Premium brands are those defined as having the highest perceived quality amongst peers, well-branded products that often are more expensive than competitive offerings. In the past few years, these companies have undergone a series of rapid, costly adjustments: frequent changes in ownership, relocation of facilities, and an increase of over 30% in flavor variety. Consequently, proving ROI has become an increasingly important requirement for capital expenditures.

Two recent Automation Associates, Inc. (AAI) clients, Beringer Blass Winery Estates and Tropicana Beverages Incorporated, in need of adapting their businesses to address these issues head on, have found that simulation modeling can be a highly effective management tool to solve their specific operational challenges.

Simulation models are built to address complex problems within an organization.

• Tropicana, for instance, faces severe competition created by the low-carb/low-sugar craze and the sheer volume of alternative choices that have emerged as a result. The availability of more beverage options and finer consumer segmentation redefined the company's product demand profile. Smaller purchased product volumes with a higher SKU count and a limited shelf life meant the company must have smaller, more frequent blending and bottling runs. The increased blending and bottling run frequency is more costly and time consuming due to the need for additional cleaning and sterilization of blending and bottling systems but offset by the cost of refrigerating finished product.
• Beringer has enormous pricing pressure due to a bumper crop of grapes and better, cheaper bottling alternatives. This is exacerbated by the availability of low-priced wine from specialty retailers such as Trader Joe's, who are conditioning consumers to give preference to low cost labels. In an effort to streamline its operations, the company evaluated the design and integration of a new bottling plant into the overall value stream of its operations. The shared use of blending and bottling system resources was analyzed to determine the blending and bottling equipment and systems required to support the operational mission.

Beginning model development requires a thorough understanding of a company's facility and processes. Inside a beverage production and bottling facility, the flow of material is usually separated into extraction, blending and bottling areas, and generally moves in semi-continuous and discrete batches through numerous combinations of blending, addition, and filtrations prior to bottling or packaging for consumer purchase. A number of variables, including FDA requirements, blending ingredient additions, cleaning and sterilization maintenance activities, packaging requirements and scheduling for the arrival of raw goods need to be well understood in order to define where problems and inefficiencies or constraints within the current processes and equipment may be present.

Four types of decision support models have been created to aid companies in replicating their current extraction, blending and bottling processes. These include:

• data model
• scheduling model
• process model
• cost elimination model

A data model is usually the start of a simulation modeling engagement and begins with information gathering. Many of today's businesses have eliminated generational apprenticeship methods of passing down knowledge on how various processes work. This collective knowledge is then reacquired by developing simulation projects, which facilitate discussions among different functional areas and can be shared among all groups. Data models have proven to be great tools for new employee orientations, as their content generally consist of a series of flowcharts and operational parameters describing the operational flows and velocities.

A scheduling model utilizes details from the data model to create the set of schedules necessary to drive the extraction, blending, bottling, and packaging operations. The packaging schedule defines the specific daily sequence for products to be manufactured on each of the filling lines. A scheduling utility supports the consideration of the line fill rate, product changeovers and maintenance events, and packaging quantity to create a schedule for each line. The packaging schedule is then aggregated by final product and across multiple fill lines and runs to create a blending schedule. The blending schedule considers the mixing steps, recipes and resources required, while creating a specific daily sequence to create each blend. Finally, the blending schedule is aggregated by raw material received or extracted to define when these components should be introduced into the facility. Each of these schedules then generates activities for the process model to attempt to execute.

A bird's eye view of current processes can be an invaluable tool for operations managers. The process model uses simulation software to replicate the facility's current and proposed operations. The data model acts as a user interface to uniquely represent the facility exactly as it exists, yet allows for client modification to represent changes to the existing equipment and processes without needing to change the process model logic. Process models can animate the time sequence status of the operations, while providing measurements for all complex interactions within the beverage production facility. Strategic and tactical ideas are defined in the data and scheduling models and executed as scenarios with the process model. Typical results of the process model include equipment utilization, blending delays and cause, and cleaning event implications.

Bearing in mind the intense focus on preserving bottom line revenue, a cost model is increasingly being paired with the process model results to form a quantifiable basis for decision-making. The cost model uses the results of numerous scenarios run with the process model in order to compare the total cost required to implement that scenario's strategy. This approach supports a detailed definition of all operational costs associated with increased assets required, avoidance of new assets and compliance to target schedules. For the beverage industry, typical scenarios determine how many pieces of equipment (such as tanks of each size) are actually being utilized, in order to further eliminate wasteful assets or identify critical resources required.

The simulation models AAI developed with both Beringer and Tropicana confirmed operational improvement opportunities at certain processes within their facilities. The models also helped the two companies better understand their own processes, and provided an effective tool to quantifiably evaluate proposed capital expenditures and scheduling changes over short and long-term periods. As a result of the three simulation models developed by AAI for Tropicana's facilities, the company has enjoyed multiple cost-avoidance benefits and continues to use simulation modeling as a means of perfecting its processes. Tropicana currently uses the models to evaluate proposed blending and packaging changes, and has used the models to broaden the process knowledge of employees assuming new operational responsibilities. Beringer, meanwhile, used AAI's simulation modeling scenarios to identify the potential to eliminate or replace35-40 tanks priced between $13,000 to $ 960,000 each. This resulted in a tremendous reduction in the capital budget of a new bottling facility, while increasing confidence in the desired operations results and return on investment.