Rank: 2286 based on 67 downloads (last 30 days) and 1 file submitted
photo

Etienne Coetzee

E-mail
Company/University
University of Bristol

Personal Profile:
Professional Interests:

 

Watch this Author's files

 

Files Posted by Etienne
Updated   File Tags Downloads
(last 30 days)
Comments Rating
13 Jul 2011 Screenshot Dynamical Systems Toolbox Bifurcation analysis of dynamical systems. Integration of AUTO bifurcation software into MATLAB. Author: Etienne Coetzee bifurcation, coninuation, coninuation, dynamical systems, fold, hopf 67 23
  • 5.0
5.0 | 4 ratings
Comments and Ratings by Etienne View all
Updated File Comments Rating
28 Feb 2014 Dynamical Systems Toolbox Bifurcation analysis of dynamical systems. Integration of AUTO bifurcation software into MATLAB. Author: Etienne Coetzee

Hi Arne

I have not tried this before, although we have done something similar with ADAMS. In ADAMS there is a setting where you can either use ADAMS for the derivative calculation or Simulink. The option needs to be set to Simulink. Maybe VLM has something similar. I cannot even offer to try and debug this because we do not have a VLM license.

You may also want to check the size of your model, compared to the size defined in Ndim. We had problems before when Ndim becomes larger than 100, and we still do not really know why.

Another option may be to construct a reduced order model from the VLM model. You can for instance see if you can use Eureqa or similar tools on a time history of your model, and see if you can get some ODE's from this. You will need to provide enough information though.

Another option may be to try and generate C-code from VLM, and see if you can link this in to vanilla AUTO.

Unfortunately the new version is not available yet. We are currently trying to generate a TLC script that will generate c-code from a Simulink model and link this in to AUTO. This is tricky and I have not had a lot of time to work on this.

Regards

Etienne

20 Dec 2013 Dynamical Systems Toolbox Bifurcation analysis of dynamical systems. Integration of AUTO bifurcation software into MATLAB. Author: Etienne Coetzee

Hi Arne

I am on holiday until the new year. I will try to look at it then. We are also working on a newer version of the toolbox, so we may update the toolbox as well in the new year. We are trying to add some new features where you can run Simulink Coder code to make the runs faster.

Regards

Etienne

14 Nov 2013 Dynamical Systems Toolbox Bifurcation analysis of dynamical systems. Integration of AUTO bifurcation software into MATLAB. Author: Etienne Coetzee

Hi Olle

I have not tested this with gfortran, which is what I assume you are using. You can get the Intel compiler for Linux for free, so maybe give that a go first.

R2012b has a new help interface. Check at the bottom of the help screen for "Supplemental Software" or type "doc -classic" on the command line. This should bring up the menu with the toolbox. docroot is only valid if you install to the actual docroot directory. Check to see if strcmp(dynasyshelproot,fullfile(docroot,'toolbox','dynasys')) is true. If not, place the whole name from fullfile(docroot,'toolbox','dynasys') into the info.xml file. If it is the same, make sure you have a $ sign and not a % in the file.

Regards

Etienne

03 May 2013 Dynamical Systems Toolbox Bifurcation analysis of dynamical systems. Integration of AUTO bifurcation software into MATLAB. Author: Etienne Coetzee

HI Sudharsana

Comment out the lines where "setstatusbar" is called. There seems to be some Java incompatability. You can also look at the getting started documentation to see where the files need to be installed. You can then bypass the GUI.

Regards

Etienne

03 May 2013 Dynamical Systems Toolbox Bifurcation analysis of dynamical systems. Integration of AUTO bifurcation software into MATLAB. Author: Etienne Coetzee

HI Rajesh

I had a problem recently where the mex file was not detected, even though it was on the path. Make sure no other mex files are on the path, even if it is mexw32 etc. This seems to make a difference.

I think the INTEGER*4 was removed and changed to INTEGER. The type was automatically set.

You should maybe check to see if you can get the INTEL FORTRAN compiler for Linux. They give it out for free as far as I know. Maybe only for personal use. Your code is obviously compiling so it must be a path error in some way.

Regards

Etienne

Comments and Ratings on Etienne 's Files View all
Updated File Comment by Comments Rating
28 Feb 2014 Dynamical Systems Toolbox Bifurcation analysis of dynamical systems. Integration of AUTO bifurcation software into MATLAB. Author: Etienne Coetzee Defraye, Arne

28 Feb 2014 Dynamical Systems Toolbox Bifurcation analysis of dynamical systems. Integration of AUTO bifurcation software into MATLAB. Author: Etienne Coetzee Coetzee, Etienne

Hi Arne

I have not tried this before, although we have done something similar with ADAMS. In ADAMS there is a setting where you can either use ADAMS for the derivative calculation or Simulink. The option needs to be set to Simulink. Maybe VLM has something similar. I cannot even offer to try and debug this because we do not have a VLM license.

You may also want to check the size of your model, compared to the size defined in Ndim. We had problems before when Ndim becomes larger than 100, and we still do not really know why.

Another option may be to construct a reduced order model from the VLM model. You can for instance see if you can use Eureqa or similar tools on a time history of your model, and see if you can get some ODE's from this. You will need to provide enough information though.

Another option may be to try and generate C-code from VLM, and see if you can link this in to vanilla AUTO.

Unfortunately the new version is not available yet. We are currently trying to generate a TLC script that will generate c-code from a Simulink model and link this in to AUTO. This is tricky and I have not had a lot of time to work on this.

Regards

Etienne

28 Feb 2014 Dynamical Systems Toolbox Bifurcation analysis of dynamical systems. Integration of AUTO bifurcation software into MATLAB. Author: Etienne Coetzee Defraye, Arne


Dear Mr. Coetzee,

With regard to my last post, I’ve already solved some problems myself. But now I’m encountering a particular problem, that have brought my entire master thesis to a complete standstill because I can’t find the fault I made myself. Could you please help me with following:

My master thesis deals with the research of using bifurcation theory for analysis of landing gear shimmy. I am using multi-body software, called LMS Virtual.Lab Motion. My goal is to research if there is a possibility to use the bifurcation toolbox in a coupled simulation with the multi-body software Virtual.Lab Motion. The coupled simulation connects Virtual.Lab Motion with Simulink using an S-function in Simulink, that brings all the equations to Simulink.

So now my problem is that I can’t run a Simulink bifurcation, wherein such an S-function is contained.
· The standalone (not using the bifurcation toolbox) coupled simulations work between Matlab and Virtual.Lab Motion.
· And I’ve also run some simple bifurcations with Simulink.
So the combination of those two doesn’t seem so farfetched to me.

However, following error keeps on popping up:
“Error using AUTO07gateway
Numerical instability, NaN or Inf detected in subroutine func, adjust the tolerances of the bifurcation
parameters

Error in auto/runautodst (line 161)
[obj.f7,obj.f8]=AUTO07gateway(obj.s,obj.c,obj.f7,obj.f8);

Error in auto/runauto (line 39)
obj=runautodst(obj);

Error in massaveersyst (line 36)
a{1}=runauto(a{1});”

I’ve already tried changing a lot of parameters, such as the tolerances, but it seems I’m stuck. Does this error rings a bell to you? Is it a common error with a specific origin?

There is one piece of code that I’m still wandering if the problem might lurk there. The piece of code is programmed in the ‘func’ file and calls the Simulink model. Could you tell me what the different function arguments are and what they mean?
o=nameofsimulinkproject(0,u,par(1:1),'outputs');
f=nameofsimulinkproject(0,u,par(1:1),'derivs');

Finally; is there already a release date for the new DST bifurcation toolbox? And will the toolbox have some features that could benefit this application?
I really hope that you could help me,

My most sincerest regards,
Arne

20 Dec 2013 Dynamical Systems Toolbox Bifurcation analysis of dynamical systems. Integration of AUTO bifurcation software into MATLAB. Author: Etienne Coetzee Coetzee, Etienne

Hi Arne

I am on holiday until the new year. I will try to look at it then. We are also working on a newer version of the toolbox, so we may update the toolbox as well in the new year. We are trying to add some new features where you can run Simulink Coder code to make the runs faster.

Regards

Etienne

15 Dec 2013 Dynamical Systems Toolbox Bifurcation analysis of dynamical systems. Integration of AUTO bifurcation software into MATLAB. Author: Etienne Coetzee Defraye, Arne

Dear mister Coetzee

I'm using your dynamical systems toolbox to execute some bifurcations, regarding to my master thesis. It's a nice toolbox, I'm most greatfull.
At first, all went well and I could run some simple examples of my own as well as the demos, provided with the toolbox.
Then, when i wanted to run the only demo with the use of simulink, matlab blocked and stopped working.

" --------------- DYNAMICAL SYSTEMS TOOLBOX ---------------------

USER NAME : ARNE
DATE : 15/12/2013 10:48:14

SIMULINK MODEL : massaveersys

<
BR PT TY LAB PAR(01) L2-NORM U(01) U(02)
1 1 EP 1 -5.00000E+00 0.00000E+00 0.00000E+00 0.00000E+00

------------------------------------------------------------------------
Segmentation violation detected at Sun Dec 15 10:48:14 2013
------------------------------------------------------------------------

Configuration:
Crash Decoding : Disabled
Default Encoding: windows-1252
MATLAB Root : C:\Program Files\MATLAB\R2012a
MATLAB Version : 7.14.0.739 (R2012a)
Operating System: Microsoft Windows 7
Processor ID : x86 Family 6 Model 58 Stepping 9, GenuineIntel
Virtual Machine : Java 1.6.0_17-b04 with Sun Microsystems Inc. Java HotSpot(TM) 64-Bit Server VM mixed mode
Window System : Version 6.1 (Build 7601: Service Pack 1)

Fault Count: 2

Abnormal termination:
Segmentation violation
...
"

The problem is that i don't always encounter this problem.
If I run a simple case with simulink, then it will work 5 times out of 30 runs. It seems to work pure randomly.
And if it runs, it also takes quite some time and seems to be stuck somewhere. it continuously repeats following message:
Warning: Initializing the model 'massaveersys' with the initial state specified in array format is not
recommended.

"The array format does not contain information about how the specified states are associated with the
respective blocks. Since the sorted order may change under different conditions, such associations are
necessary for achieving consistent results. Therefore, consider initializing the states using a structure
format.
This warning can disabled or changed to an error message by using set_param(modelName,
'InitInArrayFormatMsg', 'None') or set_param(modelName, 'InitInArrayFormatMsg', 'Error'). If you get the
initial state by using the save final state option, you can change the format to structural format to avoid
this warning too.
> In func at 14
In auto.runautodst at 161
In auto.auto>auto.runauto at 39
In massaveersyst at 35"

in the function file, at command line 14,their can be found the following:
o=massaveersys(0,u,par(1:3),'outputs');

So there seems to be an array problem that you didn't encounter. Any ideas how to solve my problems.
Thanks in advance,

Arne

ps: I have the following setup on my computer
- Windows 7
- Matlab r2012a
- Intel visual fortran xe 2011

Contact us