Re: Quantifying losses of Process Variation

Name    : 
E-Mail  : 
Subject : 

Comments:
: Krishnan,

: Actually the approach is generic enough to apply to systems with nonlinear blending properties.   Please feel free to download the paper now that I have finally posted it.

: Also, If you look at Figure 6 (?) in the paper, you can see that there is some quantifiable benefit for operating as you say, "on the safe side of the spec".  Irreversible loss occurs on both sides of the spec, but is greater on the "off-spec" side (usually).

: Tod

: 
: : Tod,
: :  I have seen your paper on this. If I remember right, it assumes a linear blend in terms of properties. Some blends would not fall in this category and there is no clear solution to correct an earlier mistake in off-spec production. A safer option by the operators would be to stick well into the good side of the spec. That, in itself, is a constraint issue which can be substantiated into benefits.

: : BTW, great forum, away from all the nonsense in Yahoo!!

: : : : Tod,
: : : : As always, I would be interested in the technique that mention.  As we can discuss at our company meeting, please don't go to an inordinate amount of work to prove your point.
: : : : Thanks
: : : : Don Snowden

: : : : : A few years back, I found myself trying to defend some of the economic benefits of advanced control.  My protagonists use an argument something like this:

: : : : : "If we experience a disturbance that brings the product either side of the operational target, we just compensate by operating on the opposite side of the disturbance for awhile.   If the impurity spikes high, then we just operate with less impurity for some time and it all balances out.  Anyway, the product we sell is blended to target, so there is not really any product loss."

: : : : : I developed a formula and technique to prove (and quantify) that there is, in fact, an irreversible loss that cannot recovered by compensation or blending.  Once I figure out how to convert and post a PDF file, I'll be able to give a more thorough description of the technique.   In the meantime, let me know if there is any interest in this type of information.

: : : : : Tod


Optional Link URL:  
Link Title:         
Optional Image URL: