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

How to fix warmup convergence failing during sizing?

asked 2015-06-08 07:04:41 -0600

Xandrika gravatar image

updated 2020-03-10 09:51:45 -0600

I am running several models using JEPLUS (parametrice study) and some models fail the convergence test. Any idea how to fix the problem?

I tried o increase tolerances but did not work out for all models. I increased the number of warmup days without any success. The challenge is that it is a building is only occupied from 18-8 hours and 12-14 hours. Thus, the HVAC system is only running during this time. Furthermore, it is mixed mode operation, i.e. that HVAC is only running when natural ventilation cannot fullfill the needs for cooling. It further has to be noted that Setpoints are controlled by operative temperature not by air temperature. For any hints how to solve the problem, I am very grateful.

~~~ * Warmup Convergence failing during sizing. * ~~~ * ...Environment(SizingPeriod)="WINTERDESIGNDAY-WEATHERFILE"

* ~~~ * ..Max Temp Comparison = 4.64E-002 vs Temperature Convergence Tolerance=0.40 - Pass Convergence

* ~~~ * ..Min Temp Comparison = 2.70E-005 vs Temperature Convergence Tolerance=0.40 - Pass Convergence

* ~~~ * ..Max Heat Load Comparison = 1.2628 vs Loads Convergence Tolerance=0.20 - Fail Convergence

* ~~~ * ..Max Cool Load Comparison = 0.0000 vs Loads Convergence Tolerance=0.20 - Pass Convergence

* Severe * CheckWarmupConvergence: Loads Initialization, Zone="ROOMWEST" did not converge after 25 warmup days.

* ~~~ * ...Environment(SizingPeriod)="WINTERDESIGNDAY-WEATHERFILE"

* ~~~ * ..Max Temp Comparison = 0.15 vs Temperature Convergence Tolerance=0.40 - Pass Convergence

* ~~~ * ..Min Temp Comparison = 0.34 vs Temperature Convergence Tolerance=0.40 - Pass Convergence

* ~~~ * ..Max Heat Load Comparison = 2.2887 vs Loads Convergence Tolerance=0.20 - Fail Convergence

* ~~~ * ..Max Cool Load Comparison = 0.0000 vs Loads Convergence Tolerance=0.20 - Pass Convergence*

In the error file the following recommendations are given:

*** ===== Final Error Summary ===== *** The following error categories occurred. Consider correcting or noting. *** Loads Initialization did not Converge *** ..1) very high thermal mass such as very thick concrete (solution: increase max number of warmup *** .. days in the BUILDING object); *** ..2) moderate mass and inadequate space conditioning such that the building keeps getting warmer *** .. and warmer on successive days (solution: add HVAC, check building thermal properties, *** .. check if infiltration is included, make sure HVAC properly controlled); *** ..3) a soil layer modeled below the concrete slab - (solution remove this layer and read about *** .. ground temperatures in the Auxiliary Programs document). *** ..4) unreasonable (too small) limits in the BUILDING object for temperature (.4 default) or *** .. loads tolerances (.04 default)

edit retag flag offensive close merge delete


Xandrika, it may be some of the input values from the range that you specified in jEPlus may be causing this. It is worth checking into each run folder (or excel sheet containing error and warning details) whether this is for all runs or just for few runs. I have also added jeplus into the tag list.

Waseem gravatar image Waseem  ( 2015-06-08 08:07:57 -0600 )edit

@Waseem, yes, you are right. It is just for some runs. However,the HVAC system is in all models the same. It is a parametric study for building materials and passive design (WWR, overhang, etc). I am using a HVAC template of energyplus.

Xandrika gravatar image Xandrika  ( 2015-06-08 15:24:43 -0600 )edit

@Xandrika: I thought it can be due to HVAC parameters but it can also be due to high thermal mass or other suggestions mentioned in the error file you posted. May be you should give a try to implement those suggestions as well?

Waseem gravatar image Waseem  ( 2015-06-08 15:31:48 -0600 )edit

2 Answers

Sort by ยป oldest newest most voted

answered 2015-06-24 08:05:12 -0600

Yi Zhang gravatar image

updated 2015-06-25 13:34:39 -0600

Hi @Xandrika, have you tried increasing the "Maximum Number of Warmup Days" in the Building object to see if it converges?

== Update ==

After experimenting with the IDF model @Xandrika has uploaded on the development server, it seems that the convergence problem was caused by the internal thermal mass construction. The original model uses the same construction of internal partition wall (NP_Partition) for addition thermal mass. The simulation result showing convergence failure can be seen here:

If the thermal construction is slightly modified for the west zone, i.e. using 2cm cement sand render as the outer layer instead of 1.3cm cement plaster sand aggregate (see NP_ThermalMass), the convergence problem disappears. This result can be seen here: Note that the east zone still has the same internal mass construction, therefore it is not the 1.3cm render layer itself causing troubles.

I thought the E+'s solver might not like symmetrical (identical) thermal mass in the east and west zones; but this is not the case. The model converges fine if both zones has 2cm cement sand render as the outside layer: So the cause of the convergence issues remains a mystery...

edit flag offensive delete link more


@Yi Zhang, yes I have increased the Number of Warmupdays to 365 and less models failed in convergence. However, there are still about 8 of 100 models that failes. I guess it has also to do with the weather data (but not sure). I am using Meteonorm Weatherdata generated with local station data. Due to the fact that design days are not available for my locations I use the DESIGN STATS generated by energyplus weather statisticts and conversions programm. Really strange this issues and I tried a lot but no progress:-( If you have any further idea, please, let me know. Thanks a lot.

Xandrika gravatar image Xandrika  ( 2015-06-25 04:13:04 -0600 )edit

Interesting problem! Could you find one of the 8 stubborn models and upload it here? I am wondering if even longer warn-up period would help in your case. Did you see the max heat load error narrowing when you increased the max warmup days from 25 to 365?

By the way, you can upload and run your model on (login: public/test), and then tell me the job_id. In this way I (and others) can see the simulation output, too.

Yi Zhang gravatar image Yi Zhang  ( 2015-06-25 07:08:46 -0600 )edit

Yes, @Yi Zhang, The Max Heat Load error is narrowing. I also put the barometric pressure in AFN Reference Crack conditions closer to the standard pressure at elevation of my weather file. Since then I get ONLY but still the convergence issue for Min Temp Comparison. However, now increasing the warmup days brought no change at all. I uploaded and run one of those STUBBORN models (Job Id is 14439). I hope you can help me. Thanks a lot.

Xandrika gravatar image Xandrika  ( 2015-06-25 16:22:54 -0600 )edit

Hi Xandrika, check the update in my answer... hope that will solve your problem. :-)

Yi Zhang gravatar image Yi Zhang  ( 2015-06-25 16:41:28 -0600 )edit

@Yi Zhang, thanks a lot for your effort but unfortunately the problem is only solved for Model LHS-000086. Without the InternalMass object even more models fails than before. I will share the model folder with you via mail.

Xandrika gravatar image Xandrika  ( 2015-06-29 17:10:03 -0600 )edit

answered 2015-06-10 15:57:56 -0600

What are your constructions? I was having this same exact issue using CBECC-Com. My exterior construction was a mass wall (8" of 140 lb/ft3 concrete). Once I changed it to a non-mass construction (metal frame), the error went away. I would start with your constructions.

edit flag offensive delete link more


Thanks, @ Anna Osborn, I am running a parametric study with high mass and low mass construction. Out of 100 model around 8 fails in convergence - high mass and low mass constructions. Any other idea?

Xandrika gravatar image Xandrika  ( 2015-06-15 03:21:47 -0600 )edit

Your Answer

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

Add Answer


Question Tools

1 follower


Asked: 2015-06-08 07:04:41 -0600

Seen: 2,180 times

Last updated: Jun 30 '15