top of page

Increasing Capping Accuracy of Bottling Lines, Niagara Bottling

Fall 2017

Bridge Inside

Overview and Objectives

Niagara Bottling is the industry leading private-label bottled water supplier in the Western United States. Our team worked to help find the potential causes of bottle cap misapplication in Niagara's bottle capping process. To determine how to reduce the number of misapplied caps, we conducted a Design of Experiments. Niagara has some of the fastest line speeds in the country making capping errors more likely to occur. Even though only 0.3 to 0.7% of all bottles have misapplied caps, at Niagara’s scale, even a small error rate means thousands of wasted bottles. For this analysis, bottle and cap design and major equipment changes were outside the scope of our project. 

​

Design Approach

Our approach was to establish potential operational settings, and then find the combination of settings that resulted in the lowest misapplication of caps. As is traditional for Juniors participating in HMC clinic programs, my participation was for only the first term of a two term project. I focused primarily on variable identification, definition, and concision. The first step, involved visiting the manufacturing facilities to collect data on the bottles where the cap was misapplied. We gathered data from the bottle capper monitoring systems, with a focus on the relative frequencies of different types of capping defects, understanding the distribution of bottle opening torques, and exploring image processing to better understand and classify defective bottles.

 

Using data processing techniques we evaluated a year’s worth of capping data, and were able to quickly find the distribution of types of cap misapplication. Further, we were able to create a distribution of bottle cap opening torques by taking random bottle samples and putting them into a machine that calculated the torque required to open a bottle. This torque was assumed to be the same as the torque required to apply the caps to the bottles. Lastly, we used a collection of images of bottles with misaligned caps, to find the average angle to which the caps were misapplied. To do this, we used an image processing software that found the top line of the caps and compared it to the angle of the picture.

 

​

​

​

​

​

​

​

​

​

​

​

Results

After investigating all possible causes of cap misapplication, we determined that there were three with the largest potential impact, including: 

  1. The type of small knife that holds the neck of the bottle when the cap is being applied. There are multiple types of knives used throughout their facilities and we thought that this, along with the age and sharpness could have an impact on bottle capping. As supporting evidence, we looked at which set of knives provided the most overall bottle stability.

  2. The plastic build up inside the chucks that hold the bottle caps. Each cap fits tightly inside the chuck, and it’s possible that small pieces of plastic could come off and build up inside the chuck.

  3. The capping torque of the cap onto the bottle. To set or change the torque of a given chuck, a mechanic must go in and manually tighten or loosen the torque. This results in very unprecise and inconsistent torques throughout the 30 chucks within each bottling line.

We started doing single variable tests of these three factors to get a sense of what settings gave good results and created a test plan for the DOE in the spring. 

Skills

Excel data processing, python image processing package, Minitab DOE analysis.

bottom of page