How much is enough?

Let’s get started with a mill “up north”. Like most mills that start automating laboratory testing the sound of more data is a primal call to the statistician in us all, nah…well maybe a few; HOWEVER, when automated test systems are implemented, part of their valid justification is more data. The question then comes, how much data is enough? Let’s just say 9 out of 10 customers fall in to the camp of being data hungry.

The more = The better

One problem, TIME. What is the cost in operational machine TIME to get that data. Let’s take the example of a 240” trim width for a machine. If I chose to test every 12” I get loads of data; approximately 20 measurements/properties and the 100's of resulting data points. Getting that data will cost me time, paper machine time and that cost big money; maybe there is a better approach.

Back to our customer up north. They wanted to maximize the data coming from their new PROFILE/Plus automated test system. They wanted it to take the most test and generate the most data, so we set it up to give them all they wanted. After about 2 months they realized that this option, while sounding good at first, wasn’t what they really needed. Like 90% of those customers before them, we sat down with them and helped them analyze their data and how they used the data. We could quickly see areas where the relevance of that much data was not beneficial. After discussions and much data analysis we agreed on some changes. Ultimately those changes cut the test data set in half, which in-turn got the data in to the mill DCS earlier than the old way and now decisions could be made sooner. Getting data minutes earlier is big, real big for a paper machine running at the speeds it does. Within one month of running under this new testing and data collection protocol the mill had reduced customer complaints/rejects by 30%.....that’ll make any mill manager happy. Optimizing your processes also includes determining how much data is enough, the most is not always the best.

Change will happen, the question is, are you ready?

36 views0 comments