Modelling diversity in single zone apartments

12 posts / 0 new
Last post

Here are a few considerations:

1. Apartments for working people will have low loads during the day and high loads at night. The offices will have the reverse pattern. Depending on the actual office schedules, the actual residential occupancy pattern and the ratio of office / residential space, 50% might work very well. 50% may also be rather close to "irresponsible" unless the local population is very tolerant of room temperatures that are a few degrees higher than normal!

2. Regardless of #1 above, a LEED project should model the building based on the architect and Engineer's "Basis of Design" (BOD) document. The scheduling and diversity patterns I mention in #1 are not commonly part of a BOD document, but in your case they sound critical. You should (strongly) request this information! If you make assumptions that differ from the Engineer's you may spend endless hours trying to reduce unmet cooling load hours (and probably will not get paid for them)

3. Once you are confident of the schedules that have been assumed by the BOD, you should be able to represent them for the energy model.

Note: Because each apartment has two fan coils, each with a thermostat, you really have two zones. This may become important for the cooling diversity.

James V Dirkes II, PE, BEMP, LEED AP

James V Dirkes II, PE's picture
Joined: 2011-10-02
Reputation: 203

You may also want to look at the multifamily building models that PNNL uses
for evaluating 90.1 changes:

(towards the bottom of the table). The 'Scorecard' XLS shows all the inputs
and where there were derived.

Drury B Crawley's picture
Joined: 2011-10-02
Reputation: 0


Thank you all for your answers. I may have confused everyone. The design
team used a diversity of 0.5 for the cooling equipment in the apartments
only. They haven't assumed this for the whole building.

Still, knowing that these are high-end apartments, I'm not entirely
convinced that people will turn cooling off when they don't occupy a space.
But if that's the assumption, I should go with it, right?

I had a look at the high-rise office/apartment model Dru pointed me to.
It's interesting to note that the apartments are modelled as single-zone
with cooling running 100% time and no setback.

So in conclusion, if I want to stick to the assumption of the design team,
I should split the apartment in 2 zones so that the FCUs in bedrooms and
living rooms can have different operation schedules.


Patrick Bivona's picture
Joined: 2012-11-12
Reputation: 0

Hi Patrick,

Another substantial source of diversity to consider, for a large building and assuming a decent thermal envelope, is that in your most extreme after-occupancy/regular operation* cooling case it's likely not every zonal unit in the building would operate simultaneously. Consider that for a simple square building, no more than 2 wall facades should be getting hit with direct sunlight at any given hour - so in turn the perimeter zones on the other two orientations would have a fraction of their design cooling loads for those hours.

I would NOT level any judgements, but 100% agree you should be collaboratively discussing with the mechanical designers understand how they came to that degree of capacity diversity so that your model matches what they are expecting to see. You may want to prepare a summary of the current model's behavior/expectations regarding chiller capacity (via an autosizing run) to help marry your sets of assumptions with respect to building occupancy/operation. As far as LEED rigor is concerned, when it comes to assumptions of design and building operation (whether in the construction documents or not), it's relatively defensible to put forward that you're matching what the engineer of record has determined for the project (whether you're discussing setbacks, setpoints, or anything similar). Be mindful many such inputs need to be applied to both baseline and proposed models uniformly.

2 systems per apartment sounds correct from the description, and I also agree from some recent experience with a mixed-use building that you'll likely want to have a separate set of occupancy and system operations schedules defined for the residential vs. nonresidential spaces, for reasons already discussed.

Best regards,


*Note from an electrical perspective, it is more typically prudent to design capacity for such cases of 100% operation, as every system may in fact need to run simultaneously, particularly after a power outage, when the building is first coming online. Sometimes BMS programming is set up to limit and/or stage system operation in consideration of electrical service size limitations, however.

Nick-Caton's picture
Joined: 2011-09-30
Reputation: 805

I don't think the issue is the separate rooms. The diversities used by the designer are most likely:

1. Different profiles for the E/S/W orientations due to solar load occurring at different times of day - should calculate within energy model

2. Each FCU is oversized compared to the peak load due to safety factors and rounding up to next size available from the factory - modeled loads should be slightly lower than the entered capacities

3. Apartments won't be occupied and/or using their appliances all at the same time - this is the tricky one that is up to you to enter in some form

4. 0.5 doesn't seem like a calculated diversity - they may be using prior projects as a guide

However if the thermostats don't have integral occpuancy controls, I wouldn't make an assumption that the residents have turned off any particular thermostat at any time, and you'll have to decide for yourself if that means changing the model to increase the quantity of zones.

You could approximate the intensity of apartment loads by having some on and some off in groups, or all of them on the same schedule using an average amount of internal gain from lights and appliances. If you are having trouble matching the plant capacity of the designers with the model, you should confirm their BOD in terms of calculating the diversity, and what quantity of internal gains they assumed, etc.

Lastly take a look at square feet per ton as a benchmark to see where the plant is coming in. Does it compare with other projects in that area?

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

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

Nick, David,

Thanks for your feedback. I can see there's room for sophistication when
wanting to calculate a proper diversity factor! After discussion with the
designers, they confirmed they considered that occupants would not be
cooling all the rooms at once and that 0.5 was not a calculated diversity
but from experience. And thermostats are under the occupants' control.

I would have been a bit more conservative but hopefully the peak load in
apartments will come (shortly) after people leave the office part. I guess
I'll be able to confirm all that once I make progress on the model and see
how much slack the chillers have around that time.


Patrick Bivona's picture
Joined: 2012-11-12
Reputation: 0

Great conversation.

One thing I didn't see mentioned in the discussion is the fact you can
diversify the internal loads within the apartment (and office spaces) to
better reflect the engineers "experience" of load diversity. This will have
a direct effect of reducing peak cooling and would save you having to split
up your apartments into separate zones.

In my experience, more zones does not always equal better results but it
always equals more headaches and a lot more work.

To setup your operating schedules to account for diversity whether it's
within a single zone or across an entire building use a weighted average
approach. For example, if the connected lighting load in the apartment is
1W/sq.ft. and the living area and the apartments have equal areas and the
lights in each area are never on at the same time then lighting power would
never be more than 0.5W/sq.ft. You would use a maximum value of 0.5 in your
lighting schedule to describe this case.

As Nick pointed out previously your diversified schedules would need to be
the same in both the baseline and proposed models.

Another thing that might help is to try and duplicate the engineers peak
cooling load calc within the model. I'm not sure what software you're using
but several tools allow you to do a "peak load calculation" for equipment
sizing that is separate from the actual 8760 simulation. If your tool has
this capability you could setup a design cooling load calculation similar to
what the engineer will do (e.g. using ASHRAE 1% CDT and all internal loads
and occupancy at 100% for the entire day). With that calculation you can
examine the theoretical building peak cooling load and then compare that to
the actual building cooling load under a diversified operating schedule.
You can then determine whether the 50% diversity is a valid assumption for
this project.

Michael Tillou, PE, BEMP, LEED AP BD+C

Michael Tillou's picture
Joined: 2011-10-02
Reputation: 0

Hi Mike,

I agree that splitting 160 apartments in 2 is work that I'd rather avoid if
I can! I was originally considering adjusting schedules to represent
diversity. That works ok for internal gains. I have trouble seeing how that
can work for cooling equipment in a single zone apartment; it's either on
or off. Having one zone with diversified internal gains and cooling on is
not the same thing, in terms of energy use, as having 2 zones, one of which
has cooling turned off. But maybe that's ok as long as both proposed and
baseline are the same?


Patrick Bivona's picture
Joined: 2012-11-12
Reputation: 0

I would think as long as you have properly accounted for the diversity via the schedules, the systems should just not run as long because there is less heat to remove from the space. The only effects you are not capturing is if only one of the apartment zones should be perimeter, but you are grouping both into a single zone that sees the combined load. That could change how often the core zone would come on versus the perimeter. At any rate, I think the cycling of the unit on less often should get you close to showing the diversity effects.

Cory Duggin, EI

crduggin's picture
Joined: 2013-04-16
Reputation: 0

Hi Cory,

Well, maybe that was just the assumption I made that needed to be
challenged. I was imagining that the combined cooling equipment in the
single zone, running at part load, would have a different energy use than
one cooling equipment running at full load in one of the two zones, while
the other is off/cycling. Maybe the difference is marginal on equipment of
the size found in apartments?


Patrick Bivona's picture
Joined: 2012-11-12
Reputation: 0

If you had DX equipment, I would be more inclined to agree that the part load performance of a combined unit would result in different energy use. I think since you have chilled water fan coils, which I assume have valves to modulate flow, the main piece of concern is the fan, and with that size system, I don't think it will make a big difference. To me the question is, how much will the energy use differ from a fan twice the size running somewhere around half the time versus two fans that are half the size not running coincidently?

Cory Duggin, EI

crduggin's picture
Joined: 2013-04-16
Reputation: 0


The 50% diversity factor for the residential occupied areas that you describe may or may not be aggressive, though it has the potential to lead to downstream problems involving lawyers. While it may be true you are not 'on the hook' for designing an undersized/underperforming engineered mechanical system design (the engineer of record is holding this bag), you have been presented a hand that if you choose to play forward, presents you with a great opportunity for showing your chops to the building owner and allows you to demonstrate the true power of hourly quasi-steady state whole building energy simulations to the community that pays our bills. I've been in residential mixed-use retrofit situations where we have downsized existing chiller capacity by more than 50% (after careful analysis of simulation output using different tools to reach consensus), but in a new construction scenario a carte blanche 50% diversity factor scenario can, if distribution piping is aggressively sized, create some problems for later easily remedying.

Once you determine what kind of internal load peaks (and appropriate time-variant schedules?) were assumed by the owner/design team, if you have the time/budget you might want to incorporate some quick investigation of schedule diversity into your proposed model to determine relative sensitivity against block load vs peak load required proposed chiller tonnage...

Some fodder for providing defendable default residential gain schedules to de-rate/deviate from to test the 50% diversity factor assumption can be found in the recently updated NREL "Building America House Simulation Protocols" document - updated in December of 2013 but still marked as a draft. This draft pdf document ships as part of the installation of the free NREL BEopt v2.1.0.0 software - released last week - get it here:

The pdf is filled with loads of simulation tool agnostic assumptions (and supporting documentation) that can be useful for generating input data for residential energy modeling activities. I hope you find it useful.

All the Best,

Chris Balbach, PE, CEM, CMVP, BEMP, BEAP, BESA

Chris Balbach's picture
Joined: 2011-09-30
Reputation: 1