The Quality Blog

I have been working on standardizing certain assembly process FMEA functions and failure modes. These Functions are verb-based, like fasten, align, locate, grasp etc…In doing this, there are certain synonyms that also can be grouped.

The new part about this is the standard control method selection based on criticality. The controls are developed around 4 levels of criticality. (criticality is the severity and occurrence combination) The highest criticality is associated with safety and regulatory conformance to the lowest where severity is below 5 and occurrences are also low.

By choosing certain practices as controls and based on criticality, the process controls (both prevention and detection) are already known and therefore just put in place on the FMEA and transferred to the control plan directly. This makes the 3rd path of FMEA development very fast but highly correlated to the FMEA /Control plan activities.

Grouping of Failure Modes by verbs can provide the basis for consistent FMEA development.

Just a quick not to share some recent experience. I was at several clients over the last two months, each with an identical issue. These clients are in different industries, aerospace, automotive, and pharmaceuticals. The issue was legacy, or lack thereof. Each company has been wildly successful over the years, (two are over 100 years old) but when it came to Product Development, a great deal of waste was present due to relearning that which has already been learned. This is partly their customer’s fault, by not accepting years of data on identical product, but nonetheless, testing and making prototypes to test at reliability and confidence levels requiring 6 or more samples, is very expensive and time consuming.

My recommendation to them was to develop a legacy based product development system. This means that all that we already know is easily retrievable and used to reduce overall time in PD. The risk is low and the confidence in these designs must be high with a good deal of history. It may be that the concept of confidence is misunderstood. The more I see, experience, etc… the more confident I am. Statistical confidence without engineering judgment and legacy will drive unnecessary testing and longer PD to market.