Controlling data variation

By Daphne Allen in Testing on March 31, 2015

When testing package seals according to ASTM F2054 (Standard Test Method for Burst Testing of Flexible Package Seals Using Internal Air Pressurization Within Restraining Plates), Thomas McNabb and his team of packaging engineers at Edwards Lifesciences would often see excessive variation in the results data as well as nonnormal data and inconsistent results over time. He decided to investigate to determine whether their testing approach could be improved to reduce variation.

McNabb, who serves as a principal packaging engineer for Edwards, spoke at HealthPack 2015 during the presentation “Burst Testing: Finding Better Value in the Data.”

McNabb explained that his team uses burst testing to evaluate seal strength during IQ, OQ, PQ as well as during in-process testing. He told the audience that if correlations were to be made between tests and testing equipment, the parameters of each test must be equivalent.

First, for this particular exercise his team held all of the factors constant: package size, material, rate of air flow into the package, pressure detection sensing mechanism, sensitivity (machine response to pressure drop), and position of test article. “The only variable was gap,” he says. “We did investigate those factors in different studies, but did not find the same sensitivity or impact on variability.”

Gap settings were then evaluated to determine whether differences had an effect on mean output value, McNabb explained. Experimenting with the gap, his team found that “gap settings do affect the value of the burst output for the same package,” he said. “As the gap increases, the burst value decreases.”

Comparing the results from testing different packages, the team found that “as the gap changes, data variation will initially increase, but the variation will eventually reach a minimum,” McNabb said.

As a result, he believes that “reduced variation leads to more accurate measurements.” He advises other users of ASTM F2054 to “keep all variables constant, including the gap.”

In addition to keeping the gap constant when measuring a particular package configuration, “the recommendation is to follow the ASTM guidance to evaluate gaps for suitability,” he adds.

McNabb’s team will next try to correlate the approach to peel testing following F88 (Standard Test Method for Seal Strength of Flexible Barrier Materials). “The idea is to improve all test methods and to get a better window into controlling data variation,” he said.

By submitting this form, you accept the Mollom privacy policy.
500 characters remaining