Question-and-Answer Resource for the Building Energy Modeling Community
Get started with the Help page
Ask Your Question
3

Long simulation time : Fully glazed room with a large number of glazing

asked 2020-07-08 09:06:43 -0500

Ttthomas's avatar

updated 2020-07-09 11:44:36 -0500

Hello everyone,

I'm modeling a building in which there is a fully glazed room with a large number of glazing. This room makes the simulation time a lot longer. When I run a simulation without this glazed room the simulation takes 2h but when I add this room the simulation takes 17 hours, and when I add Blinds controlled by solar radiation on each window the simulation doesn't stop and remains at the step "Warming up" without ever stopping.

I'm looking for a solution that would make the simulation faster without simplifying the geometry.

edit retag flag offensive close merge delete

Comments

Ummmm ... exactly how large is "large number of glazing surfaces"?

__AmirRoth__'s avatar __AmirRoth__  ( 2020-07-09 10:42:01 -0500 )edit

In the room there are around 319 different glazing and the volume of the room is around 20 000 Cubic Meter. In the actual project there are interior blinds on each glazing that react to solar radiation

Ttthomas's avatar Ttthomas  ( 2020-07-09 11:14:51 -0500 )edit

That is a lot of surfaces! And controllable blinds is not making it any faster.

__AmirRoth__'s avatar __AmirRoth__  ( 2020-07-09 11:29:55 -0500 )edit

2 Answers

Sort by ยป oldest newest most voted
2

answered 2020-07-08 12:41:33 -0500

updated 2020-07-10 02:38:13 -0500

With reference of your simulation purpose, and guessing you are using energyplus, to reduce sim time you can:

1- select an appropriate Solar Distribution algorithm in Building obj: leaving off "Reflections" increase speed time , as well to move from FullInterioAndExterior to FullExterior to MinimalShadowing. Generally FullExterior is a good compromise (at least as a starting point ) between acccuracy and sim time

2- use window multiplier rather than model each window with a fenestration obj

Blinds control on W/sqm hitting windows is compatible with the upspeeding solutions I stated. This could be different if you need a detailed interior solar distribution : e.g.

-to have daylighting illuminance maps, or

-for artificial lights dimming based on daylight illuminance level, or

-to model indirect solar daylight from a zone to an inner one through an internal window,

-etc

edit flag offensive delete link more

Comments

Thanks for your answer. I tried FullExterior and MinimalShadowing but the simulation still doesn't stop and remains at the step "Warming up" when I compute the building with interior blinds.

I also already used window multiplier to model windows of this room. It was simplier for me to model the building.

Ttthomas's avatar Ttthomas  ( 2020-07-09 11:22:48 -0500 )edit
2

answered 2020-07-09 11:38:38 -0500

The most recent version of EnergyPlus (9.3) has a new object called PerformancePrecisionTradeoffs. In that object, the field Zone Radiant Exchange Algorithm gives you the option to override the default ScriptF quadratic radiant exchange algorithm with a linear one called CarrollMRT. That performance improvement should be especially noticeable for zones with many surfaces. [Disclaimer: we are not yet entirely sure about the relationship between zone/surface configuration and CarrollMRT precision degradation.]

edit flag offensive delete link more

Comments

Thanks for your answer, I would try something else before trying to convert my model to the version of EnergyPlus (9.3). It seems that my issue (Simulation remaining blocked during warming up) is due to the regulation on solar radiation of the blinds, indeed when I add blinds on 24h/24 7j/7 instead of blinds controlled by solar radiation the simulation gives me results. Anyone has an idea of how I could create a kind of hysteresis in E+ that would avoid too frequently opening/closing of the blinds ? Maybe it is due to something else ? As I don't have any error report It is difficult to tell.

Ttthomas's avatar Ttthomas  ( 2020-07-13 04:58:21 -0500 )edit

Part of the slow-down with any shading device is a recalculation of the SciptF approximate view factors for a given zone every time the shading status of even a single window changes. Blinds also require a bit more computation than a shade layer. If you want to tailor the blind control, I would suggest running a simulation of a simple building, report hourly incident solar values, then use a spreadsheet rule to create an hourly blind control schedule that can be imported using Schedule:File.

MJWitte's avatar MJWitte  ( 2020-07-13 10:15:08 -0500 )edit

Thanks, I have tried this solution :

Run a simulation of a simple building, report hourly incident solar values, then use a spreadsheet rule to create an hourly blind control schedule that can be imported using.

But the simulation still remain at the step Warming up without stoping. Can it be due to the radiant floor which is set in autosize ?

Ttthomas's avatar Ttthomas  ( 2020-08-27 06:57:32 -0500 )edit

I would have expected some improvement (assuming your schedule does not change the blind status very often during the day). That said, radiant systems invoke extra radiant exchange calculations, so this same model with a non-radiant floor should run somewhat faster, but I would expect it to be on the order of only 2x.

MJWitte's avatar MJWitte  ( 2020-09-01 17:39:52 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Careers

Question Tools

1 follower

Stats

Asked: 2020-07-08 08:43:56 -0500

Seen: 292 times

Last updated: Jul 10 '20