Simulation results change without editing the model

asked 2016-03-09 19:28:31 -0500

509 ●2 ●9

updated 2016-03-11 14:27:25 -0500

I encountered a bizarre problem in my simulation today. I went to run a simulation that I new was working yesterday, but found that the results were different. Yesterday the zone air temperature was maintained at 20C, but today the boiler providing hot water to the radiant floor seems to have a cap beyond which it cannot provide more heat, and as such the zone isn't getting enough heat.

I use both OpenStudio and EnergyPlus for my simulations. I am modelling an aquaponics facility, so I start by making a simulation in OpenStudio and then export the simulation to EnergyPlus so I can add in a swimming pool.

I am sure that I have not edited the simulation since I put it aside in a separate folder where working simulations are safe in case I make a big mistake and need to start over.

I also considered the possibility that the building materials have somehow changed and the heat loss from the building has increased and the boilers, while sized correctly before, are now undersized. However, I have not changed the material properties or any constructions so I do not see how that could have happened.

Has anyone else encountered this problem before? I tried making a brand new simulation so that no relics of the old 'broken' simulation could affect the results, and I still end up with the same problem of the zone not receiving enough heat.

Thank you.

edit retag flag offensive close merge delete

Comments

Are you running through EPlaunch?. Same computer, same Idf, same epw, same version of E+?

David Goldwasser ( 2016-03-10 08:59:52 -0500 )edit

Yes, I run each simulation through EPlaunch on my laptop. The idf, epw and E+ version are the same. If I copy an idf to a new folder does that affect the simulation?

RChidwick ( 2016-03-10 10:35:45 -0500 )edit

No, that shouldn't matter. I assume you are not changing your design days either, as that would be changing the IDF.

David Goldwasser ( 2016-03-10 11:11:55 -0500 )edit

No, the design days are the same.

RChidwick ( 2016-03-10 14:25:38 -0500 )edit

Is it possible that updating to a newer version of E+ and OpenStudio is the problem?

RChidwick ( 2016-03-25 15:04:28 -0500 )edit

add a comment

2 Answers

Sort by » oldest newest most voted

answered 2016-03-10 11:23:10 -0500

Edwin

1357 ●3 ●10 http://github.com/myoldmopar

So as part of the EnergyPlus development process, we run tests on every commit and we get 100% repeatable results. Not just down to a tiny epsilon, but the same outputs. I don't see any possibility that it spontaneously started producing different results. The only possibility I see is that if your simulation is relying on the Energy+.ini file and it is pointing to some other folder to find some outside dependency and that accidentally changes. From an EnergyPlus standpoint, if you give the same idd, idf, and epw, (and any other outside dependency files such as csv schedules), it will produce the same results.

edit flag offensive delete link

Comments

It is possible that OpenStudio translation to IDF results in different ordering of objects in the IDF between runs, this should not cause differences but it could.

macumber ( 2016-03-10 13:32:58 -0500 )edit

Yeah, the order dependence is known, but that's all contained within a convergence loop so that the numbers eventually come out similar. For example, if you order the surfaces differently, the resulting converged state will be very close but not exact, so the HVAC cycling may occur on different time steps. Nothing like the boilers simply never running, etc.

Edwin ( 2016-03-10 14:23:07 -0500 )edit

Interesting. I will be sure to not move simulations from now on, since it sounds like it could be causing an issue. I figured that each run in E+ would produce the same results, which is why I am so confused about what I could have changed to dramatically alter the simulation. I don't remember changing flow rates, boiler size, etc. (i.e. things that would directly affect the heating provided to the space).

RChidwick ( 2016-03-10 14:31:11 -0500 )edit

It's also very possible to get confused moving files around, it's possible that is what happened?

macumber ( 2016-03-10 15:07:29 -0500 )edit

I don't think so. Each time I alter a simulation in some way that I fear may cause a fatal error, I make a copy of the simulation and rename it, appending a number on the end. The simulations all remain in the same folder, and I haven't have problems with any new iteration of the simulation. But when I moved the simulation to a new folder I encountered problems, so it seems that it could be that the simulation is relying on an Energy+.ini file from another folder. How do I figure out if that is the problem?

RChidwick ( 2016-03-11 14:32:14 -0500 )edit

add a comment

answered 2016-03-11 13:23:34 -0500

Annie Marston
4297 ●4 ●31 http://www.resustain.com

I have found when I have pulled a model from Open Studio to EnergyPlus that the OperationSchemes on the boiler have had issues. See my question and the answer here: https://unmethours.com/question/5019/...