Site icon Visionify

Leakage inspection in the beverage industry – Case Study

Background

The beverage industry is a primary contributor to global economy and it is evolving rapidly. New products are developed, new consumer tastes emerge and consumer behavior evolves. This directly impacts production of beverages and equipment used for making them.

Over a million tonnes of beer is produced worldwide each year. However, when it comes to checking quality many major manufacturers still rely on humans to manually inspect their products. In fact, while beverage industry is quite advanced in terms of analysis and sensors used during production, vision systems are surprisingly scarce in use.

In beverage industry, every manufacturer measures its liquids in metal containers. These containers are known as kegs and stores beer, wine, soda, milk, and carbonated drinks. 

Challenges

Keg-level inspection is critical during filling and packaging processes since it determines packaged beer quality. The challenge is detecting whether a keg has a beer leak and dents scratch after filling a beverage.

Solutions

Avoiding leakage in kegs is an essential goal for beverage industries. Our vision systems are used for automatic inspection of beer kegs. Leak testing is typically performed online with a vision-based leak test system to indirectly measure internal pressure during filling or bottling process. The system inspects exterior and interior of a keg, locating defects such as dents, scratches, leaks, smudges, bubbles and dirt.

Our machine vision inspection technology is intended explicitly for moist conditions, such as interior of draught beer kegs. Due to its quick cycle time it can inspect up to 1,400 kegs per hour which equates to fittings being checked in less than 3-second cycles. Leakage is detected using our technology which particularly senses leaking foam.

Minor fitting surface and condensate damage might not always result in a tight keg getting rejected. Therefore, leakage inspection is customized for foam structures. Rejection rates are typically around 0.2 percent whereas error rates are at 0.05 percent. A particular ratio between foam and structure can be achieved by capturing successive images with varying parameters.

We have installed cameras in a light protective tunnel to enable high-quality images. This will aid in maintaining image accuracy. In addition our system identifies defects and focuses on real issues with an ultrasonic bath (a conveyor system passing through a water-filled tub activated by ultrasonic activators). Our solution was one of most successful error detection methods in this application with less than 0.2 percent beta risks.

You can save error statistics from getting a sum of leaking/tight ones on every batch of kegs. The accuracy of the fittings used will be continuously documented.

Exit mobile version