The shape of things to come: ECHIP, Inc. is producing tomorrow's software for today's brewers

Modern Brewery Age, Jan 21, 1991

The Shape Of Things To Come With brewery automation advancing by leaps and bounds, computer technology has become an integral part of modern brewery operations. Computers serve a particularly valuable role in design of experiments at major breweries across the country.

According to ECHIP, Inc., a manufacturer of software for engineers and scientists, the design of experiments is one of the "best kept secrets for cost reduction and consistent quality in beer production." The company ranks Anheuser-Busch, Miller and Coors among their customers.

Although brewmasters of yore might goggle at contemporary brewing technology, ECHIP points out its necessity in today's competitive beer market.

Duane Melek, senior research statistical analyst with Miller Brewing Co., uses computer-aided design for experiments (CADOE) to optimize output. He identifies levels for such inputs as time, temperature and ingredients that make the best beer at the lowest cost.

Design of experiments is a tool that allows engineers and scientists to gather the most information in the fewest experiments. The results are analyzed by computer and provide reliable predictive knowledge in easy-to-read "maps" of the process. Brewers can understand exactly how a change in input will translate into a change in product characteristics. It's a "much more efficient way of determining an optimum mix of ingredients than traditional one-variable-at-at-time methods," says Melek. "It saves us a lot of time and effort.

From fermentation conditions and ingredient mixture selection to fill level determination, CADOE helps brewers efficiently identify the best settings for each stage of the production and packaging process. CADOE is quickly becoming an essential tool for brewing scientists and engineers.

Chief among the benefits of design of experiments, or DOE, is the ability to cut time and resources needed for the introduction of more efficient ways of manufacturing. Laboratory discoveries are quickly incorporated in the production environment, leading to advantage in the competitive beverage market.

For example, scientists at the Coors Brewing Co. wanted to understand various fillers' performance at any given operating condition. Hugo Patino, manager of fermenting and packaging research and development, set out with his team to determine if filling valves performed significantly differently from one another and by how much. An understanding of the effect of filler settings on fill and oxygen levels was sought.

Patino's team opted to use computer-aided design of experiments (CADOE) to fully characterize their filling process, winning them the Master Brewers' Association Presidential Award. The response surface methods used in ECHIP software allow the experimenter to vary a number of inputs at once, using computer-determined levels for each experiment. The number of trials that need to be made are significantly reduced, saving both time and money. This is of particular importance in experiments that take weeks to perform, such as in the colloidal stability area.

The collected data may then be quickly analyzed by computer. In this case, the resulting three-dimensional "response surface" produced by ECHIP [Fig. 1] predicts how the best trade-off of settings for the oxygen, fill and beer level loss will result from changes in the filler settings. It is easy to locate the "peak" on the response surface that corresponds to optimal settings.

A corresponding two-dimensional "contour plot" shows the projections of the 3-D surface [Fig.2]. By positioning the cursor at the optimum, the settings for inputs that produce this best case are displayed. The production line may then immediately be adjusted for peak efficiency.

When the number of possible input variables is large, a "screening design" may be run. ECHIP prescribes the experiments to be run, and then identifies which of the many variables that are important, and which are not. The technical professional is then free to design and run response surface experiments which concentrate only on those important factors and their interactions. Patino selected this method in his study.

Response surface methods are applicable in other aspects of brewing as well. For example, the effect of mash separation time and other operating conditions on beer flavor may be efficiently characterized. Optimal levels of ingredients in each successive stage of brewing may be determined. ECHIP has special designs for experiments in which the proportions of ingredients are to be studied. The resulting graphics have three axea instead of two [Figures 3a & 3b].

Greg Casey, of Anheuser-Busch's corporate research and development, used Cadoe to determine the influence of wort amino acid levels on vicinal diketone (VDK) production in lager fermentations. He, too, used screening designs to determine which of nine variables are important. Five of his inputs proved to have significant effects on VDK production during the primary fermentation. Further experiments were run to optimize the levels of each of these amino acids that would minimize VDK production.