Re: [EnergyPlus_Support] what can i do for convergence the result ?

[CleanTech Analytics <jeremiah@xxxxxxxxxxxxxxxxxxxxxx>]

Linda,

Thank you kindly for the response, I have been attempting to figure this out without bugging the E+ team, and again thank you for your input. 

Funny thing! I had the full model producing expected energy savings for three runs yesterday, but I was still getting warnings about convergence, so made some changes dumbly hit save, so lost my progress, now I am back to a model that runs all day and will probably say that convergence is not reached if it ever finishes simulating..

Currently I am using 60 timesteps, for 2 study days, for the large model, and a full year for the small.. I am attaching the files encase you have the time to take a look...

I do not exactly understand how the values of the loads and temperature tolerance values are to be correlated with the zones meeting the setpoint, and would appreciate it if you could provide more description? 

When the last run worked It had warnings about inside surface temperature convergence, so think this feature may have something to do with the issues I am having, would you have any further insight?

Also I found savings in my smaller model only when using the construction internal source, and I wonder how you feel about this feature being used with the PCM object?

Lastly,I checked out your website-- cute dogs

Thanks again

On Thu, Mar 8, 2012 at 2:59 PM, Linda Lawrie <linda@xxxxxxxxxxxxxx> wrote:

 

I'm not Dr Li but will provide a developer's/ usage perspective.

At 03:49 PM 3/8/2012, CleanTech Analytics wrote:

Field: Loads Convergence Tolerance Value/Field: Temperature Convergence Tolerance Value = value in Celsius for the difference in temperatures from previous iteration.

Assumptions: (not sure if number 1 or 2 is correct)

1. Smaller may be better as it would represent less difference
2. Larger may be better as it would provide a greater window of difference

Smaller is "usually" better.  What you really want is to make sure that whatever values you insert here have the zones meeting the setpoint.  This is one reason we suggest running design day inputs rather than just starting with weather periods.  Whether sizing or other (non-Summer/Winter Design day types), you may want to have your "report during warmup" turned on so you can see if the zone is actually converging or oscillating.

If you're not meeting setpoints (assuming your zone is controlled) or not reaching "steady state", then your convergence criteria is probably set too high.

Field: Inside Face Surface Temperature Convergence Criteria
Assumption: Lower values may further increase stability at the expense of longer runtimes, Higher values may decrease runtimes but lead to possible instabilities.

Question:

• Could this reference to stability be the cause convergence issues, and if so would it make since for me to use the lowest possible value?

This is a bit harder as this is a newly introduced criteria for the CondFD (conduction Finite Difference) algorithm.  Most important right now for CondFD is to have timesteps=20 or greater (number of timesteps in hour).  I played a little with this criteria the other day trying to solve something and it increased the number of iterations (which is in the next release a variable that can be output) but it didn't really make a substantial difference in the bottom line result.

I suspect that answer really didn't help very much.

Linda

Here is a link to a visualisation of the working Energy Plus model for a small building: https://docs.google.com/spreadsheet/ccc?key=0Ag7cD5ZV9QwRdGdmMThETElEaGJNWUQ3ZGU1MEZRcEE

Thanks


On Thu, Mar 8, 2012 at 1:30 PM, YuanLu Li <yli006@xxxxxxxxxxx> wrote:

 

What values were you comparing?

 

How do you match your experimental values to the simulated values?

 

Did you have modified current data weather file?

 

How do you mount your thermometers for temperature measurement?

 

Use real weather data measured at site is easy to say and difficult to obtain and insert into the EPW file.

 

 Dr. Li  

 

---

To: EnergyPlus_Support@xxxxxxxxxxxxxxx

From: salehsnjsaadi404@xxxxxxxxx

Date: Thu, 8 Mar 2012 11:33:50 -0800

Subject: Re: [EnergyPlus_Support] what can i do for convergence the result ?

 

Does the simulation underestimate or overestimate your experimental results?

You might need to consider the infiltration in your model. try to use simple convective heat transfer coefficient algorithm.

Use real weather data measured at site.  

what is the uncertainty in your measurements equipment. ? your results need to be within 1 deg difference. (3-4) degree difference is large.

regards,

From: Mohammad Kheradmand < mohammadkheradmand@xxxxxxxxx>

To: " EnergyPlus_Support@xxxxxxxxxxxxxxx" < EnergyPlus_Support@xxxxxxxxxxxxxxx>

Sent: Thursday, March 8, 2012 11:06 AM

Subject: [EnergyPlus_Support] what can i do for convergence the result ?

 

I simulate my model for coldest and hottest a day of August, for comparing results between experimental and simulation results, i got reasonable result for hottest day(almost 4-5 degree differences ), but for coldest day i have difference almost 10-12 degree ...what is my problem?

regards

--
Jeremiah D. Crossett
CleanTech Analytics
503-688-8951
www.cleantechanalytics.com



This document may contain valuable information proprietary to CleanTech Analytics which is private and confidential. It may not be shared, copied, stored or transmitted in any form without the prior written consent of CleanTech Analytics

--

Jeremiah D. Crossett

CleanTech Analytics

503-688-8951

www.cleantechanalytics.com

This document may contain valuable information proprietary to CleanTech Analytics which is private and confidential. It may not be shared, copied, stored or transmitted in any form without the prior written consent of CleanTech Analytics

Attachment not found:
d:\eudora\attach\NEWESTPAKTECHBIOPCM(HD1CD).idf

Attachment not found:
d:\eudora\attach\RoomStudy-BioPCM M27-Q212.idf