福建省发布年度省级重点招商项目 福州24个项目入选

Question

Bit of a soft / history question which might be answerable best by people who are retired or residents of another spiritual plane.

The first open-source mesh generators that were good enough for engineering purposes were published between 1995 and 2005. But the finite element method had been used since the 1960s and (I think?) engineers had been solving problems on pretty non-trivial geometries by the 80s. What did they do in the dark times before Triangle? Did people make meshes by hand? Was it done automatically but the software was proprietary? Were simulations that required unstructured meshes considered infeasible?

Triangle and TetGen had already been around for several years by the time I started grad school so I have no historical perspective of my own here. I talked to an old-timer once who said that in his day engineers only trusted structured grid generation techniques because unstructured mesh generators weren't reliable. Anecdotal answers are fine (great even).

Answer 1

An informal data point: I did a summer internship in 1996 at Bosch, and they were interested in simulating combustion in cylinders including the inlets and outlets. The software they had did not have a mesh generator, so that step was outsourced to specialized companies that charged on the order of \$10,000 (with inflation likely \$30k today) to create high-quality meshes. My understanding at the time was that these were semi-structured meshes created mostly by hand by gluing together pieces for simpler geometries. The cost made it clear that there was a lot of manual labor involved in creating these kinds of meshes.