Theory versus Reality
- Published: February 22, 2013
If you'd like to hear from Mark Miller's own lips rather than read his column titled, "Coating Matters | Theory versus Reality," click on his podcast below:
{mp3}theory-versus-reality{/mp3}
How well can a computer program, mathematical equation, or experience predict fluid coating performance is a question I am asked on a regular basis. As with any good question, the answer is, “It depends.” I am an advocate of using predictive devices whenever possible, but the key is to know the limits of the tools. One of the great engineers I worked with when I first started my career made sure I understood the difference between tools and technology. Technology is what engineers should concentrate on, not the fancy tools we use to analyze the technology. Technology can be modified to accomplish things we never thought possible, tools measure the results. Technology is the product of our imagination, tools are the calibrated analysis of what is currently state-of-the-art.
The way I view theoretical prediction tools is exactly this way – calibrated analysis. Prediction can interpolate data already known, but you cannot extrapolate data to predict what will happen outside what you have already accomplished. In other words, if I have found what process control settings work well for coating an emulsion at 100 meters per minute and 500 meters per minute, then I can most likely predict what my process control settings should be for 250 meters per minute. However, I cannot use this historical data to predict what my control settings should be for 1000 meters per minute. The higher line speed would be outside what this particular system had data for historically. This new process world requires statistical experimental design outside the range we have operated this system in. +/- 20% variation on each process setting is helpful to cast a statistical net wide enough to capture most of this information.
Of course, you may run into equipment limitations and software imposed limit switches when you push for more predictive data, but to grow you have to invest and work at levels that were previously uncomfortable. The bottom line is that predictive equations, software, and experience can be trusted to get you in the realm of expected results for fluid coating product and process technology. Just don’t expect an exact response when you are trying to predict behavior outside the operating world your process has lived in previously.
A good example of this is when I first started coating thinner and thinner fluids onto substrates, I was told that coating 1 micron wet with a slot die was not possible. This is an example of predicting behavior of technology based off past experience. The physics involved in coating thin fluids requires equipment tolerances and geometries that may not have allowed for thin fluid coating in the past. At the time I worked on (and succeeded) in coating less than 1 micron wet with a slot die, the predictive outcome was wrong because the technology had changed. Interpolation fails when trying new things.
Remember to use tools to understand what you are currently doing, but don’t let it limit what you imagine can be accomplished in your area of technology.
If you are interested in discussing this concept further, contact Mark D. Miller, Founder and CEO of Coating Tech Service, LLC (www.coatingtechservce.com) at This email address is being protected from spambots. You need JavaScript enabled to view it. or (612) 605-6019.
