search

NREL / EnergyPlus

EnergyPlus™ is a whole building energy simulation program that engineers, architects, and researchers use to model both energy consumption and water use in buildings.
https://energyplus.net
Other
1.15k stars 392 forks source link

Cooling Towers are not being sequentially loaded, all towers are running #4262

Closed BobHenninger closed 8 years ago

BobHenninger commented 10 years ago

With v4.0, user test file with 10 cooling towers which were drastically oversized and programmed to load sequentially were all running and equally loaded even though 1 tower would have carried the load. Same was happening with test file that had only 3 cooling towers. case with only 1 cooling tower appeared to run okay. When tested with v8.1, the 1 tower and 3 tower cases ran successfully with only 1 cooling tower running and carrying the load. The v8.1 10 tower case CRASHED. All cases were run for the 3-day period Aug 1-3. v4.0 takes a long time to run (15+ minutes) but v8.1 runs much faster. Helpdesk Ticket 1777 Test files:
091220_U147_Asls.idf which has 10 cooling towers 091214_U147_Asls.idf which has 3 cooling towers Hotel_Asls_091124.idf which has 1 cooling tower

rraustad commented 8 years ago

Looks like only the first tower operates to meet the load in V8.5.

4262

rraustad commented 8 years ago

The 10 tower file fails on sizing in V8.5.

**  Fatal  ** Autosizing of cooling tower UA failed for tower COND_COOLTWR_1
rraustad commented 8 years ago

OK, all towers were grossly oversized (1E+25) leading to NAN issue and fatal on tower UA calculation. When changing the tower capacity to be 1E+6 for each tower, only the first tower operated. When changing all towers to 1E+5, it appears the towers come on as needed to meet the load. In this figure, the first 3 towers are operating to meet the load, with other coming on line to meet additional load if necessary.

4262-10towers2

Myoldmopar commented 8 years ago

@rraustad So is there anything that needs to be done to close this issue? Protect against insane (NaN-instigating) sized towers?

rraustad commented 8 years ago

I saw that the towers operated correctly when appropriately sized and moved on. I guess something is still needed here. I'll have to circle back to this when I get some time.

Myoldmopar commented 8 years ago

OK, if you don't anticipate getting back to this during the next few weeks I'll take it.

rraustad commented 8 years ago

There was a problem with sizing and is now corrected. There were two variables used to set up the RegulaFalsi routines (Par array variables) that were not set correctly and still showed -99999 or other negative number (i.e., AutoSize * something). I ran the 10 tower defect file and see only 1 tower operating, which makes sense since all towers are grossly oversized. At least now there is no fatal or crash.

4262towersizingandoperation

rraustad commented 8 years ago

@EnergyArchmage while creating a unit test for single speed tower, copied the existing Merkel unit test and converted to single speed. With MyLoad = 0, the tower outlet is equal to the set point temp while the inlet is a little higher. Should all towers be off if MyLoad = 0 ... with outlet = inlet? See issue #4927 and existing unit test = CondenserLoopTowers_MerkelNoCooling.

EnergyArchmage commented 8 years ago

If I recall the old single speed tower controls off leaving setpoint and doesn't really respond to MyLoad, like it should. Long standing "design" feature.

rraustad commented 8 years ago

See PR https://github.com/NREL/EnergyPlus/pull/5831

mitchute commented 8 years ago

Possibly related to #5153 ?

rraustad commented 8 years ago

I think #5153 might be caused by the towers not looking at MyLoad and always meeting the set point. #4927 clued me in on that subtle operation. MyLoad is not even passed into CalcSingleSpeedTower, nor to Init but a RunFlag variable is passed to Init but not used. If the tower mass flow rate is set to 0 in Init (based on RunFlag?) then the tower won't run.

mjwitte commented 8 years ago

Closed via #5831