Fermentation is a biochemical processes used in food production, alternative protein sources,biopharmaceuticals, biofuels and waste management, to name a few. One of the oldest fermented
products produced by humans is wine, made from grapes. Today, while wine is produced at a global
scale, the rate of adoption of new technology lags compared to other mass produced products.
Unlike most other fermentations of commercial products, wine fermentations are seasonal in nature
and occur within a six- to eight-week period, only once each year. The most valuable wines come
from small batches of grapes, and this usually means that grape loads from different vineyards are
kept separate and fermented in small volumes, leading to a wide variation in nutritional components
and the resulting fermentation patterns. Increasing constraints of water, labor, energy along with
lower crop yield due to changing climates motivates innovation in the winemaking digitization.
This dissertation presents advances in two components of wine fermentation: the measurement,modeling and control of fermentation kinetics, and the measurement and control of redox potential during fermentation. In the first portion of this dissertation, a mathematical description of
fermentation kinetics from 1979 is updated to accept changes in temperature and combined with
a new nonlinear parameter estimation method implemented in the cloud. Combined with sensors
that make measurements in regular time intervals, these advances are now enabling commercial
wineries to diagnose fermentations and perform informed interventions during fermentation. The
measurement-modeling technique is then extended to implement a nonlinear model predictive controller of fermentation kinetics.
The second portion of this dissertation focuses on the redox potential of the fermenting mixture,a measurement that is linked to the electrochemical state outside the yeast cells and provides insight
into yeast metabolism. As an under-studied variable in wine research, measurement and control
of redox potential as a process parameter has not previously been demonstrated in commercial
scales. Controlling redox potential during fermentation can prevent the formation of undesirable
byproducts and has clearly been shown to influence and direct yeast metabolism. This work
demonstrates the control of redox potential at research, pilot and commercial volumes while also
exploring alternative methods for controlling redox potential that is independent of the variations
in juice chemistry.
Combined, the two focuses of this work provide the basis for commercially adoptable processcontrol strategies for commercial wineries that also have application to other fermentation industries