Hi Pierre
I have seen the video and have a question about the GH head shape. I do not
understand why the GH head changes its shape while the analysis is running,
initially it has a long narrow shape and in the end it looks more compressed
in shape. Could it be graphic related and just something seen on this video?
In AnyBody such a change in the shape is not possible while the model is
running, there would have to be several segments to replicate this.
I understand your point of not replicating the same optimization in AnyBody
and Matlab, but if you plan to impose the results found in matlab on the
AnyBody model it will be important that the models gives the same results.
Please write again if you have further questions, I was a bit unsure if the
last part of your posting was a question.
Best regards
Søren, AnyBody Support
From: anyscript@yahoogroups.com [mailto:anyscript@yahoogroups.com] On Behalf
Of Pierre-Olivier Lemieux
Sent: 19 March 2008 19:22
To: anyscript@yahoogroups.com
Subject: [AnyScript] Re: shoulder 2D model
Hi Soren,
the Matlab routine (which I’ve created entirely) is pretty
simple. For each positions, it perform a calculation of the lever arm
and the angle of the upper part of the muscle to separate the X and Y
direction of the muscle force. When the contact between the prosthesis
and the acromion occurs, the acromial reaction is possible and is
drawn on the figure. There is no recruitment muscle because there is
only one muscle in my routine (deltoid).
The strength model is a constant force model. Like you can see, the
muscle wrapping is done by using a thin rope between the origin and
insertion point.
I send you a little video of all that under (GH_Movement.zip).
I understand your idea about creating the same routine in Anybody, but
I don’t see any interest about doing that if the results between
AnyBody and Matlab are the same.
My idea was rather to find the correct movement pattern in term of
prosthetic dimensions in Matlab and then put this movement into my
AnyBody model, to find the deltoid forces. By obtaining the same
results in AnyBody and Matlab, I think this is a way to verify the
accuracy of my findings… What do you think about that ?
I have try to get the same models between AnyBody and Matlab. To do
so, I’ve put the static friction coefficient to 0 and put a constant
force model for the deltoid. Thus, I think that the differences are
just in the WeakMuscle definition. In my Matlab routine, there is no
WeakMuscle because the acromion force give us a hint that something
don’t work during the first abduction degrees (check the results.jpg).
I hope that was more clear for you.
Thank you very much.
Pierre
— In anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com> ps.com,
“AnyBody Support” <support@…> wrote:
>
> Hi Pierre
>
>
>
> I think I understand what your algorithm is doing, it finds the minimum
> deltoid force and if there is a force unbalance it moves the humeral
until
> balance is obtained.
>
>
>
> Using the build in optimizer in AnyBody something similar should be
possible
> I guess, but it will be a slightly differently approach. Basically
you will
> need a model that will always be able to balance itself through
artificial
> joint reactions, but through the optimization you can change the
motion so
> that it tries to minimize those reactions.
>
>
>
> 1 you will need the extra muscle that ensures that a force balance can
> always be obtained.
>
> 2 a design variable has to be defined, this is the translation,
maybe this
> should be done using an interpolation function…
>
> 3 an objective function has to be setup, here I would choose to
minimize the
> forces in the artificial muscles which you did not want to become
active. So
> you will be minimizing these reaction forces by moving the humeral
head to
> another position.
>
>
>
> There is a tutorial of parameter and optimization studies it might be
> helpful.
>
>
>
> I can not judge if the MatLab code is ok sorry, I can think of the
following
> questions
>
> 1 Are there any differences in the models, that in anyway could justify
> these differences?
>
> 2 What about muscles wrapping, how is this done in MatLab…
>
> 3 Muscles strength models
>
> 4 Recruitment criteria?
>
> Etc…
>
>
>
>
>
> Best regards
>
> Søren, AnyBody Support
>
>
>
>
>
> _____
>
> From: anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com> ps.com
[mailto:anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com> ps.com]
On Behalf
> Of Pierre-Olivier Lemieux
> Sent: 18 March 2008 19:18
> To: anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com> ps.com
> Subject: [AnyScript] Re: shoulder 2D model
>
>
>
> Hi Soren,
> thank you for your support. Now I have a model that represent
> more my pathology and I have also created a Matlab algorithm that find
> the movement of the humeral head and import it into my AnyBody model.
>
> My Matlab algorithm compute the minimum force in the deltoid to
> perform an increment of abduction with a moment balance (constant
> velocity). If there is a non-equilibrium of the forces and a degree of
> freedom in the direction of the force, there is automatically an
> increment of displacement. This is the case for the vertical
> displacement in my model.
>
> Now, I want to know how Anybody solve a problem. I know that for each
> step time, it use the min/max criterion to find the minimum muscle
> activity. Does it works like my Matlab routine ?
>
> I have compared the Matlab results from those obtained in Anybody from
> the same movement and in a lot of situation it gives me similar
> results. However, there is some differences that I’m not able to
> explain. Do you think that my Matlab routine is well made ?
>
> I send you my results (Matlab & AnyBody) for a particular case under
> (PO_Lemieux_shoulder_model_March_18). Just take a look and give me
> your idea about it.
>
> Thank you again.
>
> Pierre
>
> — In anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com> ps.com,
> “AnyBody Support” <support@> wrote:
> >
> > Hi Pierre
> >
> >
> >
> > The contact force you have applied works on the contact segment. This
> > element will need to be carried by the driver otherwise the model
> will be
> > unbalanced and unable to solve; it can not be left unbalanced.
> >
> > The contact force you have applied can only do a pushing force, so the
> > reactions you will see in this contact driver will only be in one
> direction.
> >
> >
> >
> > The weak muscles also need to be there, there can not be dof in the
> model
> > which are unbalanced. Even if these are supposed to have force, it
> will lead
> > to numeric problems if they are left unbalanced.
> >
> >
> >
> > The reason why the weak muscles are needed must be a moment balance
> issue
> > there appears to be situation where the deltoid muscle it self can not
> > balance the upper arm. To me it sounds reasonable, these weak muscles
> > replicated the joint reactions up/down and these would always be non
> zero.
> > In you had more muscles in the model these might be able to
control this
> > up/down dofs., and the weak muscles could be excluded.
> >
> >
> >
> > Best regards
> >
> > Søren
> >
> > _____
> >
> > From: anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com> ps.com
> [mailto:anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com> ps.com]
> On Behalf
> > Of Pierre-Olivier Lemieux
> > Sent: 20 February 2008 03:31
> > To: anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com> ps.com
> > Subject: [AnyScript] Re: shoulder 2D model
> >
> >
> >
> > Hi again Sylvain,
> > my model works well now with the Reaction.Type turned
> > on within the Contact driver. However, it makes this reaction able to
> > pull or push. For some set-up of the parameters of the ellipse, I got
> > negative forces for the Contact driver. It tells me that the permanent
> > contact with the acromion is not valid in all the case. Maybe I should
> > know the real displacement in term of the elliptic shape, prior to
> > proceed to the inverse analysis.
> >
> > Also, it is not only the deltoid that pulls on the humeral head, it is
> > also the WeakMuscles. These muscles are not there normally and when I
> > comment them, it gives me a “Muscle recruitment” error if the tStart
> > is 0. Thus, I want to know why the deltoid can’t do the complete
> > abduction by itself ?
> >
> > Thank you again.
> >
> > Pierre-Olivier
> >
> > — In anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com>
ps.com,
> > “AnyBody Support” <support@> wrote:
> > >
> > >
> > >
> > >
> > > Hi,
> > >
> > > The reason you cannot see the reaction force of
> > > ContactSurfaceLinPush is because the ContactSeg you created is not
> > > fully balanced. I was effectively talking about adding some
> > > AnyReacForce between the ContactSeg and the humerus, so that the
> > > ContactSeg is carried in all directions in term of forces and it
can
> > > transmit the reaction forces from the acromion point to the humerus
> > > head.
> > > But it’s not the only way to do it: by adding a prismatic joint you
> > > drive the ContactSeg in five directions, you just have to switch on
> > > the ReactionType of the Ellipse2ContactDriver to carry the load of
> > > the segment in the sliding direction also. Now you will be able to
> > > see appear the reaction force at the contact point.
> > >
> > > Best regards.
> > > Sylvain, AnyBody Support
> > >
> > >
> > >
> > >
> > >
> > > — In anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com>
> ps.com,
> > “Pierre-Olivier Lemieux”
> > > <pothekid@> wrote:
> > > >
> > > > Hi Sylvain,
> > > > I’ve made the changes that you told me about, but I’m
> > > not
> > > > able to make my model works well. I don’t have any
> > > > “ContactSurfaceLinPush” reaction forces and friction forces.
> > > >
> > > > In Excel, I’ve calculated the displacement of the humeral head in
> > > term
> > > > of static equations for the abduction and put those values in the
> > > > model (“SliderDriver” and “AbductionDriver”). With these values,
> > > the
> > > > displacement is upward at the beginning of abduction and downward
> > > at
> > > > the end. So, the contact between the acromion and humeral head is
> > > not
> > > > always there. For this reason, the “ContactSurfaceLinPush” is very
> > > > useful for me.
> > > >
> > > > I’m not sure to understand well when you talk about adding some
> > > > reaction forces. Is it the “AnyReacForce” object ? If it is the
> > > case,
> > > > can you help me to write it correctly ?
> > > >
> > > > My new model is named “PO_Lemieux_shoulder_3”.
> > > >
> > > > thank you for your support.
> > > >
> > > > Pierre-Olivier
> > > >
> > > > — In anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com>
> ps.com,
> > “AnyBody Support” <support@>
> > > wrote:
> > > > >
> > > > > Hi Pierre,
> > > > >
> > > > > In the file ContactSurfaceLinPush the object Base define the
> > > node
> > > > > that is applying the force. The object Target defines the node
> > > where
> > > > > the force is applied to. So by switching Base and Target object
> > > you
> > > > > are now applying the reaction forces from the center of the
> > > > > ellipsoid to p12. And the reaction forces are displayed in the
> > > Base
> > > > > frame (ellipsoid frame in this case) that’s why the arrows
> > > follow
> > > > > the abduction angle.
> > > > >
> > > > > This can be improved by applying the reaction forces from
P12 to
> > > a
> > > > > floating point on the external shape of the ellipsoid, where
the
> > > > > contact really occurs. You can create a segment (with no mass)
> > > that
> > > > > you drive to slide on the ellipsoid shape as you did with the
> > > > > AcromionDriver, and you set the distance between one point of
> > > this
> > > > > segment and P12 to 0. Then you have to create some reaction
> > > forces
> > > > > between this segment and the ellipsoid to transmit all the
> > > forces.
> > > > > So you will now have a floating contact point on the ellipsoid
> > > > > shape. By setting the Base to the new segment you should be
able
> > > to
> > > > > draw the reaction force vectors in the tangential plane of the
> > > shape.
> > > > >
> > > > > About the ContactRotPush, this applies a moment between two
> > > segments
> > > > > in contact. There are no friction forces, so this may not be
> > > what
> > > > > you need.
> > > > >
> > > > > For the external pushing force of ContactSurfaceLinPush you
talk
> > > > > about: this function actually use 5 muscles. One pushing in the
> > > > > normal direction of the surface and 4 others creating the
> > > friction
> > > > > forces in the plane of the surface combined with normal force;
> > > > > please see this link where there is a short description
> > > > >
> > > > >
> > > > >
> > > > > http://www.anybodyt
> <http://www.anybodyt <http://www.anybodyt
<http://www.anybodytech.com/116.0.html?&tx_ttnews>
ech.com/116.0.html?&tx_ttnews>
> ech.com/116.0.html?&tx_ttnews>
> > ech.com/116.0.html?&tx_ttnews[backPid]
> > > > > =37&tx_ttnews[tt_news]=23&cHash=c92c2826e9
> > > > >
> > > > >
> > > > >
> > > > > I understand you would like to activate only the friction
> > > muscles?
> > > > > That is not possible, the five muscles are linked and working
> > > > > together.
> > > > >
> > > > > And for the strange behaviour you noticed when you change the
> > > > > deltoid origin, well i don’t think it’s so strange. Because by
> > > > > changing the origin of the muscle you change the force it will
> > > have
> > > > > to develop, you also change the reaction forces in the contact
> > > > > point. But i think you don’t have a big interest in changing
the
> > > > > muscle origin and you should just keep it in its normal value.
> > > > >
> > > > > It’s correct that you did not find repository 7, it is not
> > > public, I
> > > > > forgot this, sorry. You should look in the repository 6, here
> > > you
> > > > > will find the file PedalDemoConditional in ARep\Aalborg.
> > > > >
> > > > > Best Regards
> > > > > Sylvain, AnyBodySupport
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > — In anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com>
> > ps.com, “Pierre-Olivier Lemieux”
> > > > > <pothekid@> wrote:
> > > > > >
> > > > > > Hi Sylvain,
> > > > > > thank you for your support. I’ve worked around
> > > > > > the “ContactLinePush” problem all the week-end and I’ve found
> > > that
> > > > > > by swithching the “Base” and “Target” objects, the friction
> > > > > appears.
> > > > > > However, it gives me too high results because the switch
mixed
> > > up
> > > > > > all the calculations. For exemple, the resulting friction
> > > force is
> > > > > > equal to the normal force, that is not correct. Also, the
> > > arrows
> > > > > > rotate following the abduction angle. Maybe a
“ContactRotPush”
> > > > > would
> > > > > > perform better ?
> > > > > > So, I guess that I’m on a good way to find a solution, but I
> > > still
> > > > > > need to understand how this script works, and what are the
> > > > > > accepted “RefFrame” for the “Base” and “Target” objects.
> > > > > >
> > > > > > Also, in the present case, there is an external pushing force
> > > > > > (ContactLinePush) on the shoulder upward during all abduction
> > > to
> > > > > > keep the contact between the acromion and the ellipse.
> > > > > > However, in the real case, the deltoid, while performing the
> > > > > > abduction alone, pulls on the humerus, instead
> > > > > of “ContactLinePush”,
> > > > > > and creates a contact force between the acromion and humeral
> > > head
> > > > > > (that should be the ContactLinePush). I’ve made an
> > > experimental
> > > > > set-
> > > > > > up that has shown me these results.
> > > > > >
> > > > > > Also, I’ve noticed another strange behavior. For exemple,
when
> > > I
> > > > > use
> > > > > > the acromion contact point P12 : (0,0.039,0) and deltoid
> > > origin
> > > > > > point P8 : (0,0.050,0) and use the “ContactLinPush” script,
> > > the
> > > > > > results are very different from using the same P12, but
> > > changing
> > > > > P8
> > > > > > to a closer distance (0,0.038,0) from P12. Note that the
first
> > > set
> > > > > > of values for P8 and P12 are the ones that I need.
> > > > > >
> > > > > > So, if you can help me understanding these problems, it would
> > > help
> > > > > > me a lot.
> > > > > >
> > > > > > Have a good week and thank you again.
> > > > > >
> > > > > > Pierre.
> > > > > >
> > > > > > P.S. : I’ve not found the repository 7, is it normal ?
> > > > > >
> > > > > > — In anyscript@yahoogrou
<mailto:anyscript%40yahoogroups.com>
> > ps.com, “AnyBody Support” <support@>
> > > > > > wrote:
> > > > > > >
> > > > > > > Hi Pierre-Olivier
> > > > > > >
> > > > > > > I will try to give you some more information on the file
> > > > > > > ContactSurfaceLinPush, this file can create a conditional
> > > > > contact
> > > > > > > with pushing muscles. If the two contact points are close
> > > enough
> > > > > > > (this distance is set by the user) then the muscles will
> > > have a
> > > > > > high
> > > > > > > strength and can create the contact force. If the two
> > > contact
> > > > > > points
> > > > > > > are to far from each other, the muscles will have no
> > > strength at
> > > > > > > all, and there can be no contact force.
> > > > > > > But I am not sure if this utility is useful for your model
> > > > > because
> > > > > > > you want a permanent contact of the humerus head?
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > The file also holds another utility which is the ability to
> > > > > create
> > > > > > > normal reaction force and friction force at the contact
> > > point.
> > > > > > This
> > > > > > > could maybe be interesting for your model. But you have to
> > > take
> > > > > > care
> > > > > > > to apply those forces to the acromion contact point. You
> > > > > actually
> > > > > > > applied them to the center of the humerus head instead of
> > > the
> > > > > > > contact point.
> > > > > > >
> > > > > > > I tried running the files you uploaded and it is working
> > > with
> > > > > the
> > > > > > > elliptic shape without doing any changes. However the model
> > > is
> > > > > now
> > > > > > > fully balanced and it is not possible anymore to see that
> > > point
> > > > > > you
> > > > > > > was interested in the last time (where the geometry cannot
> > > carry
> > > > > > the
> > > > > > > load) to plot the max abduction.
> > > > > > >
> > > > > > > So the solution of recreating the humeral contact
depends on
> > > the
> > > > > > > simulation you want to perform
> > > > > > >
> > > > > > > An example of using ContactSurfaceLinPush is available in
> > > the
> > > > > > > repository 7/Application/Examples/PedalDemoConditional,
as I
> > > > > > recall
> > > > > > > it the contact muscles are defined in the
JointAndDriver.any
> > > > > file.
> > > > > > >
> > > > > > > Best regards,
> > > > > > > Sylvain, AnyBody Support
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > — In anyscript@yahoogrou
> <mailto:anyscript%40yahoogroups.com>
> > ps.com, “Pierre-Olivier Lemieux”
> > > > > > > <pothekid@> wrote:
> > > > > > > >
> > > > > > > > Hi Soeren,
> > > > > > > > thank you for you recommendations. It seems to
> > > works
> > > > > > now
> > > > > > > for
> > > > > > > > the abduction. The deltoid activity is not what I want
> > > most,
> > > > > so
> > > > > > > I’ve
> > > > > > > > just commented it out.
> > > > > > > > I have an other question now. When I went to Aalborg,
John
> > > > > gave
> > > > > > me
> > > > > > > a
> > > > > > > > folder with some special scripts. One of these scripts is
> > > > > > > > “ContactSurfaceLinPush”. Like it says, it was to create a
> > > > > pushing
> > > > > > > > force on a contact node (I guess). However, I have some
> > > > > > > difficulties
> > > > > > > > to understand this script. Also, it works with a constant
> > > > > radius
> > > > > > > > humeral head, but not for my elliptic shape. Do you think
> > > this
> > > > > is
> > > > > > > > possible ?
> > > > > > > >
> > > > > > > > You can check the folder “AcromionHumerusContact” in the
> > > > > bottom
> > > > > > of
> > > > > > > my
> > > > > > > > main script (The file is “PO_Lemieux_shoulder2D.zip”). I
> > > will
> > > > > > send
> > > > > > > you
> > > > > > > > the special script, because I don’t know who has written
> > > it.
> > > > > > > >
> > > > > > > > Thank you again,
> > > > > > > >
> > > > > > > > Pierre-Olivier
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > — In anyscript@yahoogrou
> <mailto:anyscript%40yahoogroups.com>
> > ps.com, “AnyBody Support”
> > > <support@>
> > > > > > > wrote:
> > > > > > > > >
> > > > > > > > > Hi Pierre
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > I think there is a simple solution to your problem….
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > The reason why you are getting this strange solution is
> > > that
> > > > > > the
> > > > > > > > analysis
> > > > > > > > > has not been finished due to the inverse analysis, this
> > > is
> > > > > as
> > > > > > you
> > > > > > > > expect.
> > > > > > > > > The abduction angle you find by taking the max value of
> > > the
> > > > > > > > abduction angle
> > > > > > > > > function. This is fine if the analysis had finished
> > > correctly
> > > > > > > > because then
> > > > > > > > > it would look at all the values in the output function
> > > (nStep
> > > > > > > > values) and
> > > > > > > > > find the maximum one. When the analysis stops it will
> > > still
> > > > > > look
> > > > > > > at
> > > > > > > > these
> > > > > > > > > nStep values, but now some of these can
> > > contain “garbage”
> > > > > > > numbers.
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > So the remedy in this case is to change the output
> > > function
> > > > > in
> > > > > > > the
> > > > > > > > > parameterstudy to the following
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > AnyDesMeasure AbductionMax = {
> > > > > > > > >
> > > > > > > > > Val = -Main.MyModel.GH.Pos[0];
> > > > > > > > >
> > > > > > > > > };
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > This function do not take the max of any values it will
> > > just
> > > > > > > return
> > > > > > > > the last
> > > > > > > > > one … and this is what you would like in this case I
> > > guess.
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > There will be the same kind of problems with your other
> > > > > output
> > > > > > > > function in
> > > > > > > > > the parameter study as I see it. The Abduction angle is
> > > > > > reported
> > > > > > > ok
> > > > > > > > but if
> > > > > > > > > you did the same with the deltoid activity it would not
> > > work
> > > > > I
> > > > > > > believe
> > > > > > > > > because this value is not calculated in the last step
> > > > > because
> > > > > > it
> > > > > > > > fails in
> > > > > > > > > inverse analysis, it goes ok with the abduction angle
> > > > > because
> > > > > > the
> > > > > > > > kinematics
> > > > > > > > > has been solved ok.
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > So using the function
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > AnyDesMeasure DeltoidActivity = {
> > > > > > > > >
> > > > > > > > > Val = max(…Study.MaxActivity());
> > > > > > > > >
> > > > > > > > > };
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > I am not really sure what you get since you take the
max
> > > of
> > > > > > > values
> > > > > > > > which are
> > > > > > > > > initialized, and this may give strange results. So I
> > > think
> > > > > you
> > > > > > > should
> > > > > > > > > consider exporting these numbers using an output file
> > > and
> > > > > then
> > > > > > > > process them
> > > > > > > > > externally.
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > I hope this helps you move on.
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > Otherwise please write again.
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > Best regards
> > > > > > > > >
> > > > > > > > > Søren, AnyBody Support
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > _____
> > > > > > > > >
> > > > > > > > > From: anyscript@yahoogrou
> <mailto:anyscript%40yahoogroups.com>
> > ps.com
> > > > > > > [mailto:anyscript@yahoogrou
> <mailto:anyscript%40yahoogroups.com>
> > ps.com]
> > > > > > > > On Behalf
> > > > > > > > > Of Pierre-Olivier Lemieux
> > > > > > > > > Sent: 07 February 2008 18:46
> > > > > > > > > To: anyscript@yahoogrou
> <mailto:anyscript%40yahoogroups.com>
> > ps.com
> > > > > > > > > Subject: [AnyScript] shoulder 2D model
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > Hi guys,
> > > > > > > > > I’m still working with anybody to finish the first part
> > > of
> > > > > > > > > my simulations in 2 dimensions. After some revisions of
> > > the
> > > > > > > model,
> > > > > > > > > i’ve found a big error in it.
> > > > > > > > > If you don’t remember, I try to simulate the migration
> > > of
> > > > > the
> > > > > > > humeral
> > > > > > > > > head, in the case of a massive rotator cuff tear.
So, my
> > > > > model
> > > > > > > allows
> > > > > > > > > a translation of the humerus upward and a permanent
> > > contact
> > > > > of
> > > > > > > the
> > > > > > > > > prosthetic head on the acromion. This contact is
allowed
> > > by
> > > > > a
> > > > > > > driver
> > > > > > > > > (acromion driver), related to mathematical
properties of
> > > the
> > > > > > > elliptic
> > > > > > > > > shape. Because this contact is between a prosthesis and
> > > a
> > > > > > bone,
> > > > > > > it
> > > > > > > > > can’t supply a pulling force (just pushing force is
> > > allowed)
> > > > > > (It
> > > > > > > was
> > > > > > > > > the error !!). To correct that, I have switched off the
> > > > > > reaction
> > > > > > > of
> > > > > > > > > the driver and put a general “only pushing” muscle on
> > > it. So
> > > > > > > now, the
> > > > > > > > > abduction can goes until a certain degree and after
> > > that,
> > > > > the
> > > > > > > geometry
> > > > > > > > > is not able to carry the movement, and the simulation
> > > stops.
> > > > > > It
> > > > > > > is
> > > > > > > > > exactly that I want because it directly shows if the
> > > > > geometry
> > > > > > is
> > > > > > > > > efficient or not.
> > > > > > > > >
> > > > > > > > > My problem is when I try to perform a parametric
> > > simulation.
> > > > > I
> > > > > > > look on
> > > > > > > > > the maximum abduction reached during the simulation
> > > > > (kinematic
> > > > > > > > > rotational measure). The results are very strange and I
> > > > > think
> > > > > > > that
> > > > > > > > > they are not representative, probably because the
> > > simulation
> > > > > > > stops
> > > > > > > > > with an error. So, if I would get rid of this error by
> > > > > changing
> > > > > > > > > something on the model, it would maybe works.
> > > > > > > > >
> > > > > > > > > Maybe you have a little idea of what I should do (I
> > > > > hope …).
> > > > > > I
> > > > > > > send
> > > > > > > > > you my model and a picture of the results.
> > > > > > > > >
> > > > > > > > > Thank you in advance for your support and say hello to
> > > > > > everyone
> > > > > > > of
> > > > > > > > > Anybody for me 
> > > > > > > > >
> > > > > > > > > bye bye
> > > > > > > > >
> > > > > > > > > Pierre
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > [Non-text portions of this message have been removed]
> > > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
> >
> >
> >
> >
> > [Non-text portions of this message have been removed]
> >
>
>
>
>
>
> [Non-text portions of this message have been removed]
>
[Non-text portions of this message have been removed]