I’d like to model the behaviour of a DUT that has been subjected to a specific acceleration pulse. I created a simple block jig that would hold a thin DUT. I’d like to apply an acceleration pulse in the +Z direction (relative to my model) but ONLY to the bottom face of the jig block to simplify a drop shock. The idea would be to examine the transference of that g-profile to the DUT then model it’s response for some extended time afterward.
It’s similar to the more standard drop test, where maybe I’d give this system an initial velocity right at the moment of contact with a rigid/fixed block (ie the floor) and model the behaviour of the DUT as it experiences the shock. I’d rather control the acceleration pulse though, than do a “drop test”.
The closest I found was suggested for another project;
But this applied a gravity “pulse” to all parts of the model (with some parts/surfaces defined as fixed). I tried it anyhow with a simple pulse profile using table input.
However, I really just want to apply a similar idea of a pulse, but to one surface of one body (in my simplified model, to the “bottom” face of the cylinder block) and view the response of the DUT.
(actually I’m getting an error right now, not sure where, but I know the setup with the gravity “pulse” applied to all parts is not what I want anyway)
I read through this, but unfortunately it’s not what I’m after. Imagine that model with simply the concrete block, and instead of slamming the impactor into the top of the block, we define an acceleration pulse (step, triangle, whatever) that is applied to the the whole top surface (or a node, whatever), then the concrete block can be solved to show the displacement/stress wave that moves from the top to the bottom, etc. In my case, I have a DUT embedded into the block would have the acceleration pulse imparted to it from the block, so the DUT’s response can be solved for. Overall I see it’s possible to apply any acceleration pulse to all the volumes of the bodies in the model, but not a specific face (or geometry primitive) of a given body.
Am I correct that a mechanical transient (acceleration, velocity, displacement) can only be applied globally using Model|Gravity magnitude? This seems to be the only avenue I’ve seen to define a time-based table (or formula) to apply acceleration.
It would be great if there was a way to apply such a transient to select bodies in a model (or faces of select bodies) using a time-based table or formula for acceleration/displacement/velocity.
I see this is an old thread, but I think it is worth to offer a solution, in case someone else needs to perform such analysis in the future.
As noticed, we do not have an ‘Acceleration’ boundary condition, thus we need to perform a workaround to perform the shock analysis. What I recommend is to use the ‘Large Mass Method’, an old reliable in the FEA world.
The basic idea of the method is to connect a body with a very large mass to the faces where the acceleration should be applied, with rigid links, and then apply a force to this body F = ma, with the desired acceleration a, which is usually a curve to model the pulse.
In SimScale, the usual workflow is:
Create an additional body for the large mass. I usually do a small cube, which will only add a couple of elements to the mesh. This body to be located far from the current parts.
The large mass should be much larger than the mass of the focus part. Traditionally in the order of 1e8 times larger. Compute the mass and specify it using the density = mass / volume of the cube.
Connect the large mass to the faces where the acceleration will be applied, using a ‘Remote Displacement’ boundary condition. The behavior will be ‘Undeformable’, and the remote point is to be located far away from all the parts. In order for the remote point to not take the applied force as a reaction, we leave one free DoF, say in the direction of the applied load.
Apply the loading curve to all the faces of the large mass, using a Force boundary condition, with magnitude f = ma, using the mass of the large mass and the desired acceleration. For a pulse, use the table input with proper time steps and force values.
Don’t forget to setup the simulation control to properly capture the pulse and the following transient. I will prepare a demo project and share later in the forum, so be alert on it!
