BEST PRACTICES: Wild Coils

27 posts / 0 new
Last post

Apologies for the cross-post, however I wanted to ask this question from 2 angles and I feel both communities may benefit from the discussion (if I can spark one).

A common reality I've observed with "real-world" hydronic systems is that system coils and baseboard/radiator loops fall into a state coined wild coils. Rather than modulating flow to maintain a measured supply air or room temperature setpoint, flow is uncontrolled. A heating or reheat coil for example will end up dumping heat at all times the associated circulation loop is active, independent of its associated system's fan operation, cooling coil activity, or thermostat signals requesting more/less heating. Occupants in response to wild coils, when they cay, will end up using windows, propping open doorways, plugging in local space heaters / circ fans, and generally suffering in terms of comfort. In just about every case, this scenario presents a win-win in terms of improved occupant comfort potential in parallel with energy savings potential for whoever is paying the bills.

Causes for this situation I've encountered more than once include:

* Manual Control valves left in an open state, with dusty cobwebs suggesting their presence is unknown to the occupants/building operators
* Automated valves (electric or pneumatic) which have become mechanically stuck in an open, or partially open position
* Automated valves (electric or pneumatic) which are otherwise busted due to upstream pneumatic line/system issues or mechanical failures of the moving parts at the valve
* A valve was never designed and/or installed and/or wired up for control in the first place

For all of this however, I have always struggled in approximating the energy and comfort impacts of "wild" coils in my building energy simulations. Quantifying this impact with some degree of confidence is difficult, but desirable in cases where I am calibrating to existing utility bills (read: always) and/or asserting the utility savings and comfort improvement impact for fixing/addressing such situations.

For the [bldg-sim] family: Are there any 3rd party tools, models, or other energy simulation platforms with explicit options for evaluating the comfort and energy impacts of wild coil situations? Is there any research I could be pointed towards exploring this topic?

For the [eQuest-users] crowd: Can anyone share a best practice or recommendation for simulating this sort of problem-state within a doe2/eQuest model? As far as I know, the native input options are essentially limited to a pair of "working" coil modulation states: TWO-WAY and THREE-WAY. Here's an example doe2 reference entry, with language that repeats a couple times over for different scenarios:
[cid:image003.png at 01D5281B.FBA76FA0]
I personally have taken different approaches, with none being particularly satisfactory. These have included introducing process loads onto the loops concurrently with "free" internal energy source definitions to get those losses dumped into the spaces experiencing discomfort. I have also played with artificially bumping the thermostat schedules around to reflect measured, uncomfortable temperature states...

Any solutions/experiences/shared-commiseration would be very welcome!

~Nick

[cid:image005.png at 01D515A3.47EDD880]
Nick Caton, P.E., BEMP
Senior Energy Engineer
Regional Energy Engineering Manager
Energy and Sustainability Services
Energy Performance Contracting
D
M
F
E
913 . 564 . 6361
785 . 410 . 3317
913 . 564 . 6380
nicholas.caton at se.com
15200 Santa Fe Trail Drive
Suite 204
Lenexa, KS 66219
United States
[cid:image006.png at 01D515A3.47EDD880]

Nicholas Caton2's picture
Offline
Joined: 2019-03-25
Reputation: 0

I'm certain I could get IESVE to model a wild coil like that plus it has ASHRAE 55 comfort calcs built in to evaluate the results.

I'm currently modeling baseboard heaters that have no thermostat or control. They heat anytime the loop has hot water and often over heat the space, which causes simultaneous heating and cooling with the PTAC.

Cory Duggin, PE, LEED AP BD+C, BEMP
Principal | PEAK Institute
cory.duggin at tlc-eng.com
TLC ENGINEERING SOLUTIONS
12 Cadillac Dr., Ste 150
Brentwood, TN 37027
Main: 615.297.4554
Office: 615.346.1939
Cell: 931.273.8396

www.tlc-engineers.com
USGBC MidTN Branch Leadership Group Co-chair

[cid:TLCtriangle_48_9b0b2127-1fd8-4dd1-8210-7b9e72ba4dda.png] [cid:LinkedIn_d77bd78f-26ef-4f20-a83e-d6bea9af9afa.png] [cid:Facebook_8e703d49-6367-43ed-b9af-b0d3399f9ce5.png] [cid:Twitter_009986bf-8b3b-4a60-972d-6224dff1b1bb.png]

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

Hey Nick,

TL;DR I usually play on the infiltration rate, bumping it higher as needed.

Something I've encountered a lot while working on Multifamily existing buildings in NYC, where you'll encounter most often than not steam systems, one or two pipe. The "wild" coil situation is exacerbated to unreal levels, because these systems are very old, out of tune, hard to balance, and the ancient knowledge of steam systems is getting lost.
Part of our audit checklist in the winter involved walking around the building to map the windows (usually no building plans exists...) and flag the ones that? were opened (you could get in the 30% easily even a deadly cold winter day. I've personally lived on a run down building when i first got to nyc and on a well below freezing point our windows were opened to avoid sweating to death: the good old "double-hung zone valve" was the only option we had).
We'd take space temperature readings (and ideally install sensors for a couple of weeks), to set the thermostat setpoints accordingly, and play with infiltration to match our understanding of how people reacted.

The opposite is true too though, I've seen people run their gas stoves with the door open as a supplemental space heater because it was too damn cold. The cooking gas account is more often than not separated from the boiler room one in nyc thankfully, so if you find a huge spike in the cooking gas bill in the winter (a HDD correlated component of your bill when you run a regression) it probably can't be explained by "people don't go out as much and stay in to cook".

Just my two cents/a rant, but i hope this sparks a conversation anyway :)

Y'all enjoy the weekend!

Best,
Julien
?--
Sent from a mobile device, please excuse the brevity.

Julien Marrec, EBCP, BPI MFBA
Owner

Direct:?+33 6 95 14 42 13
Website:?www.effibem.com
LinkedIn (en) | (fr)?

jmarrec's picture
Offline
Joined: 2013-01-09
Reputation: 0

Nick,

Small world, I seem to run into this a lot too.

As for the eQuest side, I'll throw in a futile effort that I tried. It's
been a few years since I've tried, so it may be worth another shot with
2.3, and also I may have been doing something wrong. Under the coil
options, there is a CHW-Coil-Cap-fFluidFlow. The default looks something
like this.

[image: image.png]

DOE2 uses this curve to control the cooling valve to modulate its hourly
output to meet a setpoint. In other words, if I need X Btu/hr of chilled
water to get Y Deg-F air off the coil, how much water flow is that? My
thought was to trick DOE2 with a custom curve which has almost zero
capacity except near 100% flow. This would make the controller run nearly
100% flow through the coil regardless of load. This curve would be like
the blue one below.

[image: image.png]

You are limited to cubic, quadratic, or linear profiles with this curve, so
creating this shape can be a little tricky. This should force the valve
into nearly 100% flow all the time, but it won't model any space
temperature effects.

To try to model space temperature effects of a stuck valve that isn't
modulating, I've tried to create a curve with near constant capacity like
the green one below. It needs to be slightly linear, or else you will put
the program into a death spiral.

[image: image.png]

In theory, this coil should put out near constant cooling output regardless
of the "flow" that DOE2 puts through it. Again, last time I went down this
rabbit hole, I was unsuccessful, but it may be worth another shot.

This doesn't quite meet your explicit option, but one 3rd party option is
to use the LBL buildings library with something like open modelica. The
library has a built-in comfort model (Fanger model), and you could build a
hydronic system around it. Downside is the overhead in learning modelica,
and you will likely have to create a single-zone pet model to study the
problem in general terms.

Definitely an issue with current state of building modeling tools is the
slant towards new construction and fully functioning equipment. I
shouldn't say issue because maybe this meets 90% of modeling needs? But
those of us who live in the retrofit space run into these things all the
time. What about a pneumatic thermostat that is leaking air and so is
causing a cooling valve to sit constantly at 80% open? Or VFDs that are
put in hand at 40 HZ, or control valves that are manually overridden?

Aaron

On Fri, Jun 21, 2019 at 1:37 PM Nicholas Caton via Equest-users <

Aaron Powers2's picture
Offline
Joined: 2011-09-30
Reputation: 0

..and since you people are distracting me enough to get me thinking about a problem that I have rarely seen...

In the EnergyPlus world:

1. As Julien suggests, we could "play" with infiltration. the ZoneInfiltration:EffectiveLeakageArea object can be scheduled, perhaps, with higher rates in colder weather for opened windows
2. I think we could also create zones in such a way as to represent our understanding of the wild coil areas and assign a fictitious zone temperature setpoint of, say, 90F. At the same time, we could schedule the boiler's operation for cold weather only (which is likely to be the case anyway or people would be screaming at the building operator in the warmer months)
3. Alternatively, some zones could have their heating setpoint overridden by an EMS routine based on certain weather conditions. We could do the same for infiltration.
4. Another possibility that comes to mind is to "schedule" boiler efficiency via EMS to be lower during colder weather. This would be a "catchall" value that would be used to calibrate the model (with brute force) to utility bill actual values.

ps, This sort of issue demonstrates the power of the modeling community to share experience and creativity - so that everyone becomes smarter and better able to provide real solutions! Let's wring it out.

pps, Cory, can you share more details for the IES users?

ppps, Funny to hear someone who lives in France write "y'all" ??. Your time in NYC changed you! (But you must have traveled South to hear that term very often)

James V Dirkes II, PE, BEMP, BCxP
Team Lead - Building Performance
616 450 8653
Foresight Management
https://fsmgmt.co/

Coffee Conversation:

But to handle human lives with no agreement as to what human beings are or what the purpose of life is - that is a formula for chaos.

Jim Dirkes's picture
Offline
Joined: 2019-05-22
Reputation: 0

Hi Nick

Ah I like "wild coil". Makes this industry feel much more exciting.

We usually call it "simultaneous heating and cooling". And sometimes it is not so much wild as a poor controls design. eg For a given zone the VAV box is on one thermostat, and baseboard radiation on another.

In eQuest, from my experience, it really depends on the situation but one thing I have done is:

* With the "Cool Control" = Constant, set VAV box minimum flow high so that reheat is required to meet set point. Adjusting DAT and VAV min flow (within reason so as not to throw fan energy out of wack) until the results are reconciled with utility billing. (Noting that I back into it starting from proper setpoints )

On the fly I can't see how messing with valves (in eQuest) would help.

I usually save infiltration adjustments as suggested by Julien to model poor fan tracking (excess exhaust ,poor building pressurization)

Best

Melissa Crowe, LEED AP, BEMP
mcrowe at cmta.com
508-647-9200 x 225

Melissa
Melissa Page Crowe's picture
Joined: 2011-09-30
Reputation: 1

You could set up a model with the Modelica Buildings Library, which models
valves and dampers explicitly, and computes the flow distribution based on
the pressure distribution in the piping (or duct) network. You can then
overwrite the control signal to keep the valve always at a certain position.

All the best,

Michael

On Fri, Jun 21, 2019 at 2:06 PM Melissa Crowe via Bldg-sim <

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

GREAT ideas so far (keep em? coming)! I?m going to forward a couple Reponses that didn?t get CC?ed to both lists in short order.

[cid:image005.png at 01D515A3.47EDD880]
Nick Caton, P.E., BEMP
Senior Energy Engineer
Regional Energy Engineering Manager
Energy and Sustainability Services
Energy Performance Contracting
D
M
F
E
913 . 564 . 6361
785 . 410 . 3317
913 . 564 . 6380
nicholas.caton at se.com
15200 Santa Fe Trail Drive
Suite 204
Lenexa, KS 66219
United States
[cid:image006.png at 01D515A3.47EDD880]

Nicholas Caton2's picture
Offline
Joined: 2019-03-25
Reputation: 0

In the VE you could do it a couple different ways. Assuming the valve is stuck open, I would modify the heating coil control to control heat transfer, so it would always be providing the amount of heat the coil is designed to at that flow and water temp as long as HHW is available. Make sure to use their advanced coil model though so the available coil capacity will vary depending on the HHW temperature. If it is a changeover system, you would need to create an availability profile to put on the ?wild? coil to limit when it can go wild.

Cory Duggin, PE, LEED AP BD+C, BEMP
Principal | PEAK Institute
cory.duggin at tlc-eng.com
TLC ENGINEERING SOLUTIONS
12 Cadillac Dr., Ste 150
Brentwood, TN 37027
Main: 615.297.4554
Office: 615.346.1939
Cell: 931.273.8396

www.tlc-engineers.com
USGBC MidTN Branch Leadership Group Co-chair

[cid:TLCtriangle_48_9b0b2127-1fd8-4dd1-8210-7b9e72ba4dda.png] [cid:LinkedIn_d77bd78f-26ef-4f20-a83e-d6bea9af9afa.png] [cid:Facebook_8e703d49-6367-43ed-b9af-b0d3399f9ce5.png] [cid:Twitter_009986bf-8b3b-4a60-972d-6224dff1b1bb.png]

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

Hey Cory!

I?m thinking of the case for a fan-powered box with a reheat coil (and a stuck open valve):

What you?re describing (to my doe2-tuned mind & terminology) sounds in effect a lot like a scheduled internal process/heat load, pushing a specified amount of heat into the space for all hours the loop would be active. A shortcoming (at least in doe2/eQuest) for this approach stems from the following:

You have asserted (or maybe just inferred?) that in VE, the specified heat transfer can be tied to the activity of the associated hydronic loop. How might you supplement this approach to also capture the relatively different (perhaps not binary) amount of heat dumped into a space between states when the local fan is in operation vs. not? Is there a way to tie these operations together within the simulation that doesn?t require processing of interval/hourly data outside of the model and feeding that back in via customized scheduling?

For shared context, the issue of a ?changeover? situation such as with 2-pipe systems and/or seasonal scheduling is easily enough addressed on the doe2/eQuest side of things by specifying hydronic loop availability and/or operational states over on the hydronic tab (of the eQ interface). Happy to help anyone struggling with that over in the [equest-users] community ;-).

@ Everyone: I do, by the way, very very much appreciate this flood of contributions so late on a Friday with the weekend beckoning us to go out and enjoy life! It?s very encouraging to find validation in our shared worries and solution-paths through other?s experiences and approaches, and I?m certain I?m not the only one who feels that gratitude.

~Nick

[cid:image005.png at 01D515A3.47EDD880]
Nick Caton, P.E., BEMP
Senior Energy Engineer
Regional Energy Engineering Manager
Energy and Sustainability Services
Energy Performance Contracting
D
M
F
E
913 . 564 . 6361
785 . 410 . 3317
913 . 564 . 6380
nicholas.caton at se.com
15200 Santa Fe Trail Drive
Suite 204
Lenexa, KS 66219
United States
[cid:image006.png at 01D515A3.47EDD880]

Nicholas Caton2's picture
Offline
Joined: 2019-03-25
Reputation: 0

EnergyPlus has the notion of fault models to model specific types of faults. Perhaps this could be added to the list of faults to model?

https://bigladdersoftware.com/epx/docs/9-0/input-output-reference/group-operational-faults.html

It might also be possible to model using EMS controls?

Dan

Sent from my iPad

Macumber, Daniel's picture
Joined: 2011-10-02
Reputation: 0

It?s easy to add a scheduled load in E+, too. Nice idea.

James V Dirkes II, PE, BEMP, BCxP
Team Lead - Building Performance
616 450 8653
Foresight Management
https://fsmgmt.co/

Coffee Conversation:
But to handle human lives with no agreement as to what human beings are or what the purpose of life is - that is a formula for chaos.

Jim Dirkes's picture
Offline
Joined: 2019-05-22
Reputation: 0

And the scheduled load can be any fuel you like, including a fuel that doesn?t get measured by the electric or gas meter.

James V Dirkes II, PE, BEMP, BCxP
Team Lead - Building Performance
616 450 8653
Foresight Management
https://fsmgmt.co/

Coffee Conversation:
But to handle human lives with no agreement as to what human beings are or what the purpose of life is - that is a formula for chaos.

Jim Dirkes's picture
Offline
Joined: 2019-05-22
Reputation: 0

Nick and all,

Perhaps I can answer your question regarding what Cory has described for the IES VE, and perhaps take that a little further, as I imagine Cory might have gone home for the day by this time. J

Firstly, indeed, if a runaway source of heat can be sufficiently described as a scheduled internal gain, then most all of the simulation tools discussed here ought to be able to do that quite readily. If, however, the behavior or output at any given time for the ?wild? coil, radiator, baseboard, unit, etc. is tied to the characteristics of the of which it is a part, it may be preferable to model this as a system or terminal unit that includes a component that has just that behavior?

For example, depending upon the type of ?rogue?, ?wild?, or ?runaway? device, there are likely to be contextual influences at one or more of five different levels:

1) Room?e.g., air and/or radiant temperature of the space, and thus the delta-T seen by the rogue device, which will influence the amount of unintentional heating or cooling it does (if it is a ?room unit? that is directly exposed to the room conditions) ;

2) Component?e.g., design vs. off-design output at reduced or increased delta-T with respect to the space or airflow path that it is located within;

3) Terminal unit?e.g., interactions with FCU fan, OA path, etc. operation, volume flow rate, and entering air temperatures;

4) System?e.g., primary airflow operation, volume, and air temperature;

5) Plant: HWL or CHWL operation, supply temperature, and flow rate.

The output of a rogue Radiator with a stuck valve would, for example, be influenced to one or another degree by 1, 2, and 5 above.
The output for an FCU coil with a stuck valve, for example, be influenced to one or another degree by 2, 3, 4, and 5 above.

In the IES Virtual Environment ??specifically within the ApacheHVAC module of the VE?? the following are relevant specifics regarding control and modeling of heating and cooling coils and ?room units? (e.g., hydronic radiators, convective baseboard units, radiant heating panels, or embedded hydronic floor/ceiling slab heating loops, as well as the corresponding hydronic cooling panels and loops) if you are interested in modeling one or another form of ?runaway? or ?wild? source of heating/cooling in the context of the system or zone-level terminal unit that it is attached to. These attributes would allow you to account for any combination of contextual influences 1??5 listed above:

? Coils: In addition to the usual controlled variables for a coil (leaving air DBT and, for cooling coils, leaving air DPT), a coil can be controlled to transfer a given amount of heat to the airstream passing over the coil (or from the airstream for the case of a cooling coil). This is different than simply modeling a scheduled process load or internal gain in the space, as the heating (or cooling) done by the coil will reach the space only to the extent that the air passing over the coil goes to that space.

o If the airflow is modulated (e.g., as in a fan-coil with multi-speed or variable-speed fan, or for a re-heat coil in a fan-powered or basic VAV box) a given controlled heat transfer rate will determine how much heat is dumped into that airstream, and thus the temperature of the air as the combined outcome of heat transfer and volume airflow. If the airflow modulates to a very small flowrate and the heat transfer remains fixed, the result may be a small quantity of very warm (or very cool) air.

o If the advanced coil model is used, the amount of heat transferred to the air will be also constrained by the delta-T between the coil and the air, using a counter-flow model of the water in the coil relative to the airflow (i.e., for heating, the entering air will initially encounter the coolest water that is exiting the coil, and will lastly be heated by the hottest water entering the coil, and vice versa for a cooling coil). Thus the heat transfer will be limited to that which can be physically transferred between the water loop and air loop, given the temperatures and flow rates on both sides of this counter-flow arrangement, as well as the coil description in terms of capacity, contact factor, etc.

o The ?simple coil?, which is really intended mainly for modeling thing other than coils on water loops, has fewer constraints, as it does not attempt to account for all aspects of air and water flows and temperatures. There are some feasibility constraints in VE2018, and there is a move to improve upon this in the new future.

o Regardless of the coil model used, if the airflow over the coil is shut off completely, the heat transfer from (or to) the coil will cease.

? In the case of a Fan-Coil Unit, this will depend upon the configuration modeled: If the ventilation air to the space is modeled as passing through the FCU, the airflow over the coil will persist so long as the system (e.g., DOAS for ventilation) is operating; If the ventilation air path to the space is modeled as separate from the FCU, the airflow over the coil will be present only when the FCU fan is operating.

? VRF Indoor Unit coils are treated the same as FCU coils in this regard.

? Parallel fan-powered boxes can likewise be modeled with the heating coil either within the combined primary + secondary airstream or just on the secondary airflow path.

? Active Chilled/Heated Beams are similar: The coil can be modeled as seeing only the induced airflow (typical active beam config) or both the primary + induce airflow. While there is no local fan, and thus airflow is determined at the system level, the airflow path is relevant as this determines the total volume of air seen by the coil and thus the amount of heat transfer that is possible (given temperature and flow rates for both the water and air loops).

o Heat transfer to/from the air flowing over the coil will also be constrained or controlled by (all of which are optional):

? Schedules (consistent with the scheduled internal gain that Nick suggested). This could be set to ?on continuously? for a coil that is making or removing heat all the time, could use the HVAC system operating schedule if it does this whenever the system is operating, or could use the HVAC schedule with a seasonal availability constraint for cases wherein the coil is on a seasonal change-over system. The schedule profile can also incorporate a simple formula referencing variables such as the outdoor air temperature, etc. if this is the basis for the change-over operation.

? On/off control and hysteresis (deadband) with respect to a setpoint and sensed variables;

? On/off control as a function of logical AND/OR connections to other controllers.

? Proportional control of the heat transfer amount with respect to a setpoint, control bandwidth, and sensed variables. Both the setpoint and the control signal values at min & max sensed variable can also be varied according to a profile.

? Room unit or zone/room-level hydronic loop: These include hydronic radiators, convective baseboard units, radiant heating panels, or embedded hydronic floor/ceiling slab heating loops, as well as the corresponding hydronic cooling panels and loops

o Heating/cooling output for all of these components is constrained by the following:

? SWT and flow rate of the water loop (HWL or CHWL) to which the zone-level component is attached;

? Design max water flow rate for the component;

? Design capacity at design conditions (design room air temperature and design room radiant temperature), and thus reduced or increased capacity at off-design conditions as a function of lesser or greater delta-T.

o Controllers for these types of heating and cooling components include all of the following (optional) controls:

? Schedules, as described above for coil controls??e.g., could be scheduled as always on with HVAC system.

? On-Off setpoint control, as described above for coil controls.

? On-Off as a function of logical AND/OR connections with other controllers, as described above for coil controls.

? Proportional flow control, as described above for coil controls, but specifically for water flow control.

? Proportional temperature control, as described above for coil controls, but specifically for water temperature control.

o Surface temperature in the space: Because room units are, unlike coils, located within the room, they are exposed to room conditions?including mean radiant temperature for the space?that influence their output. Furthermore, all of the above On/Off and Proportional controls have access to one additional sensed variable:

? Floor, ceiling, or wall temperature for any one tagged surface within the room. And, as there is radiant exchange between surfaces, convective heat transfer with the air, and conductive heat transfer with adjacent spaces or exterior environments, this surface temperature is an integral part of the thermo-physical model of the room.

? This sensed surface temperature is normally used for the control of an embedded hydronic loop; however, in the case of a ?rogue? radiator or similar device that might be mounted directly on an exterior wall, this sensed variable could be used to ?control? the ?rogue? radiator output in relation to the interior surface temperature of the exterior wall. Thus, especially for an older building with poorly insulated walls, the model could be configured to estimate the output and resulting comfort and energy implications of a radiator adjacent to this externally cooled surface?I.e., increasing the heat output of the radiator as a result of the potentially high delta-T in that location.

o Zonal subdivisions: If you want a better estimate of how much heat from a rogue radiator is being lost through the exterior wall it?s mounted upon, it may be useful to divide the room into core and perimeter zones, with a rather thin perimeter zone being separated only by an ?air barrier? (simply a division between ?fully mixed? thermal zones in the model, but having no partition of any kind preventing either radiative or conductive heat transfer. And, if you wanted to take this further, you could firstly turn on MacroFlo to account for air transfer between the ?fully mixed? perimeter and core thermal zones, as driven by bulk-airflow modeling of airflow driven both by temperature differences and other airflow inputs and outputs (infiltration, exfiltration, operable windows, window crack flow coefficients, airflow from adjacent rooms, and any HVAC airflow inlets or extract).

? Direct-acting Heater/Coolers: Lastly, and perhaps truly least of all, there is also within ApacheHVAC a component called a Direct-acting Heater/Cooler that can be added to any space. This simpler ?room unit? component can be used for both heating and cooling. It is like a radiator or other room unit in terms of the controls and sensed variables, but there is no relationship to any hydronic loops and there are no detailed inputs to describe design vs. off-design output. It just does what you tell it to do, as constrained only by the controller inputs.

That was kind of fun to write! Hope you enjoyed it.

Cheers,
Timothy

[IES-email2014]

Timothy Moore
Senior Product Manager

C:

+1 (415) 810-2495

T:

+1 (404) 806-2018

T:

+1 (617) 502-2085

Internal IP ext. 8549

http://www.iesve.com

Integrated Environmental Solutions Limited. Registered in Scotland No. SC151456
Registered Office - Helix Building, West Of Scotland Science Park, Glasgow G20 0SP

Email Disclaimer

Timothy Moore2's picture
Offline
Joined: 2011-10-02
Reputation: 200

Hey Nick,

TL;DR I usually play on the infiltration rate, bumping it higher as needed.

Something I've encountered a lot while working on Multifamily existing buildings in NYC, where you'll encounter most often than not steam systems, one or two pipe. The "wild" coil situation is exacerbated to unreal levels, because these systems are very old, out of tune, hard to balance, and the ancient knowledge of steam systems is getting lost.
Part of our audit checklist in the winter involved walking around the building to map the windows (usually no building plans exists...) and flag the ones that? were opened (you could get in the 30% easily even a deadly cold winter day. I've personally lived on a run down building when i first got to nyc and on a well below freezing point our windows were opened to avoid sweating to death: the good old "double-hung zone valve" was the only option we had).
We'd take space temperature readings (and ideally install sensors for a couple of weeks), to set the thermostat setpoints accordingly, and play with infiltration to match our understanding of how people reacted.

The opposite is true too though, I've seen people run their gas stoves with the door open as a supplemental space heater because it was too damn cold. The cooking gas account is more often than not separated from the boiler room one in nyc thankfully, so if you find a huge spike in the cooking gas bill in the winter (a HDD correlated component of your bill when you run a regression) it probably can't be explained by "people don't go out as much and stay in to cook".

Just my two cents/a rant, but i hope this sparks a conversation anyway :)

Y'all enjoy the weekend!

Best,
Julien
?--
Sent from a mobile device, please excuse the brevity.

Julien Marrec, EBCP, BPI MFBA
Owner

Direct:?+33 6 95 14 42 13
Website:?www.effibem.com
LinkedIn (en) | (fr)?

jmarrec's picture
Offline
Joined: 2013-01-09
Reputation: 0

Nick,

Small world, I seem to run into this a lot too.

As for the eQuest side, I'll throw in a futile effort that I tried. It's
been a few years since I've tried, so it may be worth another shot with
2.3, and also I may have been doing something wrong. Under the coil
options, there is a CHW-Coil-Cap-fFluidFlow. The default looks something
like this.

[image: image.png]

DOE2 uses this curve to control the cooling valve to modulate its hourly
output to meet a setpoint. In other words, if I need X Btu/hr of chilled
water to get Y Deg-F air off the coil, how much water flow is that? My
thought was to trick DOE2 with a custom curve which has almost zero
capacity except near 100% flow. This would make the controller run nearly
100% flow through the coil regardless of load. This curve would be like
the blue one below.

[image: image.png]

You are limited to cubic, quadratic, or linear profiles with this curve, so
creating this shape can be a little tricky. This should force the valve
into nearly 100% flow all the time, but it won't model any space
temperature effects.

To try to model space temperature effects of a stuck valve that isn't
modulating, I've tried to create a curve with near constant capacity like
the green one below. It needs to be slightly linear, or else you will put
the program into a death spiral.

[image: image.png]

In theory, this coil should put out near constant cooling output regardless
of the "flow" that DOE2 puts through it. Again, last time I went down this
rabbit hole, I was unsuccessful, but it may be worth another shot.

This doesn't quite meet your explicit option, but one 3rd party option is
to use the LBL buildings library with something like open modelica. The
library has a built-in comfort model (Fanger model), and you could build a
hydronic system around it. Downside is the overhead in learning modelica,
and you will likely have to create a single-zone pet model to study the
problem in general terms.

Definitely an issue with current state of building modeling tools is the
slant towards new construction and fully functioning equipment. I
shouldn't say issue because maybe this meets 90% of modeling needs? But
those of us who live in the retrofit space run into these things all the
time. What about a pneumatic thermostat that is leaking air and so is
causing a cooling valve to sit constantly at 80% open? Or VFDs that are
put in hand at 40 HZ, or control valves that are manually overridden?

Aaron

On Fri, Jun 21, 2019 at 1:37 PM Nicholas Caton via Equest-users <

Aaron Powers2's picture
Offline
Joined: 2011-09-30
Reputation: 0

Hi Nick

Ah I like "wild coil". Makes this industry feel much more exciting.

We usually call it "simultaneous heating and cooling". And sometimes it is not so much wild as a poor controls design. eg For a given zone the VAV box is on one thermostat, and baseboard radiation on another.

In eQuest, from my experience, it really depends on the situation but one thing I have done is:

* With the "Cool Control" = Constant, set VAV box minimum flow high so that reheat is required to meet set point. Adjusting DAT and VAV min flow (within reason so as not to throw fan energy out of wack) until the results are reconciled with utility billing. (Noting that I back into it starting from proper setpoints )

On the fly I can't see how messing with valves (in eQuest) would help.

I usually save infiltration adjustments as suggested by Julien to model poor fan tracking (excess exhaust ,poor building pressurization)

Best

Melissa Crowe, LEED AP, BEMP
mcrowe at cmta.com
508-647-9200 x 225

Melissa
Melissa Page Crowe's picture
Joined: 2011-09-30
Reputation: 1

GREAT ideas so far (keep em? coming)! I?m going to forward a couple Reponses that didn?t get CC?ed to both lists in short order.

[cid:image005.png at 01D515A3.47EDD880]
Nick Caton, P.E., BEMP
Senior Energy Engineer
Regional Energy Engineering Manager
Energy and Sustainability Services
Energy Performance Contracting
D
M
F
E
913 . 564 . 6361
785 . 410 . 3317
913 . 564 . 6380
nicholas.caton at se.com
15200 Santa Fe Trail Drive
Suite 204
Lenexa, KS 66219
United States
[cid:image006.png at 01D515A3.47EDD880]

Nicholas Caton2's picture
Offline
Joined: 2019-03-25
Reputation: 0
Nicholas Caton2's picture
Offline
Joined: 2019-03-25
Reputation: 0
Nicholas Caton2's picture
Offline
Joined: 2019-03-25
Reputation: 0

Hey Cory!

I?m thinking of the case for a fan-powered box with a reheat coil (and a stuck open valve):

What you?re describing (to my doe2-tuned mind & terminology) sounds in effect a lot like a scheduled internal process/heat load, pushing a specified amount of heat into the space for all hours the loop would be active. A shortcoming (at least in doe2/eQuest) for this approach stems from the following:

You have asserted (or maybe just inferred?) that in VE, the specified heat transfer can be tied to the activity of the associated hydronic loop. How might you supplement this approach to also capture the relatively different (perhaps not binary) amount of heat dumped into a space between states when the local fan is in operation vs. not? Is there a way to tie these operations together within the simulation that doesn?t require processing of interval/hourly data outside of the model and feeding that back in via customized scheduling?

For shared context, the issue of a ?changeover? situation such as with 2-pipe systems and/or seasonal scheduling is easily enough addressed on the doe2/eQuest side of things by specifying hydronic loop availability and/or operational states over on the hydronic tab (of the eQ interface). Happy to help anyone struggling with that over in the [equest-users] community ;-).

@ Everyone: I do, by the way, very very much appreciate this flood of contributions so late on a Friday with the weekend beckoning us to go out and enjoy life! It?s very encouraging to find validation in our shared worries and solution-paths through other?s experiences and approaches, and I?m certain I?m not the only one who feels that gratitude.

~Nick

[cid:image005.png at 01D515A3.47EDD880]
Nick Caton, P.E., BEMP
Senior Energy Engineer
Regional Energy Engineering Manager
Energy and Sustainability Services
Energy Performance Contracting
D
M
F
E
913 . 564 . 6361
785 . 410 . 3317
913 . 564 . 6380
nicholas.caton at se.com
15200 Santa Fe Trail Drive
Suite 204
Lenexa, KS 66219
United States
[cid:image006.png at 01D515A3.47EDD880]

Nicholas Caton2's picture
Offline
Joined: 2019-03-25
Reputation: 0

Nick
I slept on it and remembered another method I use:
I add a preheat coil with a discharge temp greater than the cooling DAT. In this way the reheat/base board is modeled at the system level so this is not good for systems with only a few wild coils or if the reheat is say electric and the AHU is gas or if the reheat is only on the perimeter zones. It is good for say hospital AHU where most zones do have reheat and all heat is via one source (ie all electric, all gas).

Melissa

Melissa
Melissa Page Crowe's picture
Joined: 2011-09-30
Reputation: 1

Nick
I slept on it and remembered another method I use:
I add a preheat coil with a discharge temp greater than the cooling DAT. In this way the reheat/base board is modeled at the system level so this is not good for systems with only a few wild coils or if the reheat is say electric and the AHU is gas or if the reheat is only on the perimeter zones. It is good for say hospital AHU where most zones do have reheat and all heat is via one source (ie all electric, all gas).

Melissa

Melissa
Melissa Page Crowe's picture
Joined: 2011-09-30
Reputation: 1

There are a number of non-linearities in this kind of control. We (UK) term
the pressure related non-linearity "valve authority". This can be quite
extreme. However, with the advent of 2-port control with Pressure
Independent Control Valves (PICV) and Differential Pressure Control Valves
(DPCV) this has helped ease the issue. Previously, with 3-port control and
constant speed pumping the agreed method to limit valve authority problems
was to design a low-pressure drop system (typically 200 Pascal/metre).
There can also be transitions between laminar and turbulent flow in the
coil itself - this affects heat transfer. I don't know enough about the
various software's models to understand if either of these phenomena can be
represented.

I think Michael Wetter's work is very relevant to the problem. I also think
there could be some success with IES VE, possibly using their ERGON
software to replicate actual building variables. The latter can also define
network airflow rather than scheduled infiltration rates. Another contender
is IDA ICE. This similarly has network airflow (IDA *has* to model network
airflow), and the ASHRAE HVAC toolkit model of heating and cooling coils
can be modeled. I have little experience of this, but it appears to model
water temperature and valve position explicitly (don't ask me anything more
detailed). There is a benefit in modeling network airflow, and I think this
gives the commercial tools an advantage.
[image: ashrae_tk_coil.png]

I seem to recall this interactive web-based training software used to let
you play with valve problems: http://www.learnhvac.org/index.php

A final note is that this is entering the realms of stochastic problems, so
the work done with Annex 66 could be relevant.

On Sat, Jun 22, 2019 at 3:42 PM Timothy Moore via Bldg-sim <

Chris Yates2's picture
Offline
Joined: 2011-10-02
Reputation: 0

Hey David!

For eQuest, we are limited to just a couple options for baseboard heating. Either can be leveraged to ?fool? the simulation into ?uncontrolled,? maximum output for one or more zones, however this can end up causing issues in accurately quantifying upstream from other heating conditioning sources.

[cid:image008.png at 01D52A73.C057C1C0]

That said, baseboards at least present such workaround options (i.e. wacky reset schedules always returning 1.0)? We?re up a creek for any other situations like airside system heating/cooling coils.

Ideally, the above reference article might describe a third option ?WILD,? which would require an additional specification of a fraction between 0.0 and 1.0. That fraction would represent a constant valve position, allowing for fully uncontrolled/open flow at all hours the loop is active (1.0) in addition to some number in between (i.e. 0.65), which would allow those of us simulating existing systems to

1. work out via calibrated modeling efforts the fraction of valves in a whole system which are malfunctioning
2. model real-world cases where a manual valve is physically cracked part way open (see: select old school radiator systems with non-hidden ?spigot? valves, intended for the occupants to actively tweak the amount of heat dumped into their room)

? on further thought, an optional SCHEDULE to apply to this new WILD COIL FRACTION input might be nice for case ?b,? allowing for a seasonal approximation of manual valve positioning, however I don?t run into many cases of facilities where such manual valve systems are operated today as they may have been intended... which is to say I commonly observe dusty cobwebs which I anticipate are more than one season old around such localized valving controls.

~Nick

[cid:image005.png at 01D515A3.47EDD880]
Nick Caton, P.E., BEMP
Senior Energy Engineer
Regional Energy Engineering Manager
Energy and Sustainability Services
Energy Performance Contracting
D
M
F
E
913 . 564 . 6361
785 . 410 . 3317
913 . 564 . 6380
nicholas.caton at se.com
15200 Santa Fe Trail Drive
Suite 204
Lenexa, KS 66219
United States
[cid:image006.png at 01D515A3.47EDD880]

Nicholas Caton2's picture
Offline
Joined: 2019-03-25
Reputation: 0

Late-breaking for EnergyPlus:
The water coils allow a user-definable MINIMUM flow. I haven?t used that field for years because it didn?t work very well at the time ??.
It might work better now and would be, I think, a perfect way to sim a wild coil.

James V Dirkes II, PE, BEMP, BCxP
Team Lead - Building Performance
616 450 8653
Foresight Management
https://fsmgmt.co/

Coffee Conversation:
But to handle human lives with no agreement as to what human beings are or what the purpose of life is - that is a formula for chaos.

Jim Dirkes's picture
Offline
Joined: 2019-05-22
Reputation: 0

Hey David!

For eQuest, we are limited to just a couple options for baseboard heating. Either can be leveraged to ?fool? the simulation into ?uncontrolled,? maximum output for one or more zones, however this can end up causing issues in accurately quantifying upstream from other heating conditioning sources.

[cid:image008.png at 01D52A73.C057C1C0]

That said, baseboards at least present such workaround options (i.e. wacky reset schedules always returning 1.0)? We?re up a creek for any other situations like airside system heating/cooling coils.

Ideally, the above reference article might describe a third option ?WILD,? which would require an additional specification of a fraction between 0.0 and 1.0. That fraction would represent a constant valve position, allowing for fully uncontrolled/open flow at all hours the loop is active (1.0) in addition to some number in between (i.e. 0.65), which would allow those of us simulating existing systems to

1. work out via calibrated modeling efforts the fraction of valves in a whole system which are malfunctioning
2. model real-world cases where a manual valve is physically cracked part way open (see: select old school radiator systems with non-hidden ?spigot? valves, intended for the occupants to actively tweak the amount of heat dumped into their room)

? on further thought, an optional SCHEDULE to apply to this new WILD COIL FRACTION input might be nice for case ?b,? allowing for a seasonal approximation of manual valve positioning, however I don?t run into many cases of facilities where such manual valve systems are operated today as they may have been intended... which is to say I commonly observe dusty cobwebs which I anticipate are more than one season old around such localized valving controls.

~Nick

[cid:image005.png at 01D515A3.47EDD880]
Nick Caton, P.E., BEMP
Senior Energy Engineer
Regional Energy Engineering Manager
Energy and Sustainability Services
Energy Performance Contracting
D
M
F
E
913 . 564 . 6361
785 . 410 . 3317
913 . 564 . 6380
nicholas.caton at se.com
15200 Santa Fe Trail Drive
Suite 204
Lenexa, KS 66219
United States
[cid:image006.png at 01D515A3.47EDD880]

Nicholas Caton2's picture
Offline
Joined: 2019-03-25
Reputation: 0