Following on from the previous posts, the main problems with DIY simulation approaches lie with the following ..
(a) the AFM is the legal document against which your simulation would be tested at law after an accident or serious incident ... the lawyers can be presumed to have the time and resources to obtain antipathetic expert opinion designed to demolish the integrity of your simulation .. so, beware.
(b) the simulation must reproduce the printed carpet lines so that the values are conservative. However, any conservatism costs the operator money so we endeavour to minimise conservatism and produce numbers which are on the centreline of the printed line to a suitable level of accuracy. What is a suitable accuracy ? I suggest that a magnified scan should plot the calculated point close to, and slightly conservative with respect to, the observed centreline. Does this involve a moderate investment in time resources ? .. sure does ...
(c) having done (b) we have a good picture of the ACTUAL printed lines. However, we rarely need to calculate data which is right on the lines .. now comes the interesting part wherein one has to come up with an interpolation technique which will stand up in court.
Interpolation in a "normal" carpet is comparatively straightforward. Where there are a few discontinuities (as routinely is the case) the task gets a bit tedious. Running up interpolation routines which can handle regions where the chart carpet folds back on itself can be more than a little difficult when it comes to putting hand on heart for the validation process.
One needs a reasonable knowledge of the capabilities and limitations associated with various numerical modelling techniques. Generally speaking, 3D techniques are pretty useless as the detail simulation accuracy is just not there in the outcome ... so it is usually a case of coming up with sets of very tedious 2D simulations .. and there are various ways of approaching this problem....
About the nitty gritty detail which goes into this part of the exercise I will remain mute as the likes of Mutt and I derive a dollar from doing this sort of work ....
Is the simulation work a pain in the neck ? Absolutely ... one needs to be quite painstaking in detail while working up the simulation and similarly in validating the simulation against the printed AFM data.
Is it fun ? .. yes, in the same way that stopping after bashing your head against a brick wall is fun ...