| |||||||
| Heatsinks / fans Questions, info, results for various heatsink/fans. |
![]() |
| | LinkBack | Thread Tools | Display Modes |
| | #1 (permalink) | |
| Apex Tech Maniac | A while back, $solid$ asked me to help him work on product development for his next generation ludacrious (however the heck he spells it). As of about two weeks ago, we had one of the initial concepts designs completed, and I was going to try and work on doing the fluid analysis while I was there. Unfortunately, I had troubles isolating the volume that the air is going to flow (required for the operation of the fan), so we never got around to doing the analysis. Then I had the idea that I might be able to do get that flow volume using ICEM CFD. Never actually got around to generating that. Instead, came up with a better idea that if I were to extract the edges that bound the volume, I would be able to create a surface using those boundaries, and then create a closed surface solid/volume. And that is precisely what I did last night. I also spent the greater part of it trying to get the meshing to work properly because of sharp edges/corners in the volume. After playing around with it for a couple of hours, I decided to round off the edges that were giving me problems, to help with the meshing. Good news is that I finally got it to go. Here's the bad news though, there are 12 fan blades, which also means 12 volmes to mesh. Each of them are about 2 M elements, which means that we are looking at nearly 25 M elements for the entire mesh. And then, this morning, I got an error message that the system was requested 11 GB of hard drive space and I only have 9 GB available, so it could not complete the meshing process. In addition to that, each element is going to have 6 variables that it is going to be solving for, therefore; the entire system would be solving for about 150 M variables each iteration, for 100 iterations. And at 2.3 M elements, on a case where there's heat transfer, it took between 3.5 to 4.5 hours to solve for the whole thing; so it will translate to roughly 21 to 27 minutes PER iteration, or anywhere between 35 to 45 HOURS to complete. (Based on initial projections.) This would represent the single largest and longest analysis that I have ever done. | |
| | | |
![]() |
| Bookmarks |
| Thread Tools | |
| Display Modes | |
| |
Similar Threads | ||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| The Register // Data Analysis 2.0 | Gizmo | The Register RSS | 0 | 02-June-07 07:40 PM |
| Ars Technica // An interactive analysis of an interactive analysis of interactive media | Gizmo | Ars Technica RSS | 0 | 03-November-06 01:35 PM |
| ATI Catalyst 6.8 Performance Analysis | THRASHER2 | Video Card OC | 2 | 22-August-06 06:36 PM |
| Latest analysis stuff | alpha754293 | Anything Goes | 9 | 06-February-06 01:37 PM |
| Slashdot // A Rubric for IT Analysis | Gizmo | Slashdot RSS | 0 | 13-June-05 03:20 AM |