Impact Analysis

[tsayed]

Gentlemen,


I require some advice on how to simulate an impact analysis without having a dynamic license of Pro/MEC. I am trying to simulate a mass falling from a given height using the base module of Pro/MEC, I was thinking to calculate the force at impact and apply it as a static load to simulate an impact study. I have3 questions,


1) How accurate will this be


2) How to calculate the force of impact


3)Has anyone ever tried this and compared it to the actual dynamic results.


Any help would be appreciated.


Kind regrads,





tsayed

 

[mikes]

Hello tsayed,


I did something on this alley some time ago. I also have an article detailing a way of solving this problem. Obviously, the couple of materials in impact is paramount to determine the impact force.


You have to give each of the meaterials(ground) and the part you drop a coefficient of restitution. If the material of the ground is concrete or steel you may consider their restitution 100%, or infinit rigidity.


Also, assuming that you drop the object on a corner you have to assume a certain permanent deformation of that corner. You cannot have the impact force acting on a point, that would give you an infinit impact force, a singularity.


If you bug me I we'll look for the papers and give you more info.


Recently I took a course in Structure and Thermal and the instructor gave us and article which deals with this subject, I have it at work.


Here's my e-mail: [email protected], if you don't find anything else send me an e-mail to give you details and maybe to send you some papers related to this subject.


Cheers,


Michael

 

[TrailBarge]

Tsayed,



I don't know how to break this to you, but the exercise would be next to useless.



The point of an impact analysis is to see how the materials respond to
a shock wave, whether that be damage or vibration. The
point is that the energy of the impact takes time to travel through the
material and thus you have ever-fluxing reactions throughout the
structure.



To oversimplify, let's assume a head-on car crash. Why are cars
shorter after slamming into a brick wall? Because the front of
the car sees the wall and stops, but the back of the car has not gotten
the message yet and keeps going until the shock wave lets it know
something is up. On a larger scale, think about a train
wreck. Assume that there is 1" of slack at each joint between
cars and travelling at 1 mph.



1mile
hour 5280
feet 12
inches&n bsp;
inch

------- X ---------- X ------------- X -------------- = 17.6 -----

hour 3600
sec&nb sp;
mile&nb sp;
foot&nb sp;&nb sp;
s ec



So if we consider each car as perfectly rigid and only look at the
connections, then it takes 1/17.6 or roughly 0.057 seconds for the
message of a wreck to get from one car to another. A train of 100
perfectly rigid cars takes at least 5.7 seconds to wreck. During
that time (say 2.7 seconds), the engine could have completely stopped,
so the connection between it and the first car is at zero stress.
The connection between the caboose and the last car is at zero, since
they are coasting along as if nothing has happened.... because nothing
has happened to it yet. Somewhere in the middle, all hell is
breaking loose.



Of course, as you go faster, the message travels faster. (This is
why the speed of sound is faster in hot air.) Also, there is
really no such thing as a perfectly rigid train car, which would slow
things down a bit.



The point is... you might expect to simulate the loads in the car
roof as a static load caused by the deceleration of all the mass behind
the roof in the time it takes for the crash to happen. Not
true.... not enough. As the shock wave passes through a given
portion of the roof, the back of the roof has not seen any load
yet. It is entirely possible to have the front and back of the
car completely unloaded but a portion in the middle highly
loaded. This is the folly of trying to simulate an impact as a
static load. Herein lies the domain of the dynamic
analysis. (Well, that and harmonic responses to time dependant
input (vibration).)



After the shock wave gets the message to the back of the car... all
bets are off. The response become so chaotic that prediction of
how the car will bounce is a matter of statistical probability based on
empirical data more than mathematical modelling. Heck, the way
cars are made today, even getting to the back of the car is assuming an
awful lot.



The generalizations and calculations you would have to perform in order
to simulate an impact as a static load would be the vast majority of
the work involved in doing the whole problem by hand... and also most
of the error. At that point, running a simulation might give you
a pretty picture, but that pretty picture would be of limited use.



I'm a huge fan of computer modelling and simulation, but here is one
plce that more arcane methods would give you answers just as good, but
without the false sense of security that "that's what the FEA said"
gives you. GIGO, as we used to say.



I wish I could have given you better news.

I suspect someone out there might flame me up one side and down the
other for this post. Not currently owning a dynamic license and
having application for one, I'd be very happy to be argued out of my
position here.



Good Luck

Paul Podbielski