Are you using a high powered SPC operating system to analyze your manufacturing process and product data?
Process Monitor (PM) has a long-term track record of eliminating production variability and saving manufacturers millions of dollars a year.
“The core principle of the latent defect theory is that variation in manufacturing processes is the main culprit for defects, and eliminating variation will help eliminate defects, which will in turn eliminate the wastes associated with defects, saving money and increasing customer satisfaction.”
Bill Smith, as cited in “A History of Six Sigma”
We all know that analyzing manufacturing data can be time consuming and, if not done right, it can be counterproductive. Tight control chart limits result in excessive or nuisance alarms. Relying on trend charts for visualization of tens, hundreds or thousands of process parameters can make you feel like “I know something important is happening in that data, but I don’t have enough hours today to find the needle in that haystack”.
What we need is a software tool that can evaluate terabytes of data to efficiently display statistically abnormal information, that does not require excessive setup or maintenance time, and does not become a nuisance because of excessive alarms.
Process Monitor (PM) is a software tool, created and optimized within the Eastman Kodak Company over the last 35 years, that has been implemented in the following production environments, on five continents!
- Polymer Manufacturing
- Paper and Polymer Web Manufacturing
- Precision Chemical Making, Filtration & Delivery
- Web Coating, Conveyance & Drying
- High speed Digital Printing
- Gelatin Manufacturing & Drying
- Nanoparticle Precipitation
- Steel Casting
- Steam and Electricity Production
- Chilled Water & Brine Production
In case you were not aware, for many years Kodak has used curtain coating of up to 13 layers of light sensitive emulsions, sub & topcoats and inter-layers in order to create the photographic papers and films you have observed in theaters and in your parents and grandparents photo albums. Imagine the complexity and precision involved with curtain coating of three: slow, medium and fast speed yellow emulsions out of a single die, all with uniform thickness across a 48” width onto a web that is moving at over 1000 ft per minute! Send that web though a drying section to remove nearly all the moisture before simultaneously applying three more layers of cyan emulsions, and then repeat the process again for three layers of magenta emulsions. Then, wind the 4-foot web onto a roll to a length of over a mile before a new roll begins, about five minutes later. Also note that the entire operation must happen in the dark, because light will render the product useless.
Over a thousand parameters are measured while making; and evaluated, after producing each of the rolls described above.
Its hard for me to imagine running an operation like this without using a high level SPC operating system to let me know about abnormalities before things go out of control. Imagine the cost of making one, or two, or ten of the photographically sensitive mile-long rolls before knowing that they have to be trashed because a delivery pump on one of the thirteen layers is starting to act up.
Remember! 10 separate mile-long rolls are made in less than one hour!
I understand that the web conveyance, coating, drying and roll making system described above is probably more demanding that what you are responsible for, but understand that you can use the same Kodak-created SPC operating system to evaluate your operation and help you to maximize the quality of your products while saving you money at the same time.
Here are some of the things you can do with Process Monitor (PM):
- Use Calculator to create time-series key statistical features:
- Mean/Median, Std/MAD, COV (& many others)
- Use Functional Analysis Based (FAB) Macros to calculate features for all tags and orders
- Inspect Univariate (UV) order based SPC charts
- Discover Normal vs Abnormal activity with Auto-Limits
- Follow Corrective-Action-Guidelines (CAGS) to determine root cause of variability
- Document actions and findings in Order Notes
- Build Multivariate (MV) / Machine Learning (ML) models
- Inspect MV/ML charts
- View 2D/3D Scatter Plots
- Drill down through contribution charts to time series
Below are links to some videos that show what PM looks like and display some of its functionality. If you are interested in using a successfully proven software tool to drive out variability in the products you make, take a few minutes to learn more about Process Monitor.
Be sure to observe the 3D, Multivariate charts that display machine learning information in a new, illustrative way. The 3D charts could possibly make you dizzy, but I think it’s worth the risk.