From r.crozier at ed.ac.uk Mon Mar 5 23:11:41 2018
From: r.crozier at ed.ac.uk (Richard Crozier)
Date: Mon, 5 Mar 2018 22:11:41 +0000
Subject: [MBDyn-users] Help with simulating a magnetic gear
In-Reply-To: <76a15bf9-f2ee-4fab-c0ff-f9221df2e0f1@polimi.it>
References:
<6104e8e4-9b0f-36a4-8eec-78b772ae7884@unicamp.br>
<527e3274-13c7-59c9-91cb-2ad59c068b8d@ed.ac.uk>
<3cb65068-0e30-9fce-8be2-70aa0c1399e1@polimi.it>
<72f9d76b-0913-70c0-beb1-8ed8fe68de14@ed.ac.uk>
<360c2ed9-2414-20f1-5680-a356a2794f5a@ed.ac.uk>
<190f6273-a681-5a2b-f6c3-732273159cd5@gmail.com>
<917e0c7b-4c55-7651-0f79-39b2e35498e0@ed.ac.uk>
<76a15bf9-f2ee-4fab-c0ff-f9221df2e0f1@polimi.it>
Message-ID:
On 09/02/18 16:33, Pierangelo Masarati wrote:
> On 02/09/2018 12:48 PM, Richard Crozier wrote:
>> I've had a look in the sources under mbdyn-1.7.3/modules, but I don't
>> see any module-fab-* folders?
>
> Correct, not distributed yet.? As far as I remember, we're waiting for
> the green light from Eduardo, who wanted to apply some last minute
> changes/cleanup.? Please correct me if I'm wrong.
>
>
> Sincerely, p.
>
Following up on this, I decided to try first with standard mbdyn
elements, namely the revolute hinge, deformable axial joint, and axial
rotation. I assume that the deformable axial joint does contribute to
the Jacobian appropriately and therefore will be fairly stable? I have
got reasonable results so far.
The only problem was the lack of an appropriate scalar differentiable
sin function. One of Eduardo's files has such a function, so I modified
MBDyn to include it, I also added cos while I was at it. I didn't test
cos much yet.
I attach two patch files with the modifications to mbdyn to add sin and
cos functions. I couldn't figure out a quick way to make a diff of only
two specific files, I'm not that familiar with the tool). Are these
patches likely to make it into MBDyn? It seems to me that additional
scalar functions would be generally useful.
I have attached an example gear file using the elements as well. I am
generating my mbdyn files using Matlab/Octave, I could provide the
m-file code to anyone interested as well.
Best regards,
Richard
--
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.
-------------- next part --------------
begin: data;
problem: initial value;
end: data;
# initial value problem
begin: initial value;
initial time: 0.0;
final time: 30.0;
time step: 0.100000000000000006;
tolerance: 0.0000001;
max iterations: 20;
end: initial value;
begin: control data;
structural nodes: 3;
rigid bodies: 2;
joints: 5;
forces: 1;
gravity;
default orientation: euler123;
end: control data;
begin: nodes;
# 6 DOF structural node
structural : 1, dynamic, # label, type
position, 0.0, 0.0, 0.0, # absolute position
orientation,
matr,
1.0, 0.0, 0.0,
0.0, 1.0, 0.0,
0.0, 0.0, 1.0, # absolute orientation
velocity, 0.0, 0.0, 0.0, # absolute velocity
angular velocity, 0.0, 0.0, 0.0 # absolute angular velocity
; # end structural node
# 6 DOF structural node
structural : 2, dynamic, # label, type
position, 0.0, 0.0, 0.0, # absolute position
orientation,
matr,
1.0, 0.0, 0.0,
0.0, 1.0, 0.0,
0.0, 0.0, 1.0, # absolute orientation
velocity, 0.0, 0.0, 0.0, # absolute velocity
angular velocity, 0.0, 0.0, 0.0 # absolute angular velocity
; # end structural node
# 6 DOF structural node
structural : 3, dynamic, # label, type
position, 0.0, 0.0, 0.0, # absolute position
orientation,
matr,
1.0, 0.0, 0.0,
0.0, 1.0, 0.0,
0.0, 0.0, 1.0, # absolute orientation
velocity, 0.0, 0.0, 0.0, # absolute velocity
angular velocity, 0.0, 0.0, 0.0 # absolute angular velocity
; # end structural node
end: nodes;
begin: elements;
# ----------------------- first section is the actual gear elements ----------------------- #
joint : 4, revolute hinge,
1, # node 1 label
position, reference, node, 0.0, 0.0, 0.0,
orientation, reference, node, matr,
0.0, 0.0, 1.0,
1.0, 0.0, 0.0,
0.0, 1.0, 0.0,
2, # node 2 label
position, reference, other node, 0.0, 0.0, 0.0,
orientation, reference, other node, matr,
0.0, 0.0, 1.0,
1.0, 0.0, 0.0,
0.0, 1.0, 0.0,
friction, 0.01, # friction radius
preload, 1.0, # friction preload
discrete coulomb,
"constant roller bearing inner", const, 0.0018, # friction model
simple # friction shape function
; # end revolute hinge
joint : 5, deformable axial joint,
1, # node 1 label
position, reference, node, 0.0, 0.0, 0.0,
orientation, reference, node, matr,
0.0, 0.0, 1.0,
1.0, 0.0, 0.0,
0.0, 1.0, 0.0,
2, # node 2 label
position, reference, other node, 0.0, 0.0, 0.0,
orientation, reference, other node, matr,
0.0, 0.0, 1.0,
1.0, 0.0, 0.0,
0.0, 1.0, 0.0,
scalar function elastic isotropic, "sin wave", sin, 11111.111111111111313221, 32.0, 0.0
; # end deformable axial joint
joint : 6, axial rotation,
2, # node 1 label
position, reference, node, 0.0, 0.0, 0.0,
orientation, reference, node, matr,
0.0, 0.0, 1.0,
1.0, 0.0, 0.0,
0.0, 1.0, 0.0,
3, # node 2 label
position, reference, other node, 0.0, 0.0, 0.0,
orientation, reference, other node, matr,
0.0, 0.0, 1.0,
1.0, 0.0, 0.0,
0.0, 1.0, 0.0,
mult,
node, 2, structural, string, "omega[1]", direct,
const, 8.0
; # end axial rotation
# one-mass body
body : 7, 1, # label, node label
28.899254466411971265, # mass
0.0, 0.0, 0.0, # relative centre of mass
matr,
0.332283627854804864, 0.0, 0.0,
0.0, 0.559210207009893878, 0.0,
0.0, 0.0, 0.559210207009893878 # inertia matrix
; # end one-mass body
# one-mass body
body : 8, 3, # label, node label
45.374530377170245288, # mass
0.0, 0.0, 0.0, # relative centre of mass
matr,
1.744741442062950343, 0.0, 0.0,
0.0, 1.489524833534827009, 0.0,
0.0, 0.0, 1.489524833534827009 # inertia matrix
; # end one-mass body
# ----------------------- next section is for testing ----------------------- #
joint : 9, revolute pin,
1, # node label
position, reference, global, 0.0, 0.0, 0.0, # node relative position
orientation, reference, global, matr,
0.0, 0.0, 1.0,
1.0, 0.0, 0.0,
0.0, 1.0, 0.0, # node relative orientation
position, 0.0, 0.0, 0.0, # pin absolute position
orientation, matr,
0.0, 0.0, 1.0,
1.0, 0.0, 0.0,
0.0, 1.0, 0.0 # pin absolute orientation
; # end revolute pin
joint : 10, angular velocity,
1, # node label
1.0, 0.0, 0.0, # rotation axis
ramp,
0.069813170079773182, 0.0, 7.5, 0.0
; # end angular velocity
couple : 11,
follower,
3,
component,
ramp,
-1629.629629629630017007, 22.5, 30.0, 0.0,
inactive,
inactive
; # end structural couple
gravity : uniform, single, 0.0, 0.0, -1.0, const, 9.810000000000000497;
end: elements;
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ScalarFunctionsImpl.h.patch
Type: text/x-patch
Size: 1294 bytes
Desc: not available
URL:
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ScalarFunctionsImpl.cc.patch
Type: text/x-patch
Size: 3023 bytes
Desc: not available
URL:
From panos.cabin at gmail.com Fri Mar 16 18:33:22 2018
From: panos.cabin at gmail.com (Panos Athanasiadis)
Date: Fri, 16 Mar 2018 18:33:22 +0100
Subject: [MBDyn-users] Nonlinear dynamics of point mass on elastic catenary
Message-ID:
Dear engineers,
Is MBDyn superior to SfePy?
I am asking this because I am trying to understand which software
package best fits my needs.
I have a solid background as a physicist and I am also versed in Python
for scientific computing,
though, I am a novice regarding the use of finite elements. I am
interested in modelling the dynamics
(initial value problem) of an elastic catenary with a point mass
attached to it through a sliding link with
friction. I presume I should begin with a simpler problem. Any help
would be very much appreciated.
I would be happy to pay for instruction, if applicable.
I can also travel to Milano :-)
Thanks a lot,
Panos
From louis.gagnon at polimi.it Mon Mar 19 10:21:09 2018
From: louis.gagnon at polimi.it (Louis Gagnon)
Date: Mon, 19 Mar 2018 10:21:09 +0100
Subject: [MBDyn-users] Nonlinear dynamics of point mass on elastic
catenary
In-Reply-To:
References:
Message-ID: <6c459813-87f5-995c-2e07-574bfd32657b@polimi.it>
Dear Panos,
Welcome to the MBDyn mailing list.
First of all, MBDyn is not finite element software as in the more
general sense of calculating the stresses and strains in tridimensional
bodies. MBDyn is a multibody dynamics software.
It does however have a beam element which uses the finite volume
approach to compute deformations, stresses, and natural frequencies.
If by catenary, you are looking to do something like this:
https://www.youtube.com/watch?v=EzoHrf-RoUE and have a sliding element
attached to it, it is feasible in MBDyn.
As for friction, there are friction functions in MBDyn, but they are not
implemented in every element yet and thus the amount of work involved
depends on what exactly you are looking to achieve.
Finally, as for SfePy, I do not know this software so I can't evaluate
it comparatively. I seem to understand it is a finite element package...
Kind regards,
-Louis
On 16-Mar-2018 06:33 PM, Panos Athanasiadis wrote:
> Dear engineers,
>
> Is MBDyn superior to SfePy?
>
> I am asking this because I am trying to understand which software
> package best fits my needs.
> I have a solid background as a physicist and I am also versed in Python
> for scientific computing,
> though, I am a novice regarding the use of finite elements. I am
> interested in modelling the dynamics
> (initial value problem) of an elastic catenary with a point mass
> attached to it through a sliding link with
> friction. I presume I should begin with a simpler problem. Any help
> would be very much appreciated.
>
> I would be happy to pay for instruction, if applicable.
>
> I can also travel to Milano :-)
>
> Thanks a lot,
>
> Panos
> _______________________________________________
> MBDyn-users mailing list
> MBDyn-users at mbdyn.org
> https://mail.mbdyn.org/cgi-bin/mailman/listinfo/mbdyn-users
--
Louis Gagnon, Ph.D.
Postdoctoral fellow
Dipartimento di Scienze e Tecnologie Aerospaziali
(Department of Aerospace Science and Technology)
Politecnico di Milano, ITALY
+39 02 2399 8308
http://louisgagnon.com/research/
louis.gagnon at polimi.it
From panos.cabin at gmail.com Tue Mar 20 19:00:54 2018
From: panos.cabin at gmail.com (Panos Athanasiadis)
Date: Tue, 20 Mar 2018 19:00:54 +0100
Subject: [MBDyn-users] Nonlinear dynamics of point mass on elastic
catenary
In-Reply-To:
References:
Message-ID:
Dear all,
A kind colleague helped me get started.
I installed blender & mbdyn and began the exploration.
Am I right in assuming that a long, homogeneous rope can
be modelled as a long series of rigid bodies with elastic joins
in between? I understand that this is what has been done in
this article: http://www.scielo.br/pdf/lajss/v10n1/09.pdf
(referring to the Rigid Finite Element Method).
I would very much appreciate your advise.
What are the respective "elements" and
options that I should use from mbdyn?
Best wishes,
Panos
> I am asking this because I am trying to understand which software
> package best fits my needs.
> I have a solid background as a physicist and I am also versed in
> Python for scientific computing,
> though, I am a novice regarding the use of finite elements. I am
> interested in modelling the dynamics
> (initial value problem) of an elastic catenary with a point mass
> attached to it through a sliding link with
> friction. I presume I should begin with a simpler problem. Any help
> would be very much appreciated.
From a.zanoni.mbdyn at gmail.com Wed Mar 28 11:00:38 2018
From: a.zanoni.mbdyn at gmail.com (Andrea Zanoni)
Date: Wed, 28 Mar 2018 11:00:38 +0200
Subject: [MBDyn-users] Nonlinear dynamics of point mass on elastic
catenary
In-Reply-To:
References:
Message-ID: <0eed27ac-ca48-fb30-0c41-29abd0172236@gmail.com>
Hi Panos,
On 03/20/2018 07:00 PM, Panos Athanasiadis wrote:
> Dear all,
>
> A kind colleague helped me get started.
> I installed blender & mbdyn and began the exploration.
You might want also to take a look at Blendyn, if you (just) want to
post-process your output using Blender:
https://github.com/zanoni-mbdyn/blendyn/wiki :)
>
> Am I right in assuming that a long, homogeneous rope can
> be modelled as a long series of rigid bodies with elastic joins
> in between? I understand that this is what has been done in
> this article: http://www.scielo.br/pdf/lajss/v10n1/09.pdf
> (referring to the Rigid Finite Element Method).
Yes, I think that is the best approach when the bending stiffness of the
rope is negligible with respect to the axial stiffness.
It really depends of which kind of rope you want to model and which
systems does it belongs to, the degree of accuracy, etc...
Actually the rope doesn't even need to be homogeneus, to be modeled in
this way.
>
> I would very much appreciate your advise.
> What are the respective "elements" and
> options that I should use from mbdyn?
The simplest "rope" model might be obtained with a combination of
structural displacement nodes and rods.
If you want to introduce some lumped-parameters bending stiffness and
damping, you might want to switch to standard structural nodes (i.e.
possessing also rotation degrees of freedom) and introduce also some
deformable axial joints or deformable hinges (depending if your model is
planar or not).
The last step, to take into account a more complex material behavior,
might be the use of beams, but my guess is that it would pay off only in
the case of a cable in which bending stiffnesses are quite significant.
Hope this helps.
Cheers
Andrea
>
> Best wishes,
>
> Panos
>
>> I am asking this because I am trying to understand which software
>> package best fits my needs.
>> I have a solid background as a physicist and I am also versed in
>> Python for scientific computing,
>> though, I am a novice regarding the use of finite elements. I am
>> interested in modelling the dynamics
>> (initial value problem) of an elastic catenary with a point mass
>> attached to it through a sliding link with
>> friction. I presume I should begin with a simpler problem. Any help
>> would be very much appreciated.
>
>
> _______________________________________________
> MBDyn-users mailing list
> MBDyn-users at mbdyn.org
> https://mail.mbdyn.org/cgi-bin/mailman/listinfo/mbdyn-users
--
---
Andrea Zanoni
Dipartimento di Scienze e Tecnologie Aerospaziali
Politecnico di Milano
Ph. (+39) 02 2399 8035