Query on mesh size

[shashank306]

Hello everyone,

I have a query regarding mesh sizes in shock loadings. Based on literature review I have come across a criterion that Mesh size must at least be 1/10th of the wavelength (corresponding to highest frequency) traversing through the medium. How do I satisfy this condition in the analysis when I don't have a defined input motion like earthquake etc. I simply have an EoS to simulate blast load.

 

[FEA way]

What kind of analysis is that ? CEL in Abaqus/Explicit ? What element types are you using ?

 

[shashank306]

It is CEL analysis in Abaqus explicit. C3D8R elements for Lagrangian elements are being used. Basically my doubt is, Is time step determination through abaqus explicit solver enough by itself?

 

[FEA way]

CEL is quite specific, in such applications, it's more like FSI (CFD + solid mechanics) where a material moves through a stationary mesh of the Eulerian domain. Mesh size considerations here mostly revolve around the size of the smallest Lagrangian part features and the interface between the two domains. There are some specific recommendations like using at least 3-5 Eulerian elements per such smallest Lagrangian domain feature.

There's a nice example "stress wave propagation in a bar" in "Getting Started with Abaqus/CAE" part of the Abaqus documentation. It discusses mesh selection in Abaqus/Explicit wave propagation analyses. It covers regular, Lagrangian-only approach but maybe you could try making a similar benchmark for CEL.