03-01-2010 09:27 PM
Hi,
I have been using DIAdem to perform numerical integrations of force with respect to displacement and pressure with respect to displacement (all acquired signals). These calculations are required to determine actuation and output energies associated with a hydraulic cylinder. The problem is I am getting impossible numbers (work out more than work in). I was wondering if it might be due to an accumulation of roundoff errors, as most of the integration intervals are about 5,000 points. Can anyone tell me with what precision DIAdem does numerical integration? The online help does not specify this. I'm guessing it's all 64-bit double precision, in which case the roundoff error is unlikely to be serious, but I'm running out of explanations.
Any advice much appreciated,
PorridgeMan.
Solved! Go to Solution.
03-02-2010 12:24 PM
Hi PorridgeMan,
All numeric channel data values are imported into the DIAdem Data Portal as DBLs (64 bit Real), regardless of their original data types on disk. All ANALYSIS routines in DIAdem that create numeric channels create DBLs, so all numeric channels in DIAdem are always DBLs and all numeric calculations in DIAdem are always carried out with DBL precision. It must be something else. Would you care to send over your data set and VBScript?
Brad Turpin
DIAdem Product Support Engineer
National Instruments
03-04-2010 03:03 AM
Hi Brad,
Thanks for that information. I tried to attach an example of the data and script I'm using, but was told the files did not have valid extensions for attachments. They were just .tdx, .tdm and .vbs files.
Anyway, I think the error is in my model, which is complicated to explain and not appropriate for this forum.
Cheers,
PorridgeMan.
03-04-2010 10:39 AM
Hi PrridgeMan,
Yeah, that's annoying about the attachments. For future reference, you can always ZIP up several files and post the resulting ZIP file, that's a "valid extension" for this forum.
Brad Turpin
DIAdem Product Support Engineer
National Instruments