[esp-r] Re: Simulation Results with Zone Temp = nan

[Jon Hand]

A recent question about a model generating NAN...

::I have generated a simple model of a 1 storey, 1 zoned building. The zone geometry, construction and operation details are all ::set. I have also set a zone control with a basic controller with 3 control periods for a weekday and 1 for Saturday and Sunday. On ::the Sunday I have set the control to 'free float'.
 
::When I run the simulation it completes very quickly - suggesting its not doing much. Looking at the results no heating load is ::associated with the zone and the zone temperature value is 'nan'. As far as I can tell I have set everything up correctly. I've ::checked the .ctl file is being picked up (by running the simulation interactively) and I have compared files against a different ::model that is working and can't see any obvious differences (i.e. omitted data).
 
::If anyone has an idea what might be causing this, or if there are a list of steps I should follow to ensure the model's validity I ::would be grateful for the advice.

There have been a few discussions about model checking.  A summary:

a) generate a model QA report, print it out and review it in conjunction
    with the model details shown in the interface
b) get someone else to look over your model (pass them the QA report)
c) reports of NAN are most often linked to controls so try running the model
    with no control (defererence the control file).
d) rebuild the zone construction files (in case the layers got confused)
e) run the model for short test periods to see if the fault is related to
    boundary conditions
f) run another model to prove that esp-r is still working (you might have
   changed something not related to the model)
g) use the 'trace' facilities in the simulator to look at the intermediate
    results

Actually NAN tends to slow down computations. And a one zone model
will often run on modern machines in only a few seconds.

Regards, Jon Hand