Residential System Sizing

13 posts / 0 new
Last post

?Hello?All,

I have a problem with DOE2.1E cooling system sizing.??The cooling capacity reported by DOE2.1e in?SV-A and LS-B reports are quite different and shows?different sensitivity trends depending locations.??

Question 1
For the case with out door air flow ratio input?set to zero?DOE2.1E?calculated?cooling capacity and the peak coolig load are shown below.?? I assume that the?out door air contribution to the system load is zero for zero?out door air fraction and?with this?assumption?the system cooling capacity will be different from?the peak cooling load only due to difference in ARI?and the?peak cooling load?hour weather conditions.? What else could?cause the cooling capacity to be different from the peak load for zero out door air flow rate case?? I have looked into to DOE2.1E manual but wasn't able?connect the steps?from the peak cooling load to the system capacity??

Question 2
Does the?change from the peak cooling?load?to the ?system cooling capacity seem a reasonable for the different loacations given below?
The building?is?2100 sqft conditioned floor area,?two story, single family house.??Each floor is modeled as a single zone.

? ? Report AV-A Report LS-B ? ?
Locations Supply Flow, CFM Cooling Capacity, kBtu/hr Peak Cooling Load, kBtu/hr Outdoor DB Temp at Peak Load, oF Outdoor WB Temp at Peak Load, oF
?Baltimore, MD? 850.0 25.7 22.1 93 73
?Houston, TX? 813.0 27.0 20.2 93 78
?Chicago, IL? 1054.0 29.3 20.3 89 72
?Minneapolis, MN? 1226.0 33.5 20.3 89 79
?Duluth, MN? 1258.0 34.8 18.1 87 72
?Phoenix, AZ? 1281.0 37.4 31.6 114 71

Thank you in advance for the help,

Bereket

Bereket Nigusse's picture
Offline
Joined: 2011-10-02
Reputation: 0

Bereket:

I would not expect the values in SV-A and LS-B to agree for most buildings. There are many reasons for this.

First, peak loads from LOADS do not include ventilation loads as would be the case in SYSTEMS, only infiltration.
Second, you may be looking at different days on the weather file between LOADS and SYSTEMS, so this would
need to be checked.

In addition, DOE-2 sizes for the absolute peak, whereas other methods, such as Manual J, have some undersizing
built into the procedures, hence you'll almost always get a much larger system size in DOE-2.

Both the SV-A and LS-B lists you show seem reasonable, with the exception of the MN #s, which I suspect that you
may have a fair bit of 24 hour infiltration in your input file...(i.e., low night time temps in MN?)
However, without looking at the file I can only guess.

There are 100s of reasons for the funny little differences you show. You're probabloy better off running a Manual J
calculation on the house if it is residential. DOE-2 is known for doing a very bad job of sizing for residential.

Jeff S. Haberl, Ph.D., P.E

Jeff Haberl's picture
Offline
Joined: 2011-10-01
Reputation: 0

Another reason that the results are different is that the space temperature may be different in the "loads" part of the DOE-2 program than in the "systems" part of the program. In "loads", the space temperatures are set equal to the space design temperate. In "systems", the space temperature is based on the zone setpoint schedule and whatever temperature is actually achieved in the space. Thus, if you have a set up or a setback from the design space temperature, it will be different.

Ellen Franconi, Ph.D., LEED AP

Bereket:

I would not expect the values in SV-A and LS-B to agree for most buildings. There are many reasons for this.

First, peak loads from LOADS do not include ventilation loads as would be the case in SYSTEMS, only infiltration.
Second, you may be looking at different days on the weather file between LOADS and SYSTEMS, so this would
need to be checked.

In addition, DOE-2 sizes for the absolute peak, whereas other methods, such as Manual J, have some undersizing
built into the procedures, hence you'll almost always get a much larger system size in DOE-2.

Both the SV-A and LS-B lists you show seem reasonable, with the exception of the MN #s, which I suspect that you
may have a fair bit of 24 hour infiltration in your input file...(i.e., low night time temps in MN?)
However, without looking at the file I can only guess.

There are 100s of reasons for the funny little differences you show. You're probabloy better off running a Manual J
calculation on the house if it is residential. DOE-2 is known for doing a very bad job of sizing for residential.

Jeff S. Haberl, Ph.D., P.E

Ellen Franconi's picture
Offline
Joined: 2011-10-02
Reputation: 0

Ellen,

Not exactly. What you wrote is definitely true during the SYSTEMS simulation and peak loads reported in SS-F (?), but Bereket was referring to the SV-A, which is the systems sizing report. Assuming that he's doing "auto-sizing", the DOE-2 sizing routine will take the peak loads from LOADS and compute the design cooling capacity at the COOLING-DESIGN-TEMPERATURE. (I may have the keyword wrong, since I don't have a DOE-2 manual handy). Thus, the termperature discrepancy between the single LOADS temperature and the COOLING-DESIGN-TEMPERATURE does exist, as you've mentioned, but transient loads due to tempeature set-up or set-back or interzone heat flows are not reflected. That's why an autosized SYSTEM will often still result in a few undercooled or underheated hours.

My suggestion to Bereket in making sense of the numbers is to compare the TEMPERATURE under SPACE-CONDITIONS for the LOADS reference temperature, to the COOLING-DESIGN-TEMPERATURE (I may have this keyword wrong) is ZONE-CONDITIONS in SYSTEMS. What kind of HVAC are you modeling? It sounds like a residential system (RESYS), but that system doesn't have any OUTSIDE-AIR, does it ?

Joe Huang

Joe Huang3's picture
Offline
Joined: 2011-10-02
Reputation: 0

Jeff,

Thank you for the explanation.

Yes I am modelig a residential building and using RESYS as a system. Infiltration is specified and the zone load calculation reflects infiltration contribution to building load. Since the outdoor air fraction is set to zero, I assume that ventilation contribution to the building load or system capacity is also zero. The space temperature for load calculation is set at 75 oF while the Design-Cool-T is 78 oF. I beileive, this has the effect of decreasing the system cooling load.

Jeff the peak design load day is different from that of the peak system days for some of the loacations. However, the cooling load differences due to peak load and system peak day differences are quite small compared to the change I am looking at the peak load and system cooling capacity.

It is not obvious to me as to how the low temperature night cooling (due to 24 hour infiltration) results in in increasing the cooling capacity for the system?

Bereket

Bereket Nigusse's picture
Offline
Joined: 2011-10-02
Reputation: 0

Hi All,

I am modeling a residential building with RESYS as a system, and using auto sizing for system sizing.

The space temperature (space design temperature) for load calculation is set at 75 oF while the Design-Cool-T is set at 78 oF. I believe that this has the effect of decreasing the system cooling capacity. There is no set-back or set-up applied. No ventilation outdoor air is specified. Infiltration, which is already captured in the building load calculation, is part of the building load.

How does the system cooling capacity increases by more than 50% (in the case of Minneapolis and Duluth, Minnesota) compared to the building peak load in the absence of any out door air load for ventilation?

Thanks,

Bereket

Bereket Nigusse's picture
Offline
Joined: 2011-10-02
Reputation: 0

Bereket:

There are many other reasons could explain the differences:

1. "System-Sizing-Ratio". For many system, eQUEST uses 1.15 "System-Sizing-Ratio".
2. Fan power. Different from space load calc, the system cooling capacity has to account for fan power heat as part of the cooling loads.
3. How about latent cooling? Have you accounted for latent cooling in your comparison?
4. Make sure the zones in LS-B and SV-A report are exactly the same. Has the loads from roof/ceiling be accounted for in the LS-B report?

I am surprised to see in your results that Minneapolis, MN has slightly higher cooling loads (20.3 kBtu/hr in LS-B) than Houston, TX (20.2 kBtu/hr in LS-B).

Jeff:

In your e-mail, you mentioned that "DOE-2 is known for doing a very bad job of sizing for residential". Is there any report/paper detailing this?

Xiaobing

Xiaobing Liu2's picture
Offline
Joined: 2011-10-01
Reputation: 0

Hi all;

Can any one advice on the best HVAC system configuration for a house with a
high ceiling in the great room (other than installing a fan in the ceiling).
I have faced this problem many times. The way I was doing it is by over
sizing the AC unit for the second floor, and over sizing the furnace for the
first floor. But still, the thermostat in the second floor will not help in
making the lobby of the first floor much cooler that what it should be. The
same thing applies for the heating. Another way I was doing it is by having
one system for the entire house and run the fan continuously, but this
system is also in efficient, because I can not zone the house. Is there a
better way to do it and is there any research in this area?

Thank you.

H. Abaza

Abaza Hussein's picture
Offline
Joined: 2011-10-02
Reputation: 200

Xiaobing,

I?tried to incorporated the changes you (Xiaobing) suggested as follows but the results remained the same.?
1. The sizing ratio?was set?set to 1.0
2.? Fan power is accounted separately.? For this purpose I set the?supply air temperature rise (Supply-Delta-T = 0.) to zero
3.??The latent cooling load contributed by?occupants and infiltration has been accounted for?in?Load calculation.? The LS-B report show that.
4.? The zones in LS-B and SV-A report are different.?

My problem is that I could get to the bottom of the cause for the difference between?LS-B and SV-A reports.

Any suggestion is highly appreciated.

Bereket

Bereket Nigusse's picture
Offline
Joined: 2011-10-02
Reputation: 0

Hi All,

I have attached the input file for DOE2.1E.

Thanks,
Bereket A N

Bereket Nigusse's picture
Offline
Joined: 2011-10-02
Reputation: 0

Hussein,
Best practice for building configurations that involve temperature stratification (atriums, high ceiling spaces etc...) is to design them using CFD for load and air distribution. If you regularly come across this configuration, it might be worth it to get a proper CFD study done on it. It must include heat transfer by radiation which become very important as the difference in temperature increase between the house sufaces.

It will clarify how the air distribution performs in the space, it will show what capacity you need and where you should locate the sensors. By helping to visualise what's happening in the house and comparing it with heat loads calcs, it will give you a feeling for things that you will be able to apply to your next projects.

Best Regards

Guillaume Jolly

Guillaume Jolly's picture
Offline
Joined: 2011-10-02
Reputation: 0
Douglas Hittle's picture
Offline
Joined: 2011-10-02
Reputation: 0
Jesse Thompson's picture
Offline
Joined: 2011-10-02
Reputation: 0