George Laird's blog

LS-DYNA: Observations on Explicit Meshing

This is the 3rd in a series of informal articles about one engineer’s usage of LS-DYNA to solve a variety of non-crash simulation problems. The first was on LS-DYNA: Observations on Implicit Analysis, the second was on LS-DYNA: Observations on Composite Modeling and the fourth was LS-DYNA: Observations on Material Modeling.

I come from a background in implicit analysis where element quality can often be swept under the rug by the use of dense meshes and since models run quickly no one really cares about model size, i.e., ten million DOF. Whereas, what I enjoy about explicit is that it demands the upmost model preparation from the choice of element types to the creation of perfect quad and hex dominant meshes having the absolute minimum number of DOF.

What I have been noticing over the last couple of years in the explicit world is the creation of gigantic meshes that are justified by saying, “It runs just fine in four hours using 32 CPU-cores.” Although it runs, I wonder how much time was spent in debugging this beast, and also whether the mesh density was justified by experience or by the economy of using an off-shore meshing service.

What’s the Point?

CFD Modeling from Electronic Cooling to Bag House Flow to Water Treatment Basin

Thermal-Fluids or computational fluid dynamics (CFD) consulting services has been part of our practice for over 20 years with a few of our success stories provide in our CFD Case Studies section. One of our strengths is that we do a wide range of CFD projects on a month-to-month basis – it has its pros and cons. The pros are that we can cross-pollinate between projects and bring new perspectives to a client’s CFD project whereas the cons are that we are not deep experts in super-sonic flow over wings or the combustion kinetics of gases. What we like to say is that we cover 90% of the CFD marketplace with a keen understanding of the strengths and limitations of thermal-fluid modeling.

In the last couple of months we have finished projects that dealt with thermal regulation of high-density electronics packaging for device burn-in and then switched gears to the study of air flow in a large industrial baghouse. In the first project, the goal was to obtain an even temperature gradient across the devices while in the second project it was to avoid entrapment of dust particles in low-velocity or recirculating regions. Both projects had a common theme: the keen understanding of air flow through constricted passages.

The last project was for a water treatment facility where excess turbidity was noted in the outflow of the basin and the CFD study focused on obtaining near laminar flow through a large water basin and then up through a multi-layered filtration system to trap the flocculent. It was interesting work with a null result. Essentially, it wasn’t a flow problem but an upstream chemical mixing issue.

LS-DYNA: Observations on Composite Modeling

This is the 2nd in a series of informal articles about one engineer’s usage of LS-DYNA to solve a variety of non-crash simulation problems. The first was on LS-DYNA: Observations on Implicit Analysis, the third was LS-DYNA: Observations on Explicit Meshing, and the fourth was LS-DYNA: Observations on Material Modeling

My academic background is in micro-mechanics and I have a good understanding of how many angels can dance on the head of a pin. A lot of my academic work was on fracture mechanics from a theoretical aspect and whenever I got into the laboratory, it was often a crazy chase in trying to correlate real-world fracture behavior to numerical models. I find the same behavior in composites. Ask ten composite experts and one can get 20 opinions. We have a myriad of theories that I don’t even want to start mentioning. My favorite reference to the uncertainties of composite analysis is that of the World Wide Failure Exercise (WWFE) where brave scientists were given raw composite data and had to make failure predictions without having access to the experimental data. Given that they couldn’t curve fit or pick their preferred layups or whatever, the reality was that theory could match experiment within 20%. This really isn’t as bad as it sounds since experimental data has a typical range of 10%.


Our Partners

Siemens PLM Software
Livermore Software Technology Corporation
Applied CAx
Sherpa Design