[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [TRNSYS-users] GenOpt variations on Type56

Dear William,

GenOpt requires data is a specific format and writes the parameters to the TRNSYS input files in a specific format. Did you check the consistency of thta.

My experience with ESp-r-GenOpt is that extremely high values of the objective function can be handled by ESP-r but not always by GenOpt.

good luck!
________________________________________
From: William.Logie@solarenergy.ch [William.Logie@solarenergy.ch]
Sent: 17 November 2008 18:01
To: thornton@tess-inc.com
Cc: trnsys-users@cae.wisc.edu; Elimar.Frank@solarenergy.ch
Subject: Re: [TRNSYS-users] GenOpt variations on Type56

Thanks for the tips Jeff,

Through their use I was able to determine that my problem occurs when our non-standard type receives a non-zero flow input. With a near- or non-zero flow initialisation a NaN iteration error is caught in simulation studio which led me to the discovery that until the loop pump is activated, the GenOpt-called simulation runs (to the 106th hour to be precise). Enforcing a permanently zero flow to this type allows GenOpt to run TRNExe without hanging. The fact that simulation studio runs with our specific flow rate but a GenOpt-called run doesn't baffles me a little.

This leads me to the question; why does our non-standard type incur problems when called from GenOpt and not when run from within simulation studio; how does the data stream between these programs bring about such behaviour? We are willing to accept that our type is not perfect (that it has some NaN or overflow by near-zero flow or other such circumstances) but I'm not sure how to isolate the problem further than this.

Cheers,
________________________________
William Logie
Forschung / Research
Institut für Solartechnik SPF: Adresse<http://www.solarenergy.ch/spf.php?lang=de&fam=44&tab=3> / Website<http://www.solarenergy.ch/spf.php?lang=de&fam=41&tab=1>
Hochschule für Technik Rapperswil HSR
Tel. +41 55 222 48 34
Fax +41 55 222 48 44
________________________________



________________________________
From: Jeff Thornton [mailto:thornton@tess-inc.com]
Sent: Thursday, 13 November 2008 5:18 PM
To: Logie William (William.Logie@solarenergy.ch); trnsys-users@cae.wisc.edu
Subject: Re: [TRNSYS-users] GenOpt variations on Type56

A few things come to mind:

1) Turn on the Overwrite Check and NAN Check commands using the appropriate control cards; they may catch a serious error that is causing the simulation to hang.

2) Make sure that all of your assign statements are fully described (full path names) as GenOpt will often write the input file into it's own directory structure and locally defined path names will fail (.\Data.dat for example).  As an alternative, move all external data files into the directory where Genopt is creating the input and BUI files.

Let us know if any of these options help.

Jeff

Jeff Thornton
President - TESS, LLC

2916 Marketplace Drive - Suite 104
Madison WI 53719 USA

Phone: 608-274-2577
Fax: 608-278-1475
E-mail: thornton@tess-inc.com
Web: www.tess-inc.com



----- Original Message -----
From: "" <william.logie@solarenergy.ch>
Sent: Thu, November 13, 2008 3:44
Subject: [TRNSYS-users] GenOpt variations on Type56

Dear Trnsys Users,

I am introducing myself to the features of TRNSYS gradually, the most recent project with which requires some optimisation of parameters in a Type56 multi-zone building. I am using GenOpt to call a *.bui template (the *.dck remains the same and as such is not altered by GenOpt) to alter the parameters of an active slab heating layer; in particular, variations over the pipe spacing and thickness of the slab.

I have tried the example from Michaël Kummert (Using GenOpt with TRNSYS16 and Type56 - 2007-02-15) which works perfectly. However when I apply the method to my project, GenOpt parses the GenOptRun.dck to TRNExe.exe (in which my *.bui file is defined) but the simulation hangs in the first time step indefinitely. The last lines printed to the GenOptRun.lst file are following:

*** The TRNSYS components will be called in the following order:
      Unit #  109   Type #  109
      Unit #   69   Type #   69
      Unit #  331   Type #   33
      Unit #  200   Type #    2
      Unit #   15   Type #  201
      Unit #   56   Type #   56
      Unit #   28   Type #    2
      Unit #   27   Type #    2
      Unit #   23   Type #    3
      Unit #   21  ;  Type #   11
      Unit #   26   Type #  869
      Unit #   24   Type #    3
      Unit #   20   Type #    5
      Unit #   22   Type #   11
      Unit #   35   Type #   65
      Unit #   36   Type #   65
      Unit #   34   Type #   65
      Unit #   31   Type #   28
      Unit #   33   Type #   65

Interestingly enough when I call the same GenOptRun.dck that GenOpt tries to parse to TRNExe from a command line directly (e.g. 'C:\Trnsys16_1\Exe\TRNExe.exe "C:\Trnsys16_1\MyProjects\project1\GenOptRun.dck" /n') the simulation runs without hanging.

Any ideas?

Thanks in advance,
________________________________
William Logie
Forschung / Research
Institut für Solartechnik SPF: Adresse<http://www.solarenergy.ch/spf.php?lang=de&fam=44&tab=3> / Website<http://www.solarenergy.ch/spf.php?lang=de&fam=41&tab=1>
Hochschule für Technik Rapperswil HSR
Tel. +41 55 222 48 34
Fax +41 55 222 48 44
________________________________


________________________________

_______________________________________________
TRNSYS-users mailing list
TRNSYS-users@cae.wisc.edu<javascript:bodyCreateMail('TRNSYS-users%40cae.wisc.edu')>
https://www-old.cae.wisc.edu/mailman/listinfo/trnsys-users



Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm