Using Vision Inspection for Production Process Adjustments
Consumers expect consistency from snack foods, baked goods, breaded meat products, and other processed foods. Many of these products enter a complex series of automated steps where machines interact with the product throughout the process.
For example, many processed food products have ingredients applied outside the product, including seeds, chocolate chips, glazes, colored seasonings, and others. Additionally, some products may have branded designs, logos, and other cosmetic features stamped onto the product at certain stages of their process.
Food processing machines are calibrated and frequently checked to ensure they are performing their specific application consistently. Overtime and for various reasons, these machines can clog, run out of alignment, or over-apply ingredients, resulting in an out-of-spec product.
In food production operations that produce thousands of products daily, many rely on manual product inspection before packaging. However, since manual inspection is highly subjective and operator-dependent, it may be challenging to detect the warning signs of an issue in the production process. It usually is not until a production run is too far to notice these variations and adjust, leading to wasted product or worse.
A Seemingly Minor Insignificance Prompts a Process Malfunction
One cupcake producer experienced how a little visual defect could signal a critical problem in the production process. After baking, each cupcake receives a layer of chocolate frosting over the top, followed by a series of icing swirls applied by a mechanical icing applicator. The position and thickness of the icing is important for brand standards.
As part of the company’s quality control protocol, once every hour of their shift, a quality assurance specialist is required to hand-pick 10 cupcakes from the production line to inspect for product specifications. Along with the size, shape, and color of the overall cupcake, they must also examine the position, sequence, and thickness of the icing swirl.
On this day, one QA operator failed to recognize that the icing swirl began to display subtle inconsistencies from one hour to the next. He ensured that the number of swirls on each cupcake was the proper amount but did not notice that the applied icing was slightly thinner from one hour to the next. The swirls also began to move closer to one side of the cupcake instead of down the middle.
 |
 |
 |
|---|---|---|
| Hour 1 Icing is symmetrical, centered, and at the ideal thickness |
Hour 3
Icing is slightly misaligned and thin in spots. Some spattering from the malfunctioning applicator appears. |
Hour 5
Icing is thin and inconsistent, missing coverage due to misalignment. Even more spattering appears. |
While the production issue begins minor, without an objective method to detect these variations, the cupcake producer gradually found themselves in a quality crisis.
Furthermore, upon completing his shift for the day, he neglected to share these findings with the next shift’s QA operator.
After only minutes in the next shift, the icing applicator began running completely out of spec, leading to a total shutdown of production and several wasted cupcakes and ingredients. Because the previous QA operator was not keen enough to notice these gradual subtleties from unit to unit, the result was a critical error that could have been avoided with a simple mechanical adjustment on the machine.
Sights Set on a Better Way to Proactively Detect Difficult-to-Measure Production Flaws
The company’s search led to the discovery of KPM Analytics’ Vision Process Control solutions for baked products. Vision Process Control solutions are installed at key production stages where data visualization and direct feedback integration can automatically act to keep process performance on target. Each system is custom programmed to analyze products at full-line speeds for size, shape, color, as well as more detailed features, like the cupcake swirl.
Vision Process Control helps enable advanced defect detection applications and provides feedback to operators to adjust machinery if products begin to drift out of specifications.
Since the icing appearance is such an important feature of their cupcake, with the help of KPM Analytics, the company programmed a custom in-line vision inspection application to analyze products after each lane of the icing applicator to ensure ideal swirl placement, sequence, and thickness. With its quality assurance team armed with this objective data, the company has improved quality and reduced waste while enabling the production team to make informed decisions about their process.
Are you interested in learning more Vision Process Control technologies for baked goods? Contact us today!
