Program Book - Master Brewers Association of the Americas

P-170
Controlling fills in the brewing industry: Does Hot water
jetting make a difference?
RUTH DUFFY-KRYWICKI (1)
(1) Miller Brewing Company, Albany, GA
A “DOE” method of investigation was conducted to determine if hot
water jetting vs. ambient water jetting was effective at controlling
fills on a high-speed bottling line process. There is only anecdotal
evidence that suggests that better fill control can be derived from hot
water jetting. Optimizing the package filling step is of considerable
interest because of regulatory compliance as well as the potential
for beer loss. This particular study was performed on a bottle line
which had hot water jetting capabilities, but routinely used ambient
temperature jetter settings because of mineral deposits which
buildup in the jetter nozzle when using hot water. Furthermore,
several different products with varying alcohol strengths were
bottled on this line, which prevented the comparison of data simply
by using long term averages. The bottle line studied demonstrated
good performance for fill height and fill volume, but did not have
the capability to sample by valve. Therefore, a screening experiment
was devised to select four adjacent valves that demonstrated
reproducibility. The data from the valve screening experiment was
used to define the appropriate sample size to avoid the possibility of
Type II error. A randomized complete block design (RCBD) type of
experiment was performed where each valve was a “block” tested
at two different levels (jetter temperature). The experimental set up
required controlling at least nine parameters which could influence
the overall data and “confound” the results. The samples from each
of the four valves were measured for fill height using an Akitek fill
height measurement apparatus, and then gravimetrically assessed
for fill volume using the appropriate conversion factors. The results
were analyzed for ANOVA using the StatGraphics 5.0 statistical
software package. The data indicate that there was no statistically
significant difference in either fill height or fill volume for samples
that were hot water jetted compared to those which were ambient
water jetted. Furthermore, the data suggest that the RCBD method
of experimental set-up where the valve to valve variability was
blocked was prudent.
Ruth H. Duffy-Krywicki earned a B.A. degree in chemistry from
Immaculata University and a Ph.D. degree in synthetic organic
chemistry from Bryn Mawr College, both of which are located in
the Philadelphia, PA, area. She recently completed a M.S. degree in
industrial/quality engineering from Lehigh University, Bethlehem,
PA. Duffy-Krywicki has over 16 years of combined experience in
the food additive, pharmaceutical, and brewing industries. She
is currently the packaging laboratory manager at Miller Brewing
Company’s Albany, GA, brewery, where she focuses on supporting
corporate world-class manufacturing initiatives, utilizing six sigma
DMAIC methodologies.
P-171
Development of a hybrid system for automatic recognition of
particulate foreign matter in filled food on the basis of multicontact
excitation
ANDREAS KASPRZYK (1), Judith Forstner (1), Rainer Benning (1),
Sascha Bach (2), Jens Peter Majschak (3), Antonio Delgado (1)
(1) Institute of Fluid Mechanics of the University Erlangen
Nuremberg, Erlangen, Germany; (2) Fraunhofer-Applications
Center for Processing Machines and Packaging Technology
AVV, Dresden, Germany; (3) Institute of Processing Machines,
Agricultural Machines and Processing Technology of the Technical
University of Dresden, Dresden, Germany
150
The presence of foreign matter in containers filled with food
represents an extremely significant problem for producers and
bottlers as well as suppliers and trade in the relevant industrial
branches. Apart from image damage, the risks that arise from
product liability—especially damage to persons—and consequential
recourse claims have to be considered, as well as possible refusal of
retailers to sell the products. The presented project focuses upon
solid particulate foreign matter that cannot be handled by the usual
optical detection systems. Presented are especially those cases
where pieces of broken glass, here with characteristic dimensions
>1 mm, in a glass container represent a high risk of causing injuries
to the consumer. Although the emphasis is put upon this specific
application, the system can also be used for containers of arbitrary
optical accessibility and various materials, e.g. PET bottles, and a
multitude of foreign particle materials, including metal splinters.
All of these cases are handled with a solution approach, where the
diagnosis of the existence of a solid particulate foreign body in food
that is sufficiently capable of flowing, e.g. beer, juices, yogurt, is
carried out on the basis of the vibrational response of the system
food-package-foreign body to mechanical excitation by means
of a neuronumerical hybrid. This system consists of numerical
simulations and artificial neural network (ANN). Before excitation
the particle is positioned near the wall of the container by an
accelerated movement. The registration of the contact between a
foreign particle and the package is realized optically and by piezo
principle. Assignment of the response signal into the classes
“particle detected” or “no particle detected” is done by the ANN.
Numerical simulations on one hand are used for training of the ANN
by producing a sufficient amount of training data. On the other
hand they build the basis in the design of the experimental process
parameters by estimating the impact of the transport induced flow
upon the behavior of the particle and by simulating the reaction of
the particle to the induced oscillation of the wall of the package. An
important goal is the integration of the system into existing filling
equipment, taking into account limiting parameters, e.g. cycle
times, and various methods of vibrational decoupling. Additionally,
different acceleration and excitation parameters are systematically
investigated. The presented work was conducted with cooperation
between the Institute of Fluid Mechanics, the University Erlangen-
Nuremberg, and the Fraunhofer Applications Center for Processing
Machinery and Packaging Technology (Fraunhofer AVV).
From 1994 to 1997 Andreas Kasprzyk apprenticed as a brewer
and maltster at the Paulaner Brewery GmbH & Co KG in Munich.
Afterward he was employed at the Spaten-Franziskaner-Bräu
GmbH as a brewer. In 2001 he began his studies on brewing and
beverage technology at the Technical University of Munich (TUM)
in Weihenstephan. He completed his Dipl.-Ing. (Univ.) degree in
2006. After graduation he began employment with Versuchs- und
Lehranstalt für Brauerei in Berlin e. V. as a scientific assistant at the
Research Institute for Engineering and Packaging (FMV). In 2007
he moved to the University Erlangen-Nuremberg (FAU), Institute
for Fluid Mechanics (LSTM). There he is working on a Ph.D. on
“Damage Detection of Returnable Goods” in the group process
automation of flows in bio- and medical technology.