Production supervisors can build comparisons, assessing how batches are running at each site and comparing that information to other locations or to corporate standards. In this way, a plant can standardize the processes of a particular product at the optimal recipe and performance level.
You Might Also Like
Explore This IssueApril/May 2010
Also By This Author
This same manufacturer can apply manufacturing intelligence to determine, for example, why one machine requires more time to dry a product. By looking at multiple data sources—from the humidity in the air to changes in raw material—and using the advanced analytics to determine correlation, the manufacturer calculates and compares process trends over time. The manufacturer uses performance equations to derive information that was not obvious, helping the company determine the root cause of a problem.
Step Two: Define and Enforce Your Quality Process
Step two in improving food quality is ensuring that the product is repeatedly made with the same quality process. Using the correlated information gathered during step one, you can make improvements in the overall production process and establish best practices.
Once the best practice is defined, its use must be enforced so that product is created the same way every time. A procedural control application will ensure the use of best practices related to the raw materials, the processing equipment, and the manual and automatic operational procedures.
One food manufacturer applied a procedural control application to create manual spice kits. This control procedure combined manual operations and automatic weighing equipment to reduce quality-related expenses by more than 10% by enforcing specific requirements for ingredient type and amount, kit processing, and proper identification and tracking of the completed spice kits.
Step Three: Apply Predictive Quality
The third step for improving food quality involves applying predictive quality. Predictive control technologies use advanced modeling techniques based on timely in-process measurements to simulate processes, run what-if scenarios, and determine how changes will impact output. Setting output variable targets and letting the models determine optimum input targets can also perform steady-state optimizations.
One of the world’s leading dairy producers wanted to increase throughput to handle growing raw milk supplies. The company looked for ways to improve the operating efficiency of existing evaporators and dryers before making new capital investments. The wealth of historical data for the wide product mix produced in the dryer allowed for creation of a model that accurately reflected moisture ranges for each product. Installing an application that biased the prediction hourly with in-process testing data from the online grading analysis system enhanced the simulation.
Audit results revealed that the predictive control solution exceeded its objective, reducing the variation in the chamber outlet temperature by approximately 43% and the sifter moisture by 52%. The dairy producer successfully increased the yield and quality of its nutritional, whole milk, and skim milk powders.
Bringing It All Together
Food manufacturers increasingly see the inherent value and tangible returns of integrated quality management systems. The information derived not only empowers employees with improved vision, but also helps them correlate and trend data and achieve marked improvements through prediction and control—all of which help reduce risk and improve profitability. ■