Subject: Beer Color Prediction Algorithm
Date: 1992-03-31 13:35:32 GMT
(I sent this reply on Friday last, but it seems to have been
lost in the bowels of someone's mail system. I appologize in
advance if it should eventually repeat.)
smithey@rmtc.Central.Sun.COM (Brian Smithey) writes:
> >>>>> On Wed, 25 Mar 92 19:10:02 CST, firstname.lastname@example.org (George J Fix) said:
> George> After the book was off to Brewers Publ., Darryl Richman sent me
> George> a remarkable new formula for the a priori prediction of wort
> George> color.
> George> We sure hope Darryl makes his new software available (it includes
> George> a new hop bitter estimation scheme as well).
> Any chance that these will be made available for public consumption,
> Darryl? Those of us who write our own software or (shudder) calculate
> by hand can use all the help we can get.
George was far more generous with the credit than necessary. I
was rereading his article in the Fall 88 Zymurgy about measuring
beer color. In that article, there is a graph presented showing
the color of a Michelob Dark (MD) as it is gradually diluted
with distilled water. MD is known to be 17 SRM. The idea in
George's article is that by diluting MD until it matches the
color of your beer, you can determine your beer's actual color.
Sort of a homemade set of Lovibond cells.
Also, in the article, George comments on the common color
prediction algorithm of beer color = grain color * pounds of
grain / gallons of beer. He says that this works for very pale
beers in the 2-4 SRM range, but becomes dramatically incorrect
as the color gets darker.
My approach is to take the grist and determine how big a batch
would be required to produce a beer with 2 SRM. Then, take the
difference from this volume and the actual batch size and use
that to move back on the curve and predict a color.
I ran this on several batches of beer from my logs and it seemed
reasonable. Until, however, the beer is darker than MD. A stout
that I had made, which was very dark but not completely opaque,
came up at about 600 SRM. The fault here is entirely my own
since I naively extrapolated the curve in a straight line.
Obviously, the curve flattens out dramatically somewhere above
17 SRM. I haven't had the opportunity to look at the data
beyond 17 SRM, so I can only advise: Beware the dark side of
the curve, Luke.
The posts that comprise the Homebrew Digest Searchable Archive remain the
property of their authors.
This search system is copyright © 2008 Scott Alfter; all rights reserved.