Mathematical modeling of beer foam.

Melm, G., Tung, P. and Pringle, A.

Abstract
The foam stability of 49 different bottled beer brands was measured, using an automatic pourer to simulate normal dispensing, and the relationships between foam stability and the original gravity, real degree of fermentation, alcohol by volume, bitterness units, high molecular weight protein content, real extract, finished beer specific gravity, colour, iso alpha acid content and pH of the same beer were evaluated using three different procedures (single factor linear correlations, multiple regression analysis and neural network modelling). The bitterness unit rating was found to be the most important single factor, while in the multiple regression and neural network mathematical models, bitterness, pH and real extract combined to account for 87% of variation in foam stability (although pH, the second most important factor in both models, showed very little correlation with foam stability when considered as a single factor).
Keywords : beer foam stability model simulation statistical analysis