Does eQuest properly account for latent loads in infiltration & ventilation?

8 posts / 0 new
Last post

We are interested in using eQuest to evaluate the impact of improved air-sealing or changed outdoor air ventilation rates. Even with very simple test case buildings and run-of-the-mill HVAC systems, we get results that make it seem as if eQuest is not accounting for the latent component of outdoor air brought into a building.

We are using Lew Harriman et al.'s 1997 ASHRAE Journal paper on the "Ventilation Load Index" as a reality check.
The authors characterize the total load (latent and sensible) for bringing 1 cfm continuously into the building for a number of cities. It's interesting to consider two cities with comparable CDD but very different latent loads for ventilation air:

New Orleans CDD=2776 VLI=12.3 [latent] + 1.8 [sensible] =14.1 ton-hrs/yr
Tucson CDD=3017 VLI= 1.5 [latent] + 3.0 [sensible] = 4.5 ton-hrs/yr

The VLI is the cumulative load to bring 1cfm from whatever the hourly condition is in the TMY2 file to 75?F, 50% RH (65gr/lb). Its units are ton-hr/yr for convenience, converted to annual kBtu by multiplying VLI by12 kBtu/ton-h.

So, for example, if we have a building with 2,000ft2 of floor area and a volume of 20,000ft3 with a ventilation rate of 1ACH, outdoor air is being introduced to the building at 20,000/60 = 333cfm. The annual cooling & dehumidification load associated with this airflow is 333 x 14.1 = 4,695 ton-hrs/yr in New Orleans and 1,383 ton-hrs/yr in Tucson. If, for simplicity, we assume the air system brings this air to 75?F, 50% RH with a SEER of 13 kBtu/kWh in both cities, then the energy consumption associated with 1ACH for this 2000ft2 building is 4,334 kWh in New Orleans and 1,383 kWh in Tucson. The impact on building site EUI will be 7.4 kBtu/sf/yr in New Orleans and 2.4 kBtu/sf/yr in Tucson - a difference of about 5kBtu/sf/yr. If one were to consider ventilating at 10ACH, then the impact would be 10x as large-74kBtu/sf/yr in New Orleans vs 24 kBtu/sf/yr in Tucson-a difference of ~50 kBtu/sf/yr.

When we run an eQuest model of a hypothetical 2000ft2 building with no windows (to rule out solar gain differences) and no heating system, only cooling, we find that the EUI does increase with ACH at the two locations-but there is almost no difference in how the predicted EUI rises with ACH for the two cities with wildly different VLI. The EUI rises by the same ~35 kBtu/sf/yr when going from 1ACH to 10ACH at both locations.

Results summarized below for total building EUI (kBtu/sf/yr):

eQuest Simple model using VLI
ACH New Orleans Tucson New Orleans Tucson
1 42 40 36 30
10 76 74 126 74
------ ----------------------- -----------------------
1->10 34 34 90 44

(This windowless test building had R10 walls, R20 roof, R5 floor - but all we are interested in is the *difference* in total EUI associated with the increased ACH, which shouldn't depend on these choices).

Z Smith, AIA, LEED AP BD+C | Director of Sustainability & Building Performance | Eskew+Dumez+Ripple | 365 Canal Street, Suite 3150 | New Orleans, LA 70130 | 504.561.8686 |

Z Smith's picture
Joined: 2013-03-13
Reputation: 0

One note about using heating and cooling degree days for that purpose is that ventilation air is usually locked out during unoccupied times so the annual total would be much lower. From a quick read of the paper you cited it seems that they used all 8760 hours. You would need to isolate the occupied hours when the economizer is not likely to be in use to get a more accurate number.

Brendan Hall

Hall, Brendan's picture
Joined: 2011-09-30
Reputation: 1

In eQuest, we were using the '24 hours high use" setting building operation schedule, which is that the building is always occupied-so all 8760 hours. We also set the system to "no economizer".


Z Smith's picture
Joined: 2013-03-13
Reputation: 0

I would think without the ability to heat or reheat you aren't
controlling your humidity to 50% RH in the New Orleans building.
Bruce Easterbrook P.Eng

Bruce Easterbrook's picture
Joined: 2011-09-30
Reputation: 0

The VLI is meant as a method of comparing different climates and informing the designer about suitable technologies when looking at design loads, as opposed to being used as a direct predictor of energy usage.

The VLI approach used for energy calculations is also a little bit of an oversimplification to only supply room neutral air - the system in eQUEST (or any other software) would typically be supplying cooler air at the cooling coil, and then some combination of the load in the space and reheat brings the total quantity of air back to the room conditions - obviously not the same calculation as the VLI method.

If the goal is to recreate the VLI calculation, then you probably want to have a room that doesn't have ANY exterior walls, as well as removing internal gains from occupants and lights, etc. You may also want to zero out the fan power, which may be included in the SEER rating. For instance in your 1 ACH case the room may not be in control if the thermal load requires more than 1 ACH at the peak times. Another issue is the VLI is operating in terms of load, while the modeled building is using equipment efficiencies which are going to vary over the range of annual conditions. The SEER approach tries to account for that, but the modeled buildings would have varying tempeature and humidity ranges which affect the system efficiencies in different proportions - but mainly to say that using the same SEER in the hand calc won't necessarily give the same results as an energy model, unless you had also replaced all of the equipment efficiency curves.

Anyway, not knowing what the purpose of the effort is, to recreate the VLI you'll need to strip the model of many other factors, and there is likely to be a combination of environmental conditions and system efficiency causing the Arizona / Louisiana results not to behave as you expected. Start with verifying the AHU configuration and setpoints, and whether the modeled spaces are maintained at the same temperature and humidity between the two cases, secondly verify if the conditions were the temp/humity that were desired. From your description of no heating system I'd expect some variation from the constant conditions used in the VLI calculation.


David S. Eldridge, Jr., P.E., LEED AP BD+C, BEMP, BEAP, HBDP

David Eldridge's picture
Joined: 2012-05-08
Reputation: 1

Thanks to both David and Bruce, whom I'll answer together:

The beauty of the VLI metric is that it doesn't depend on how a particular HVAC system will get the ventilation air to the comfortable condition (75?F / 50%RH)-it just quantifies the enthalpy difference between the outdoor condition and the desired endpoint inside the building. It could be that an Outdoor Air unit is cooling warm moist outside air from to 55?F and then doing re-heat, or it could be that the re-heat is being provided from internal gains or solar gain-but the VLI doesn't care.

In our first message, we turned off heat for clarity. When we have also run cases where we use DX cooling + heat pump, and also DX cooling and electric resistance heat-but we always find that eQuest predicts that New Orleans and Tucson (which have similar HDD and CDD but very different VLI) should have very similar EUI, and the very high Outdoor Air rates (at which we would expect New Orleans to show off a much higher EUI) do not show a different predicted EUI for the two cities. They should have different EUI vs ACH, but they don't-they match in slope and value.

To David's point about ACH and fan power: The ACH in our discussions are not recirculating air rates but the rate of Outdoor Air. We plot up EUI vs OA ACH; you would expect fan energy for handling skin loads to fall away as an effect once we get to high ACH. Our results are very similar whether we introduce this outdoor air as deliberate ventilation or as infiltration-so the fan energy associated with actually introducing the outdoor air is not a big effect compared to the energy to condition this air.

To the question of what we're really trying to do here: When we work in hot-humid climates, we are concerned that eQuest will show the correct benefit for either decreased infiltration or improved energy recapture from systems like Enthalpy Recovery Ventilators. We have set up the Tucson / New Orleans comparison as a thought-experiment where common sense tells us that we should see a big difference but eQuest shows none.


Z Smith's picture
Joined: 2013-03-13
Reputation: 0

I have a hard time believing that eQUEST /DOE-2 has been missing this important cooling
load component (latent load from outside air) all these decades.
How are you modeling this outside air flow? In LOADS as infiltration or in SYSTEMS as
mechanical ventilation? If it's the latter, do you have the FAN-SCHEDULE
set to always on? If you're simply trying to see the change in loads from a constant
change in outdoor air flow, why don't you model it as INFILTRATION, and that
way you can see the impact on sensible and latent loads due to INFILTRATION as well as
circumvent any of the interactions between this outside air
flow and the system operations. I would be quite surprised if indeed DOE-2 has been
missing this latent component, and that you're not seeing some other
unexpected or unaccounted for effect.


Joe Huang's picture
Joined: 2011-09-30
Reputation: 406


See the LS-C report in the sim file for all your runs. It shows the sensible and latent components of all your loads, including infiltration, but excluding mechanical ventilation.

See the SS-N report for the average relative humidity of all the zones served by each of the systems.

Also see the DOE-2 help menu description for MAX-HUMIDITY. This is a system setting, typically set to "100%" by default, meaning there is no humidity control. You can only control excess humidity by 1.) using a system type that has a heating coil downstream of the cooling coil, and a humidistat, and 2.) setting the MAX-HUMIDITY to less than 100%. You can tell if you have a humidistat assigned by the little rain-cloud symbol on the system schematic.



Bill Bishop's picture
Joined: 2012-02-25
Reputation: 7