Open AccessRegulating a distilled malt spirit fractioning process using Harrington’s desirability function
Author(s) -
В. А. Романов,
Nadezhda V. Barakova
Publication year - 2019
Publication title -
vestnik voronežskogo gosudarstvennogo universiteta inženernyh tehnologij
Language(s) - English
Resource type - Journals
eISSN - 2310-1202
pISSN - 2226-910X
DOI - 10.20914/2310-1202-2018-4-219-224
Subject(s) - fractionation , mathematics , function (biology) , sample (material) , distillation , fraction (chemistry) , process (computing) , quality (philosophy) , chromatography , organoleptic , process engineering , chemistry , computer science , food science , engineering , biology , philosophy , epistemology , evolutionary biology , operating system
In the study, a relation is found that eases the regulation of a distilled malt spirit fractionation process. The reasons are given for considering both the objective indicators (impurity content) and the organoleptic properties as quality criteria, which implies the need for a single quality criterion based on expert opinion and having a numerical expression. Using the Harrington desirability function method, such a criterion was found, tied to the model process of distilled malt spirit fractionation. The spirit was obtained by distilling twice on a Doctor Guber pot still a wash made from light barley malt and fermented with reactivated DistilaMax MW dry yeast for 70 hours at 24°C. The mode of distillation of the wash was constant, but during the second fractional distillation, the first sample of alcohol was divided into fractions according to the initial (reference) regime, and the next sample was fractionated in small batches, with a time step of 20 min, to estimate the dynamic uptake of by-products: aldehydes, esters, and higher alcohols. After impurity determination with gas chromatography, a generalized criterion was derived with the linear convolution formula from the by-product concentrations and the respective significance coefficients. The coefficients were found via a set of fuzzy rules constructed with the Harrington function method. The logic of collegial decisions regarding the first (standard) and the next (small batch) spirit samples was thus fixated in a simplified model, and the relation between stillage time and the generalized criterion was found, which provided feedback from the expert evaluated spirit sensory characteristics to the mode of fractioning. With the feedback loop, it is possible to model the logic of expert evaluation of the spirit and thus regulate the fractioning process without further need for expert evaluation.