you can find hier), we said that you can either use estimates on a yearly or monthly time scale, or you can make a dynamic simulation of your building with tools like Vabi, Linear, Energy Plus, IESVE, Vicus buildings, IDA ICE …

Since making a dynamic simulation for your building requires creating a full 3D model of the building and making assumptions for the occupancy behaviour, it is quite time-consuming and not necessarily more correct than simple rule-of-thumb (bullshit in = bullshit out), but it can be a powerful tool to learn about the behaviour of your building and answer questions like:

• Is there a risk of overheating in summer?
• What is het peak demand I really need?
• If I have automated solar shading, what would be the effect on my building demand?

Besides these architectural and MEP engineering benefits, these dynamic simulations also give you the demand profile of the building in a CSV file with an hourly resolution, which you can use in GHEtool.

Benefits of using hourly data

When you run a simulation with a monthly resolution, you enter both the peak demand and the monthly demand separately. In order to simulate the temperature profile, you need to estimate the peak duration (see also this article). This peak duration tells you how long the heat pump will operate at maximum power during one continuous run. The longer this peak duration, the more the ground will be put under stress. When you use hourly data, this information is already embedded in the data itself, which means no further estimations are needed: you can rely on the building physics directly.

As we already mentioned, using hourly data does not necessarily make your geothermal simulation more accurate (you can still make incorrect assumptions about building occupancy, for example), but it will make it more insightful. This will be demonstrated through two case studies below.

Two case studies

To compare the simulation of a temperature profile using both monthly and hourly data in GHEtool, two different buildings were selected:

• One apartment building with a collective borefield. This building has a significant imbalance in extraction due to the higher heating and domestic hot water demands.
• One auditorium building, which has a high peak cooling demand (since the system is all-air in cooling), but still an overall imbalance in extraction.

Both buildings were dynamically simulated, so we have the hourly demand profiles. Using some post-processing, this hourly data was converted into yearly (peak) demands for both heating and cooling. By designing the borefield with both datasets, we eliminate all other uncertainties and can focus solely on the design differences resulting from the use of either resolution.

Residential project

As a reference, the simulation with a monthly resolution was done first. The default monthly distribution was used, with the maximum power for heating and cooling set to 62 kW and 77 kW respectively, and yearly demands of 120 MWh and 19 MWh. The domestic hot water demand was 60 MWh/year. Using an average peak duration of 8 hours, the following profile was obtained with a 7×4 borefield and boreholes up to 150 m deep.

As was clear from the building demand itself, there is a significant imbalance in this borefield of around 116 MWh/year in extraction. However, this borefield is slightly more limited by the cooling demand in the first year (peak temperature of 16.81 °C) rather than by the peak heating demand in the 25th year (2.63 °C).

!Hinweis
One could argue that exceeding the maximum temperature limit for cooling in the first year is not a major issue, since you are cooling the ground regardless, and this problem may disappear over time. However, when working with a monthly profile, you don’t know how significant this exceedance will be. Will the temperature limit be crossed for just one hour, or for an entire week? Therefore, when using monthly data, it’s better to stick to what you do know: whether or not a certain limit is crossed, but not for how long.

When we run the same simulation with the hourly data, we see that the same borefield is indeed sufficient (so our design was correct), but for a slightly different reason. Here, the minimum peak temperature in heating is 2.09 °C, and in cooling it is 16.37 °C. Whereas before we designed the borefield based on the peak cooling demand, it is now clear that heating is the real limiting factor.

Where does this difference come from? The peak duration.

With the hourly data, we see that the peak cooling temperature only occurs for one hour, with the second-highest temperature already below 16 °C. Assuming the peak power lasted for 8 hours was, in this case, an overestimation. For the peak heating demand, the story is the opposite: since the building uses floor heating, the temperature remains low for a much longer period (as shown in the figure below). In this case, a peak duration of 8 hours for heating was actually an underestimation.

!Caution
One might think that, to overcome this difference, you could simply adjust the peak duration for heating and cooling to match the hourly demand. While this would indeed work, it only works a posteriori, once an hourly simulation has been performed. The peak duration that reflects the real behaviour of the building can only be known through simulation (or measurement, if the building already exists). This value varies from building to building, and fine-tuning it based on one project can lead to significant discrepancies in another.

Auditorium

For the auditorium building, an initial simulation was also carried out using a monthly resolution, with a peak demand of 32 kW for heating and 90 kW for cooling, and yearly demands of 38 MWh and 3.9 MWh respectively. As shown in the figure below, this borefield is clearly limited by the peak cooling demand in the first year, and 9×4 boreholes, each 150 m deep, are required to meet the building’s needs. The maximum average fluid temperature reaches 16.85 °C.

When the same simulation is performed with an hourly resolution, the profile below is obtained. In this case, the maximum temperature drops to 16.16 °C, indicating that the required borefield size was overestimated in the monthly simulation. Since the cooling system in the auditorium is all-air, the peak power is typically highly variable.

When we significantly reduce the borefield size to 7×4 boreholes (cutting the investment cost by 22%), we end up with a borefield that experiences a peak average fluid temperature of 17.28 °C — slightly above the allowed threshold. However, thanks to the hourly resolution of the data, we can see (as shown in the figure below) that this peak temperature occurs only once during the entire simulation, while the other peak temperatures remain well below 17 °C. Using this hourly resolution, we can therefore confidently reduce the required borefield size and avoid uneconomical oversizing.

Fazit

This article discusses the differences between designing a borefield using the traditional monthly resolution and designing one with hourly data. It is clear that using hourly data provides far more insight into how the building (and thus the borefield) behaves. Since there is no need to estimate the peak duration, potential over- or undersizing can be identified more easily. So, although a higher resolution does not necessarily result in a more accurate design, it does allow you to design with greater confidence and insight.

Literaturverzeichnis

• Sehen Sie sich unsere Videoerklärung auf unserer YouTube-Seite an, indem Sie klicken hier.