Re: [EnergyPlus_Support] Simulation Time

Touraj,

I agree with everything that Jeremiah has said. Here are some additional considerations:

o Consider making a much smaller “dummy” file with enough zones for testing. This smaller file should run a lot faster and enable faster debugging.

§ The idea is to test small portions of the facility before testing the entire facility. Prove that each of the smaller portions works properly before combining them into a single large model.

§ The “dummy” file must follow all of the normal requirements, including correct zone names.

§ Certain multi-zone systems will need adjustment of inputs to get the zone / system air balance in line

§ If you are autosizing anything, you may find that, for example, the autosized supply air flow is less than the specified exhaust air flow, causing a system to run at 100% OA continuously. (Supply air flow is small because your “dummy” file represents a smaller building with less load.)

o These might also be helpful (but are not tested yet) for models that take a long time to run, even when using the above suggestions:

§ Set Building, Solar Distribution to “MinimalShadowing”

§ Set ShadowCalculation, Calculation Frequency to a larger value than default (say, 30 days)

§ Set Timestep to 1. Accuracy is reduced, but this should help for speed until you are confident that all problems are resolved.)

§ Set ConvergenceLimits, Minimum System Timestep to a larger value (say, 15 minutes. Same caution as for timestep above)

o Some time ago on this forum, someone mentioned that they wrote a Ruby script to run (4) three month periods simultaneously on 4 PCs and then combine the results. e.g., Run Jan-Mar, Apr – Jun, Jul – Sep, Oct – Dec on 4 PCs. The script may even be in the “Files” area of the forum. This, of course, would be used after testing all of the smaller systems. In my experience, I often run the final model several times before it’s “perfect” so this may be valuable to you.
I don’t know Ruby, so if I were to do this, I’d probably use Excel to combine the results – that should not be difficult.

James V Dirkes II, PE, BEMP, LEED AP
www.buildingperformanceteam.com
Energy Analysis, Commissioning & Training Services
1631 Acacia Drive, Grand Rapids, MI 49504 USA
616 450 8653

From: EnergyPlus_Support@xxxxxxxxxxxxxxx [mailto:EnergyPlus_Support@xxxxxxxxxxxxxxx] On Behalf Of Jeremiah Crossett
Sent: Thursday, July 25, 2013 4:29 PM
To: EnergyPlus_Support@xxxxxxxxxxxxxxx
Subject: Re: [EnergyPlus_Support] Simulation Time

Touraj,

It is likely that you could simplify this model quite a bit and get faster runtime. I would suggest starting with creating thermal blocks of areas that have the same loads, schedules and orientations. Then if possible use zone and window multipliers. Also the voids you speak of might be able to be removed if they are not thermodynamically significant, as could the many sub surfaces you mention. A good model is as simple as possible whilst maintaining thermodynamic accuracy.

I would bet that there are many many zones that could be merged, and that there is much other simplification that could be done and your model would still be reasonably accurate (possibly more accurate) and would be more useful for parametric study.

If I were you I would first rezone to minimize surfaces, then check the simpler model against the more complex in terms ot EUI.

Kind regards

Jeremiah D. Crossett | Senior Analyst | Phase Change Energy Solutions
120 E. Pritchard St. | Asheboro, NC 27203 | Mobile 503-688-8951
www.phasechange.com

On Thu, Jul 25, 2013 at 1:10 PM, Touraj Ashrafian <t.ashrafian@xxxxxxxxx> wrote:
>
>
>
> Thanks for your reply. I am going to apply your recommendations however I am interested whether is this long simulation time normal for such a building or something going wrong with my model. As I said before my model is a huge shopping center with lots of void in the central corridor and about 300 zones; I am doing the simulation for whole of a year. My laptop has core i7 CPU and 8GB RAM. Me and my colleagues previous less or more similar works take shorter time to simulate although my last work that was an office building and I modeled in the DesignBuilder and export it to the EnergyPlus take a long time as well!!!
> About jE+, maybe I had a problem with my EnergyPlus v8 when I was using the jE+! I am going to check it again.
>
> --
> Touraj Ashrafian
> PhD Student - Istanbul Technical University
> Phone:+90(536)479-8007
> Email: t.ashrafian@xxxxxxxxx, ashrafian@xxxxxxxxxx
> Web: http://alpha-arch.blogspot.com , http://alpha-arch2.blogspot.com
>
>
>
> On 7/25/2013 6:57 PM, Jeremiah Crossett wrote:
>
>
> Agreed about the use of sample days such for optimization. Also you might create surrogate models using zones with ideal loads systems then use the information from them on your larger model. My models use FD@60 timesteps, so I have had to find tricks to run parametric and optimization studies. For my projects the zone surrogate model has been the most useful.
>
> Abut your statement "JESS-Plus works with EnergyPlus v.7 and is not working with v.8 I"
>
> jE+ and the JESS server work fine with Energy Plus v8.1, (I know because I use it.)
>
>
>
> Jeremiah D. Crossett | Senior Analyst | Phase Change Energy Solutions
> 120 E. Pritchard St. | Asheboro, NC 27203 | Mobile 503-688-8951
> www.phasechange.com
>
>
>
>
>
>
> On Thu, Jul 25, 2013 at 5:51 AM, Richard Raustad <RRaustad@xxxxxxxxxxxx> wrote:
>>
>>
>>
>> Are you running an annual simulation? If so, you could optimize using a
>> few days from the weather file.
>>
>>
>>
>> On 7/25/2013 2:34 AM, Touraj Ashrafian wrote:
>> > Hi All,
>> >
>> > I am simulating a huge shopping center with lots of voids, zones and
>> > sub-surfaces. Unfortunately, the time of simulation is too long in core
>> > i7 computer, about 12 hours, and it is really boring and any
>> > optimization make me really tired. Before completing the constructions
>> > and adding some sub-surfaces, the simulation time was about 5 hours. As
>> > the JESS-Plus works with EnergyPlus v.7 and is not working with v.8 I
>> > can not use it! I wanna know that how can I reduce the time or simulate
>> > the model in more than one computer and combine the results?
>> >
>> > Best Regards,
>> > Touraj
>> >
>>
>> --
>> Richard A. Raustad
>> Senior Research Engineer
>> Florida Solar Energy Center
>> University of Central Florida
>> 1679 Clearlake Road
>> Cocoa, FL 32922-5703
>> Phone: (321) 638-1454
>> Fax: (321) 638-1439 or 1010
>> Visit our web site at: http://www.fsec.ucf.edu
>>
>> UCF - From Promise to Prominence: Celebrating 50 Years
>>
>
>
>