Fan coil units with hard-sized hot water coil throw UA sizing fatal error

[mjwitte]

Issue overview

User file with hard-sized hot water coils fails with UA sizing error:

   ** Severe  ** Autosizing of heating coil UA failed for Coil:Heating:Water "COILHEATING 23"
   **   ~~~   **   Bad starting values for UA
   **   ~~~   **   Lower UA estimate = 1.477957 W/m2-K (0.1% of Design Coil Load)
   **   ~~~   **   Upper UA estimate = 1477.957406 W/m2-K (100% of Design Coil Load)
   **   ~~~   **   Zone "ZN L01_ FRONT OFFICE N" coil sizing conditions (may be different than Sizing inputs):
   **   ~~~   **   Coil inlet air temperature     = 16.600 C
   **   ~~~   **   Coil inlet air humidity ratio  = 5.860E-003 kgWater/kgDryAir
   **   ~~~   **   Coil inlet air mass flow rate  = 9.428277E-002 kg/s
   **   ~~~   **   Design Coil Capacity           = 0.000 W
   **   ~~~   **   Design Coil Load               = 1477.957 W
   **   ~~~   **   Coil outlet air temperature    = 32.200 C
   **   ~~~   **   Coil outlet air humidity ratio = 1.514E-002 kgWater/kgDryAir
   **   ~~~   **   Inadequate water side capacity: in Plant Sizing for this hot water loop
   **   ~~~   **   increase design loop exit temperature and/or decrease design loop delta T
   **   ~~~   **   Plant Sizing object = HOTWATER LOOP
   **   ~~~   **   Plant design loop exit temperature = 60.000 C
   **   ~~~   **   Plant design loop delta T          = 22.222 C
   ** Severe  ** Developer Error: sizing of U-Factor Times Area Value [W/K] failed.
   **   ~~~   ** Occurs in Coil:Heating:Water COILHEATING 23
   **  Fatal  ** Preceding water coil sizing errors cause program termination

Related to #11078

Tested the defect file with the fix in #11079. The simulation runs to completion. Not sure how to evaluate if it is a good solution for this case.

Operating System (Multiple choices)

Any

Operating System Version

any

Version of EnergyPlus

25.2.0-0200beacac

Unmethours link or helpdesk ticket number

Helpdesk ticket 16885

Defect file

EnergyPlusDevSupport\DefectFiles
Weather: Napa CA

[rraustad]

The idea for #11079 is to size the water coil UA as large as possible since the design inputs would not allow the coil to size properly. This is not ideal, given the resulting large HX area, but some have stated that proceeding would be better than terminating. That would allow the user to review the results, in addition to the err warning, to make an informed decision as to what to do. In this defect file when the UA is equal to the load, the water coil exit air temp = 26 C, which is less than the design outlet air temp of 32 C. The coil capacity at this condition was 893 W (HeatingCoilLoad), which was not greater than the load/UA assumption. SolveRoot failed with -2 because of this. That branch keeps the large UA, understanding that the coil will not meet the load, and proceeds.

 Component Sizing Information, Coil:Heating:Water, COILHEATING 23, Design Size Rated Capacity [W], 457.37378
 Component Sizing Information, Coil:Heating:Water, COILHEATING 23, User-Specified Rated Capacity [W], 1477.95741
 Component Sizing Information, Coil:Heating:Water, COILHEATING 23, Design Size Maximum Water Flow Rate [m3/s], 1.61636E-005
 Component Sizing Information, Coil:Heating:Water, COILHEATING 23, Design Size U-Factor Times Area Value [W/K], 1477.95741

[mjwitte]

So are you saying that the user inputs for capacity and water flow rate are impossible given the specified temperatures?

[rraustad]

I didn't say that, the water coil model did. Is there something in this HW coil model that causes this? I find that sort of hard to believe but anything is possible. This defect file has autosized water flow and hard-sized capacity.

  Coil:Heating:Water,
    CoilHeating 23,          !- Name
    Always On Discrete,      !- Availability Schedule Name
    Autosize,                !- U-Factor Times Area Value {W/K}
    Autosize,                !- Maximum Water Flow Rate {m3/s}
    HotWater Loop Demand Splitter Outlet Node 61,  !- Water Inlet Node Name
    CoilHeating 23 Water Outlet Node,  !- Water Outlet Node Name
    L01_ FRONT OFFICE N_FPFC Cooling Coil Outlet Node,  !- Air Inlet Node Name
    L01_ FRONT OFFICE N_FPFC Outlet Node,  !- Air Outlet Node Name
    NominalCapacity,         !- Performance Input Method
    1477.95740687852,        !- Rated Capacity {W}

If Rated Capacity is also autosized the coil sizes, without warning, to 457 W. I would not have expected that. So the other coil model internal parameters (water flow rate, UA, other?) must be calculated in a way that does not allow the coil to size. Here the water flow rate is 30% of what it was when Rated Capacity was hard-sized. It seems the hard-sized capacity of 1477 W caused the maximum capacity, given other model inputs, to be 893 W, which was less than the user-entered value. Here the zone load of 457 W can be met easily with a reasonable UA. So is there an issue with the HW coil model sizing when Rated Capacity is hard-sized and water flow is autosized (i.e., sizing assumptions and/or parameter sizing order)?

 Component Sizing Information, Coil:Heating:Water, COILHEATING 23, Design Size Rated Capacity [W], 457.37378
 Component Sizing Information, Coil:Heating:Water, COILHEATING 23, Design Size Maximum Water Flow Rate [m3/s], 5.00203E-006
 Component Sizing Information, Coil:Heating:Water, COILHEATING 23, Design Size U-Factor Times Area Value [W/K], 24.43356

 Zone Sizing Information, ZN L01_ FRONT OFFICE N, Cooling, 972.68842, 972.68842, 7.13051E-002, 7.13051E-002
 Zone Sizing Information, ZN L01_ FRONT OFFICE N, Heating, 457.37378, 457.37378, 2.68475E-002, 2.68475E-002

[rraustad]

On the SolveRoot call with UA = 1477.95, the exit water temp = 16.6003, very near the entering air temp. That does make sense with such a high UA. At a leaving air temp of 25.9 C, and an "autosized" air flow rate lower than what would be typical of a 3x larger coil, the coil output cannot exceed 893 W. And given the user request for 1477 W, SolveRoot fails. Now, had the SolveRoot upper UA been lower (e.g., 50% of capacity instead of 100%), the exiting water temp would have been higher?, the exiting air temp would have been higher?, and the HW coil could have provided a larger output (higher than 1477, I don't know, no, it's still 893 W at UA = 50% of load and 893 W at 25% of load). It looks like the coil capacity can't get any higher unless the air flow increases. So there's the bad design inputs, air flow and capacity, with these design air/water temps.

Real64 UA0 = 0.001 * this->dataCapacityUsedForSizing;
Real64 UA1 = 1.0/0.5/0.25 * this->dataCapacityUsedForSizing;

So there is definitely a relationship between these parameters (air flow rate, load/capacity, UA, water temps, other?) to "reverse engineer" this model for sizing. Maybe it as simple as an air-side air flow/capacity relationship that breaks when autosizing and hard-sizing is mixed. As air flow drops the coil can only provide so much capacity no matter what the UA is. Capacity seems max'd out here based on water-side properties. I recall F. Buhl doing some permutations on sizing but can't recall the details. These equations seem to dictate this limit.

    TempAirOut = TempAirIn + Effec * CapacitanceMin * (TempWaterIn - TempAirIn) / CapacitanceAir;
    TempWaterOut = TempWaterIn - CapacitanceAir * (TempAirOut - TempAirIn) / CapacitanceWater;
    HeatingCoilLoad = CapacitanceWater * (TempWaterIn - TempWaterOut);

I still don't see the answer for what to do in a case like this (other than comparing max coil output to requested/hard-sized Rated Capacity) and provide that information here in the warning (or why design coil capacity shows 0 here, this should be max coil capacity?):

**   ~~~   **   Design Coil Capacity           = 0.000 W
**   ~~~   **   Design Coil Load               = 1477.957 W

[rraustad]

So the limit for this case can be calculated as:

HeatingCoilLoad = CapacitanceWater * (TempWaterIn - TempWaterOut);
MaxCoilCapacity = 20.58 * (60 - 16.6) = 893 W
MaxCoilCapacity = Cp,water * WaterMassFlowRate * (TempWaterIn - TempAirIn);
MaxCoilCapacity = 4185 * 0.004918 * (60 - 16.6) = 893 W

Setting the UA very large still works here (because an 893 W coil can meet a 457 W load) but adding more information to the warning will help identify the issue.