nikhilgupta10 imported these comments from Sourceforge:

owner changed from dchassin to ftuffner
status changed from new to assigned ,
status changed from assigned to accepted

Attempts to replicate the bug have failed. Contacting user for more information.

, Replying to dload:
I was unable to replicate this under both debug and release versions of the win32 client. Do you have any specifics on when this occurs? Are you running the Linux or Windows build, x32 or x64 version? Are you running the nightly build, or are you downloading and compiling the source yourself? You mention you are running this as part of the validation tests. Are you running the validate.py or validate_dev.py directly, or manually calling each file, or through some other means? It works fine for me, but with the additional details I can do some extra testing and see if I can track down the issue.

-Frank

When running the validation tests, I get errors from test_balanced_stepup_D-grY_phB.glm . Its complaining about o swing bus found\, however, in the GLM file, there is exactly one node with bustype SWING.

This is with the latest SVN release (revision 1209).

I'm thinking this is actually a bug. Please let me know if I can provide any further information.

Thanks,
Jim

, I'm running 64 bit Linux and my binary is also 64 bit; 'file' says this:
ELF 64-bit LSB executable, AMD x86-64, version 1 (SYSV), for GNU/Linux
2.6.4, dynamically linked (uses shared libs), for GNU/Linux 2.6.4, not
stripped

I'm building Gridlab-d from SVN sources. This is currently happening for
me with SVN version 1210.

I've noticed the error with 'make validate'. When I run 'python
validate_dev.py', I can see that the error messages indicating the swing
bus problem. I can confirm this by cd'ing to the directory where the test
was run and then manually running the test.

HTH,
Jim

, Replying to dload:

I'm running 64 bit Linux and my binary is also 64 bit; 'file' says this:
ELF 64-bit LSB executable, AMD x86-64, version 1 (SYSV), for GNU/Linux
2.6.4, dynamically linked (uses shared libs), for GNU/Linux 2.6.4, not
stripped

'file' is being run against the gridlabd binary here.

I'm building Gridlab-d from SVN sources. This is currently happening for
me with SVN version 1210.

HTH,
Jim

component changed from Unknown to Powerflow
milestone changed from _Release win32-2_10 to _Release src-2_00

Frank, you're going to have to run the test on the Superdome to get a similar system.

, Are you still experiencing this issue? We finally managed to get an x64 Linux build up and running on our test computer. I've tested this under both SVN 1209 and the current subversion (1257). I've encountered other errors with the make validate (having to deal with pathing issues), but not any swing bus issues.

Replying to dload:

I'm running 64 bit Linux and my binary is also 64 bit; 'file' says this:
ELF 64-bit LSB executable, AMD x86-64, version 1 (SYSV), for GNU/Linux
2.6.4, dynamically linked (uses shared libs), for GNU/Linux 2.6.4, not
stripped

I'm building Gridlab-d from SVN sources. This is currently happening for
me with SVN version 1210.

HTH,
Jim

, Hello,
Yes, I'm still getting the same swing bus error as before with SVN version 1257:
download@pequena:/work/download/work/mammoli/trunk/powerflow/autotest/test_balanced_stepup_D-grY_phB> gridlabd test_balanced_stepup_D-grY_phB.glm
WARNING [INIT] : locale TZ is empty
WARNING [INIT] : gridlabd.conf was not found
WARNING [INIT] : installation type not specified
ERROR [INIT] : init failure: NR: no swing bus found
ERROR [INIT] : model initialization failed
FATAL [INIT] : shutdown after simulation stopped prematurely
dump to 'gridlabd.xml' complete
FATAL [INIT] : environment startup failed: No such file or directory

, It appears to me that the problem is related to line 184 of powerflow/node.cpp which reads:

if ((buslist==NULL) (buslist->hit_count==0)) //Not a meter either, see if it is a load

It appears to me that buslist->hit_count is not properly incremented???

HTH,
Jim

, This is a likely candidate and where I've been trying to replicate the problem. Unfortunately, I haven't had any luck on that front so far. On the trac interface for this ticket, I'm going to attach a slightly modified node.cpp file. This file just adds a quick printout of what the buslist variable finds. If you can put this file into your powerflow folder, rebuild GridLAB-D, and then run the test_balanced_stepup_D-grY_phB.glm again, it should have some extra lines of output. If you can send those lines back, it may help me determine why this is failing, or what I need to change inside the powerflow objects to handle this case.

Replying to dload:

It appears to me that the problem is related to line 184 of powerflow/node.cpp which reads:

if ((buslist==NULL) (buslist->hit_count==0)) //Not a meter either, see if it is a load

It appears to me that buslist->hit_count is not properly incremented???

HTH,
Jim

, I replaced node.cpp with the one that you attached here. Here is the new output (Note: there is nothing in the buslist->result string):

download@pequena:/work/download/work/mammoli/trunk/powerflow/autotest/test_balanced_stepup_D-grY_phB> gridlabd test_balanced_stepup_D-grY_phB.glm
WARNING [INIT] : locale TZ is empty
WARNING [INIT] : gridlabd.conf was not found
WARNING [INIT] : installation type not specified
0 3
ERROR [INIT] : init failure: NR: no swing bus found
ERROR [INIT] : model initialization failed
FATAL [INIT] : shutdown after simulation stopped prematurely
dump to 'gridlabd.xml' complete
FATAL [INIT] : environment startup failed: No such file or directory

, Sorry for the delay in getting back to you on this. I'm not sure if you have had any luck fixing it or not.

This issue continues to confuse me. I've tried a few things to replicate the issue here, but have had no luck. I'm curious if the gl_find_objects code is working at all, or simply has problems with the \SWING\ find. I've attached a GLM file to this thread. It uses the Forward-Back Sweep solver, so shouldn't need the gl_find_objects code anywhere to solve a powerflow. The file includes a voltdump object, which also uses the gl_find_objects function. If you try to run this file, it should dump a comma-separated-variable output with 6 lines (2 header and 4 node voltage sets). If this file ends up empty (or mostly empty), it will indicate gl_find_objects is having larger problems than matching enumerations.

Please give the file a try and let me know the results (unless this is already fixed).

Replying to dload:

I replaced node.cpp with the one that you attached here. Here is the new output (Note: there is nothing in the buslist->result string):

attachment _test_unbalanced_stepdown_D-grYVoltDump.glm added

Standard FBS test file with voltdump object to test gl_find_objects further

, Using SVN version 1304 with your new glm file produces the following output:

gridlabd --debug -v test_unbalanced_stepdown_D-grY_VoltDump.glm

WARNING [INIT] : locale TZ is empty

... file 'gridlabd.conf' is 1097 bytes long
... include_file(char incname='gridlabd-download.conf', char buffer=0x0x7fffdfd859aa, int size=10214): search of GLPATH='/users/download/work/mammoli/gridlabd/rt:/users/download/work/mammoli/gridlabd/etc:/users/download/work/mammoli/gridlabd/lib:/users/download/work/mammoli/gridlabd/etc/gridlabd:/users/download/work/mammoli/gridlabd/lib/gridlabd' result is 'gridlabd-download.conf'
... gridlabd-download.conf(1): included file is 1072 bytes long
... 0 object loaded
... file 'test_unbalanced_stepdown_D-grY_VoltDump.glm' is 3611 bytes long
... /users/download/work/mammoli/gridlabd/etc/gridlabd/tzinfo.txt loaded ok
... module.c(225): module 'powerflow' not yet loaded
... module.c(301): module 'powerflow' memory allocated
... full path to library 'powerflow' is '/users/download/work/mammoli/gridlabd/lib/gridlabd/libpowerflow.so'
... module.c(333): module 'powerflow' loaded ok
... module.c(346): module 'powerflow' exports init()
... class powerflow_object registered ok
... unitfile.txt loaded ok
... class powerflow_library registered ok
... class node registered ok
... class link registered ok
... class capacitor registered ok
... class fuse registered ok
... class meter registered ok
... class line registered ok
... class line_spacing registered ok
... class overhead_line registered ok
... class underground_line registered ok
... class overhead_line_conductor registered ok
... class underground_line_conductor registered ok
... class line_configuration registered ok
... class relay registered ok
... class transformer_configuration registered ok
... class transformer registered ok
... class load registered ok
... class regulator_configuration registered ok
... class regulator registered ok
... class triplex_node registered ok
... class triplex_meter registered ok
... class triplex_line registered ok
... class triplex_line_configuration registered ok
... class triplex_line_conductor registered ok
... class switch registered ok
... class substation registered ok
... class pqload registered ok
... class voltdump registered ok
... class series_reactor registered ok
... class restoration registered ok
... class frequency_gen registered ok
... class volt_var_control registered ok
... class fault_check registered ok
... class motor registered ok
... class billdump registered ok
... module.c(407): module 'powerflow' intrinsic create_node found
... module.c(407): module 'powerflow' intrinsic create_link found
... module.c(407): module 'powerflow' intrinsic create_capacitor found
... module.c(407): module 'powerflow' intrinsic create_fuse found
... module.c(407): module 'powerflow' intrinsic create_meter found
... module.c(407): module 'powerflow' intrinsic create_line found
... module.c(407): module 'powerflow' intrinsic create_line_spacing found
... module.c(407): module 'powerflow' intrinsic create_overhead_line found
... module.c(407): module 'powerflow' intrinsic create_underground_line found
... module.c(407): module 'powerflow' intrinsic create_overhead_line_conductor found
... module.c(407): module 'powerflow' intrinsic create_underground_line_conductor found
... module.c(407): module 'powerflow' intrinsic create_line_configuration found
... module.c(407): module 'powerflow' intrinsic create_relay found
... module.c(407): module 'powerflow' intrinsic create_transformer_configuration found
... module.c(407): module 'powerflow' intrinsic create_transformer found
... module.c(407): module 'powerflow' intrinsic create_load found
... module.c(407): module 'powerflow' intrinsic create_regulator_configuration found
... module.c(407): module 'powerflow' intrinsic create_regulator found
... module.c(407): module 'powerflow' intrinsic create_triplex_node found
... module.c(407): module 'powerflow' intrinsic create_triplex_meter found
... module.c(407): module 'powerflow' intrinsic create_triplex_line found
... module.c(407): module 'powerflow' intrinsic create_triplex_line_configuration found
... module.c(407): module 'powerflow' intrinsic create_triplex_line_conductor found
... module.c(407): module 'powerflow' intrinsic create_switch found
... module.c(407): module 'powerflow' intrinsic create_substation found
... module.c(407): module 'powerflow' intrinsic create_pqload found
... module.c(407): module 'powerflow' intrinsic create_voltdump found
... module.c(407): module 'powerflow' intrinsic create_series_reactor found
... module.c(407): module 'powerflow' intrinsic create_restoration found
... module.c(407): module 'powerflow' intrinsic create_frequency_gen found
... module.c(407): module 'powerflow' intrinsic create_volt_var_control found
... module.c(407): module 'powerflow' intrinsic create_fault_check found
... module.c(407): module 'powerflow' intrinsic create_motor found
... module.c(407): module 'powerflow' intrinsic create_billdump found
... module.c(225): module 'assert' not yet loaded
... module.c(301): module 'assert' memory allocated
... full path to library 'assert' is '/users/download/work/mammoli/gridlabd/lib/gridlabd/libassert.so'
... module.c(333): module 'assert' loaded ok
... module.c(346): module 'assert' exports init()
... class double_assert registered ok
... class complex_assert registered ok
... class enum_assert registered ok
... module.c(407): module 'assert' intrinsic create_double_assert found
... module.c(407): module 'assert' intrinsic create_complex_assert found
... module.c(407): module 'assert' intrinsic create_enum_assert found
... 28 objects loaded
... starting up batch environment
... initializing objects...

ERROR [INIT] : init failure: incrementing links on nodes
ERROR [INIT] : model initialization failed
FATAL [INIT] : shutdown after simulation stopped prematurely
dump to 'gridlabd.xml' complete
FATAL [INIT] : environment startup failed: Numerical argument out of domain

... shutdown complete
... elapsed runtime 0 seconds

HTH,
Jim

, Hello,
Using SVN revision 1317, I no longer get the previous error! In fact, gridlabd runs without errors(!) and produces a 6 line estout.csv\ file containing this:

testout.csv run at 1999-12-31 19:00:00 EST on 13 nodes

node_name,voltA_real,voltA_imag,voltB_real,voltB_imag,voltC_real,voltC_imag
node:6,7199.558000,0.000000,-3599.779000,-6235.000000,-3599.779000,6235.000000
node:11,7126.141061,-34.056870,-3611.972067,-6134.934213,-3527.390270,6178.588312
node:16,1933.792859,-1227.142677,-2029.480332,-998.012755,186.130661,2206.111447
load:24,1782.947572,-1213.714443,-1782.686956,-755.432411,528.544736,1772.280581

However, I am still having problems with o swing bus found\ (on many glm files in the test suite, including test_balanced_stepup_D-grY_phB). Here is the output from the file originally referenced:
> gridlabd test_balanced_stepup_D-grY_phB.glm
WARNING [INIT] : locale TZ is empty
WARNING [INIT] : gridlabd.conf was not found
WARNING [INIT] : installation type not specified
0 3 result: []
ERROR [INIT] : init failure: NR: no swing bus found
ERROR [INIT] : model initialization failed
FATAL [INIT] : shutdown after simulation stopped prematurely
dump to 'gridlabd.xml' complete
FATAL [INIT] : environment startup failed: No such file or directory

, This is good to hear. The previous error had me extremely confused, so at least we are back to just the /No swing bus found\ problem. The output indicates the gl_find_objects routine is at least working on some level, so that is good to know. The next question should be fairly obvious, but I want to ask it just to be safe. Have you tried this on a completely fresh copy of the GridLAB-D source code, or are you simply updating a previously downloaded code set?

Replying to dload:

Hello,
Using SVN revision 1317, I no longer get the previous error! In fact, gridlabd runs without errors(!) and produces a 6 line estout.csv\ file containing this:

testout.csv run at 1999-12-31 19:00:00 EST on 13 nodes

, I've been simply running 'svn up' followed by 'make distclean; autoreconf -isfv ; ./configure ... && make && make install'. I was under the (naive?) impression that this would build the latest & greatest :)

I'm checking out a 'fresh' copy from SVN now and will build it and let you know how it goes.

Jim

, I checked out and built a fresh copy of SVN revision 1319 and still am having the same problem of o swing bus found.

HTH,
Jim

, Every once in a while, it is something simple. Unfortunately, this does not appear to be one of those cases. I'll see if I can come up with any better testing ideas to figure out why the swing bus is not being found. Thank you for the information so far, it helps narrow the problem down.

, Hello,
I'm more than glad to help out. I'm working on a (funded!) project that /needs/ gridlabd to work on our platform! :) So, anything I can do to help you out, PLEASE DO NOT HESITATE TO ASK!!! :)

I'm fairly new to using SVN (I'm still stuck in CVS land most of the time). For my edification, do you think that 'make distclean' followed by 'svn up' should get me a clean set of sources?

Thanks,
Jim

, As far as I know, using 'make distclean' and 'svn up' should be sufficient (followed by the appropriate autoreconf and such commands, of course) to get an updated copy of GridLAB-D up and running. The fact that it resolved the secondary issue that popped up (with the voltdump file I sent) seems to reinforce this. I typically just use a 'make clean', but 'make distclean' is more thorough. If anything, it helps increase the probability that you are working on a fresh compile.

, I don't think make distclean will do anything useful (if it does anything at all).

svn update will only fix up outdated files but it will not replace things you've modified. You may want to find out what you modified, say with svn diff

, Replying to dchassin:

I don't think make distclean will do anything useful (if it does anything at all).

svn update will only fix up outdated files but it will not replace things you've modified. You may want to find out what you modified, say with svn diff

Thanks David! FWIW, it appears to me that 'make distclean' essentially does 'make clean' plus removing any of the autoconf created files.

Jim

attachment [node.cpp](/attachment/ticket/357/node.cpp) added

Changed node.cpp to try different gl_find_objects types

, Replying to dload:
Sorry for the delay in getting back to you again. I've put a new node.cpp on the ticket that tries a couple different searches for the SWING bus. I get expected outputs on a Windows build, but I was unable to test it on an x64 Linux build due to issues with our test machine.

If you can try this and let me know the output, that will help narrow this down a little further. This problem still confuses me, but hopefully we are getting closer to resolving it.

Thanks,
Frank

, Here is the output from my machine:

> gridlabd test_balanced_stepup_D-grY_phB.glm
WARNING [INIT] : locale TZ is empty
WARNING [INIT] : installation type not specified
SAME SW - 0 3
SAME 2 - 1 3 ?
EQ 2 - 1 3 ?
LIKE SW - 0 3
EQ SW - 2 3
SAME sw - 0 3
(null) SWING
(null) 2
ERROR [INIT] : init failure: NR: no swing bus found
ERROR [INIT] : model initialization failed
FATAL [INIT] : shutdown after simulation stopped prematurely
dump to 'gridlabd.xml' complete
FATAL [INIT] : environment startup failed: No such file or directory

, Replying to dload:
This looks like the enumeration is not being properly passed or parsed inside gl_find_objects. As a work-around to see if other things work, you can try replacing the

FINDLIST *buslist = gl_find_objects(FL_NEW,FT_CLASS,SAME, ode\,AND,FT_PROPERTY,ustype\,SAME,\SWING\,FT_END);

line in node.cpp (line 170 of the current revision) with

FINDLIST *buslist = gl_find_objects(FL_NEW,FT_CLASS,SAME, ode\,AND,FT_PROPERTY,ustype\,SAME,\,FT_END);

and see if the powerflow file now runs and passes its autotest. Based on the output you provided, this should do the same thing (I assume it is finding the same bus). I need to consult with one of our other programmers about why this may not be passing or checking things correctly.

, Originally posted by: in reply to:/u2191/u00a024; ftuffner

Replying to dload:
I was informed that the way we were searching for SWING buses was an \it should work, but may not be reliable\ approach. As part of changeset:1354, I've adjusted the way SWING bus nodes are detected. Please update your GridLAB-D to the latest revision (be sure to remove/revert the node.cpp we've been adjusting first) and see if this has fixed the problem.

, Replying to ftuffner:
Ensure you get at least changeset:1355. changeset:1354 was actually missing a file needed to make this work.

, changeset:1355 does in fact fix the problem!

Thanks!
Jim

status changed from accepted to closed
resolution set to fixed , Replying to dload:
Glad to know it finally works. Sorry it took so long to track down.

nikhilgupta10 / GridLAB-D

#357 test_balanced_stepup_D-grY_phB -- init failure: NR: no swing bus found, #1096

testout.csv run at 1999-12-31 19:00:00 EST on 13 nodes

testout.csv run at 1999-12-31 19:00:00 EST on 13 nodes