[Bldg-sim] loop to loop heatpump in equest

3 posts / 0 new
Last post

I've chatted with Rashmi and feel like others may want to know how this
can be done in the future... Also copying [Equest-users] on this
discussion as advised by a list administrator.

See attached screengrab for a simple visual of how I accomplished
this...

Setting up the 4-pipe system sourced from a closed ground loop:

1. Create a 'chiller' of the type 'water-to-water heat pump'
2. Create an independent circulation loop of the type 'lake/well'
3. Create a ground loop heat exchanger of the type 'lake/well'
4. Select as the chiller CW loop and GLHX circulation loop the
independent loop you created - they should connect without errors
5. Assign a custom temperature schedule for the lake/well ground
loop in the screen that pops up when you double click the GLHX (read on)
6. Make sure the default head/pump properties for the circulation
loop and GLHX make sense for your system - adjust as necessary. (i.e.
if it's a closed ground loop, there shouldn't be static head)

The tricky thing is that both Rashmi and I were dealing with closed-loop
ground heat exchanger (not lake/well). The solution I came to model
this was to temporarily save the project in a separate directory to keep
my work, then go back to the wizards and heat/cool the same building
reasonably using the DX/DX heating/cooling ground source heatpump
option, with all appropriate ground and field properties well defined.

Once this is done, you can have eQuest additionally generate an hourly
report over of the resulting ground loop temperatures throughout the
year (actual procedure eludes my memory, but it's in the archives).
Once this is done, I used that raw data to come up with the resulting
weekly average temperatures in the ground loop using excel. I then used
these figures to create a custom temperature schedule for the actual
project, which was assigned to the GLHX as mentioned above. This way
you're reasonably accurately representing what temperatures the
water-to-water equipment is seeing coming out of the ground throughout
the year based on your actual geographic, ground loop, and soil
condition variables as defined.

The final logical step, and one I'm currently working on learning for
myself, is defining a custom performance curve for the water-to-water
equipment that's representative of the real thing. I have been advised
the default chiller curves as well as the water-to-air heatpump curves
from the Climatemaster eQuest add-on are not appropriate for modeling
water-to-water heatpumps.

Best of luck!

PS: Apologies for the screenshot's quality - I had issues making it
small enough for the mailing list to accept.

NICK CATON, E.I.T.

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

Nick/Rashmi-

One limitation of the Well-Water-to-Water Heat-Pump (or W2WHP, TYPE =
HEAT-PUMP) is that only a chilled water or 2-pipe circulation loop can be
attached to it. This can be an issue if the central heat pump is designed
to supply both hot and chilled water.

produce both hot and cold water simultaneously, I gather that analysts often
opt to use loop-to-loop heat pump (or L2LHP, TYPE = LOOP-TO-LOOP-HP)
instead, since you can connect both a chilled water and hot water loop to
the chiller. However, from experience, I know that many of the
water-to-water heat pumps available are NOT loop-to-loop heat pumps, but
rather a series of water-to-water heat pumps that are manifolded to allow
some circuits to operate in heating while others are operating in cooling,
with all circuits sharing a common condenser water loop. If you read the
description of the L2LHP chiller closely, the L2LHP operates much
differently than the W2WHP machine. Also, I would be wary of using the
default DOE-2 performance curves for loop-to-loop and well-water-to-water
chillers, as I found that the default performance curves were quite
different from what is currently available in the US market.

The approach you have outlined, I believe, is on the right track, if the
correct chiller type is being simulated. However, it seems unlikely that
estimating the ground loop loads/temperatures using a building model with
single-zone WSHPs will produce the same results as the W2WHP/chilled beam
system you described.

Here are some reasons:

1) Does not account for pump heat (both positive and negative impact
depending on it the system is in heating or cooling) on the building
circulation (chilled beam) loops.

2) Using the "DX/DX heating/cooling ground source heat pump" option you
mentioned, you will basically get single-zone WSHP systems, where the WLHP
loop is connected to the eQUEST GSHX (well-field) model. This method may be
taking too much credit for periods when there is high diversity in the
building. If a small fraction of the building requires cooling, the chilled
water system would have to run and the W2WHP would impart a load on the
ground loop and GSHX. However, in a building model with single zone
water-to-air heat pumps, I believe the load on the loop could, in theory, be
balanced by a fraction of units in heating and cooling, so that the ground
temperature leaving the building is the same as it was coming in.

3) Finally, part-load performance curves for water-to-air heat pumps
are likely not appropriate for a water-to-water heat pump system.

Here is a suggested alternative (that I have fully implemented in eQUEST):

1) Take the hourly LOADS on the lake/well loop model you have created.

2) Apply them, with the appropriate schedules, as a internal energy
source in a single space building.

3) This space is served by a single zone, water-to-air heat pump, which
has a WLHP condenser loop and a GSHX, as would be typically be created by a
the wizard.

4) Make all the surfaces of the space adiabatic, zero out all other
loads and the system fan energy, and set you cooling/heating EIRs to 1.0.
With a few other adjustments (not enough time to list them all), you can
more or less get the load on the WSHP loop to be the same as internal energy
source load applied to the space.

In essence, you are taking the loads directly from the lake-well loop and
transferring them to the other model what includes the GSHX model of the
well-field. From this secondary model, you can then extract the
temperatures returning from the well-field. The rest is similar to you
described, except you could even get as detailed as 8760 hourly load and
temperature schedules in each model.

Any thoughts from others who have simulated/studied these systems before?

David Reddy

David Reddy2's picture
Offline
Joined: 2011-09-30
Reputation: 0

Nick/Rashmi-

One limitation of the Well-Water-to-Water Heat-Pump (or W2WHP, TYPE =
HEAT-PUMP) is that only a chilled water or 2-pipe circulation loop can be
attached to it. This can be an issue if the central heat pump is designed
to supply both hot and chilled water.

produce both hot and cold water simultaneously, I gather that analysts often
opt to use loop-to-loop heat pump (or L2LHP, TYPE = LOOP-TO-LOOP-HP)
instead, since you can connect both a chilled water and hot water loop to
the chiller. However, from experience, I know that many of the
water-to-water heat pumps available are NOT loop-to-loop heat pumps, but
rather a series of water-to-water heat pumps that are manifolded to allow
some circuits to operate in heating while others are operating in cooling,
with all circuits sharing a common condenser water loop. If you read the
description of the L2LHP chiller closely, the L2LHP operates much
differently than the W2WHP machine. Also, I would be wary of using the
default DOE-2 performance curves for loop-to-loop and well-water-to-water
chillers, as I found that the default performance curves were quite
different from what is currently available in the US market.

The approach you have outlined, I believe, is on the right track, if the
correct chiller type is being simulated. However, it seems unlikely that
estimating the ground loop loads/temperatures using a building model with
single-zone WSHPs will produce the same results as the W2WHP/chilled beam
system you described.

Here are some reasons:

1) Does not account for pump heat (both positive and negative impact
depending on it the system is in heating or cooling) on the building
circulation (chilled beam) loops.

2) Using the "DX/DX heating/cooling ground source heat pump" option you
mentioned, you will basically get single-zone WSHP systems, where the WLHP
loop is connected to the eQUEST GSHX (well-field) model. This method may be
taking too much credit for periods when there is high diversity in the
building. If a small fraction of the building requires cooling, the chilled
water system would have to run and the W2WHP would impart a load on the
ground loop and GSHX. However, in a building model with single zone
water-to-air heat pumps, I believe the load on the loop could, in theory, be
balanced by a fraction of units in heating and cooling, so that the ground
temperature leaving the building is the same as it was coming in.

3) Finally, part-load performance curves for water-to-air heat pumps
are likely not appropriate for a water-to-water heat pump system.

Here is a suggested alternative (that I have fully implemented in eQUEST):

1) Take the hourly LOADS on the lake/well loop model you have created.

2) Apply them, with the appropriate schedules, as a internal energy
source in a single space building.

3) This space is served by a single zone, water-to-air heat pump, which
has a WLHP condenser loop and a GSHX, as would be typically be created by a
the wizard.

4) Make all the surfaces of the space adiabatic, zero out all other
loads and the system fan energy, and set you cooling/heating EIRs to 1.0.
With a few other adjustments (not enough time to list them all), you can
more or less get the load on the WSHP loop to be the same as internal energy
source load applied to the space.

In essence, you are taking the loads directly from the lake-well loop and
transferring them to the other model what includes the GSHX model of the
well-field. From this secondary model, you can then extract the
temperatures returning from the well-field. The rest is similar to you
described, except you could even get as detailed as 8760 hourly load and
temperature schedules in each model.

Any thoughts from others who have simulated/studied these systems before?

David Reddy

David Reddy2's picture
Offline
Joined: 2011-09-30
Reputation: 0