when I add Natural Ventilation of 12C min_out_temp_ the heat consumption is way higher.
Also the infiltration is raised. That doesn’t make much sense to me.
Actually infiltration should drop or even disappear during ventilation, doesn’t it?
The issue with ‘Natural Ventilation’ and ‘Infiltration’ I can only guess at, but I suspect that it is the same issue that you have found with the heating energy going up as you institute openable windows as a cooling strategy: as far as I can tell - and I want to have a grad student test this more thoroughly - what happens when the energyplus model is running is that if the indoor temperature is above your setpoint, then the window opens, and, depending on the ventilation model selected, the volume of outdoor air provided by the window opening size and wind pressure coefficient enters the room at say 12C! As EnergyPlus is an hour by hour calculation this could mean significantly cooling the room so it now needs heating.
What I want a grad student to check is the relationship of this to the E+ timestep. As I understand it, the energy balance being redone at, say, 6 times each hour (the default), will mean that the really cold air will only be brought into the room for 10 minutes if the indoor temperature drops too drastically. I would try your model with a timestep of say 30 times per hour (2 minute intervals) to assess the effect on cooling and heating…
Do tell me what you find
Hello @MichaelDonn
thank you for your thoughts on this topic. They correlate to what I already observed in several simulations.
During summer(warm)time something is triggering a strong heat demand when natural ventilation is used for cooling.
But before I follow your suggestion I have to clarify a mistake I made in the first post.
I just plugged in a single floor in the simulation without ventilation, therefore the result in the first chart was to low.
Here is the right one
So it looks like there is still something missing. I think what you described (partially cooling and reheating) should be part of the equation because it happens in reality and is for sure significant in spring and autumn conditions. But in the climate used in this simulation there is in reality for sure no (or almost no) heat demand in July.
I’m looking at the Berlin climate file in your GH script, and your heating demand in July makes perfect since once you plot the exterior drybulb temperature in relation to your setpoints.
You can see in the graph below that the July exterior drybulb (red line) goes far below the heating setpoint for your program of 21.7C (blue line), almost all of which you are ventilating indoors with your minimum outdoor ventilation setpoint of 12C (purple line). Even with the temperature increase from internal heat gain, that is going to trigger your heating system. For reference, I added the program cooling setpoint (green line) to illustrate where your target indoor temperatures need to be in order to avoid triggering the mechanical cooling and heating.
So all you need to do is fix your ventilation setpoints relative to your heating/cooling setpoints (i.e. bring the minimum outdoor temp to ~22C), and you’ll get the energy reduction you’re looking for.
@SaeranVasanthakumar I can understand there will be times when heating is required, even in the depths of a Berlin summer. However, I am puzzled by the recommendation that the minimum outdoor temperature should be ~22C. This is suggesting that only when the air outside is almost as warm as the air inside, will I open the windows to cool the interior? This does not seem sensible? What am I missing in your logic?
You’re right, it should be lower, to whatever the actual balance point temperature is. For a modest residential building I suspect that balance point temperature is going to be close to the 21.7C heating setpoint, but I could be wrong.
@SaeranVasanthakumar thank you for your reply and ideas. Am I right you refer to this equation?
I calculated it and if I got it right the result is in this case roundabout 17C in July.
I think this is a mayor point here. Usually you can be sure in a Berlin climate heatings are switched of anywhere latest from mid June until August/September. The massive buildings keep the heat gains easily until next morning. Usually there are no ventilation systems in use, cooling is done only by natural ventilation. But this doesn’t seem to work out same way in E+.
Yes, that’s one version of the balance point temperature, but keep in mind it’s just a guide as its still a summarization of a timeseries, so it’s not surprising it won’t give you zero heating/cooling for all of July.
I think it’s a lot clearer to simply plot the temperatures against the setpoints, and see where your zone air is activating heating. I’ve plotted temperatures for just July below to illustrate the problem. Green = Exterior air temperature, Purple = Zone air temperature, Light Green = Cooling setpoint at 24.4C, Red = Heating Setpoint at 21.7C. The separate graph below the temperatures shows the heating load (red line).
Without ventilation (just ideal loads), you can see there’s definitly some influence of midday internal heat gain that is keeping the zone air higher then the exterior air temperature, but despite that your heating load in July is much larger then your cooling load. Which means that there’s not a lot of hot temperatures to cool down in the first place.
I didn’t look at your energy balance closely enough and missed this, but the main problem is really just the heating (system) dominated summer.
The too-low minimum ventilation setpoints are, unsuprisingly, just compounding the heating in July.
Here’s the ventiltaion with a 12C minimum ventilation temperature. The zone air is at the heating setpoint the vast majority of time. At colder exterior temperatures, the zone air actually goes lower then the zone heating setpoint. Caveat: this is with very minimal parameter tuning (shutting down some windows, reducing operation area, adding a room minimum temperature setpoint), but there’s so few cooling days I don’t think it matters.
Yes, that’s the first thing you need to fix then. It sounds like you may need to massively increase your internal heat gains, and insulation to reflect the building condition you are looking for, maybe even find a more representative EPW. One parameter that seems particularly unrealistic to me is the default heating setpoint of 21.7. That’s high for european residential occupants using natural ventilation, since I think they should tolerate a broader range of temperatures.
But I don’t think there’s anything unusual in the heating and natural ventilation loads as it is currently, that’s just the logical outcome of your current EPW and BEM combination.
What I observe is the indoor temperature follows the outside temperature rather precisely. You can see the heating kicks in, the moment inside (and outside) temperature goes below 20C.
I agree with you that the outcome is logical
But I have second thoughts if the basic logic is indeed following reality?
If the inner wall is made of 1mm sheet metal and insulation. Then an indoor temperature drop like this with open windows would make sense to me.
But with an internal heavy brick wall all day loaded with solar energy I doubt the internal temperature behavior shown in this graph has something to do with reality.
I would expect a slight decline of the indoor temperature, in spring and autumn also below the heating point sure, but rarely in summer.
I also tried to minimize the fraction of the openings to 0.1, but that has hardly an effect on the internal temperature.
In my opinion what’s missing here is thermic inertia.
Or I just have no idea yet how to initiate it in a model?
It depends on the construction. If the brick wall is just a rainscreen or sufficiently insulated the thermal lag isn’t going to have as large an impact on the building. Do you have exposed thermal mass in your interior constructions? You can play around with the HB Internal Mass component to add some:
I also think you should close your windows and test to see if you get the thermal lag you’re looking for before adding ventilation. EP will add heat transfer from air (convective, infiltration, ventilation, and zone-mixed) directly to zone air, whereas the effects of conduction and radiation is first scattered, absorbed/emitted and then contributed to the zone air. So it’s possible you’re just overpowering the thermal lag due to air. I know you’ve already lowered the operable fraction, but purely from the perspective of debugging strategy, the first thing to do is simplify your parameters to find your bug, which in this case is the thermal lag not matching your desired behaviour, so removing it completely will help.
Overpowering seems the right word for what I observe. If you have a look on the left side of the picture. There is this strong peak of heating while in the cold phase afterwards there is a steady low heating towards way lower outside temperatures. Of course the windows are closed, but I still think the response in E+ to incoming air is for some reason way to high.
By the way, what mass is actually taken into account by E+? Are external and internal walls right away part of the simulation or have they or part of them to be added (later)?
You’re seeing the room air be this close even when the windows are open? Can you share your latest GH file, I’m not finding that at all when I close the windows (room air temp is purple):
I just noticed something and think you may be right that something is wrong in one of the model assumptions, but it’s not from EP, it’s from ASHRAE: the default infiltration rate for this program is really high and having a significant impact on the correlation between interior/exterior air temperature. I think you should try to find what your actual building’s infiltration rate is, because if the default assumption being used is inaccurate it’s really throwing off your simulation. Default infiltration rates are notoriously unreliable, and have high uncertainty, so I feel it’s likely this is where your model and real-world expectations are diverging.
Alternatively, you can also switch to the AFN-based natural ventilation where you have a higher degree of control in selecting what kind of leakiness you expect in your building (i.e. via the crack template). From my initial test, this seems to reduce the infiltration and thus interior/exterior air temperature correlation a lot.
By the way, what mass is actually taken into account by E+? Are external and internal walls right away part of the simulation or have they or part of them to be added (later)?
All the mass is taken into account and will contribute per time-step to the zone heat gain/loss relative to temperature difference and its specific heat capacity. My point is just that when air mass is directly filtered into the zone air (i.e. natural ventilation, air-exchanges with neighboring zones, infiltration) or convected off of surfaces, the heat gain/loss immediatly gets summed into the zone air heat balance, so the nature of airflow impact means you’ll see a more instantatenous response ot the mean zone air temperature. It’s the same reason why air-based HVAC systems can more quickly reach comfort setpoints, whereas radiant-based HVAC systems take a lot longer, especially in the prescense of high amounts of thermal mass.
There are some properties of EP that exaggerates this effect. Specifically, EP assumes each zone represents a well-mixed parcel of air (there’s no cold/hot spots) equal to the mean zone temperature, so there’s an additional bias in it zone air heat balance where all heat gain/loss from airflows are (almost) immediately reflected in the mean zone air temperature, whereas in real life the room air pressure dynamics may prevent that. This leads to inevitable zone air temperature inaccuracies if a thermal zone is too large, or the stratification of air temperature in tall zones (i.e. atriums) are important.
