| QG16: Measurability (Metric 1 of 7) |
|
Measurability is the beginning of knowledge building in CPD&M. If you cannot measure your NUD functions and part/material characteristics that integrate to enable them, you are not going to be able to conduct the tasks of CPD&M. Measurability is the basis for the whole method.
.jpg) | | Minitab Gage R&R Menu |
Measurement System Analysis includes all elements of variation that can characterize the risk present in trusting a data acquisition system. Conducting a Gage Repeatability and Reproducibility (GR&R) Analysis is the dominant method to prove the data you are taking has more signal than noise in it and that you can adequately discriminate between the two. It is advisable that every NUD functional response have a documented GR&R Study conducted on it and that the development team can display the output chart seen above to prove they know how capable their measurement systems are. This chart is from Minitab's Gage R&R menu for continuous variable data gathered in a "crossed" DOE format for non-destructive or destructive testing of samples. A Gage R&R Study is a designed experiment (DOE) conducted on a measurement system which includes the measured samples, people, environment as well as the instrumentation. ANOVA statistical analysis is used to decompose the components of variation within a measurement system.
The table below illustrates a tiered structure of % Knowledge-In-Hand to rate the risk a NUD Function or Characteristic is carrying due to the current level of measurability.
|
How to Assign % Knowledge-in-Hand for MEASURABILITY
| |
100%
|
I have GR&R at approx. 10% to prove I have very good signal-to-noise control of the measurement system. I have a control plan in place to maintain approx. 10% GR&R.
| |
90%
|
I have GR&R at approx. 30% to prove I can measure the function within reason. I can certify the measurement system and have a plan to lower the GR&R.
| |
80%
|
I can routinely measure functions using continuous variable data and am able to conduct Measurement Systems Analysis using Gage R&R studies.
| |
70%
|
I have prototype or COTS data acquisition hardware and software under active development/evaluation. I have not certified the measurement system.
| |
60%
|
I have one or more ideas on how to measure direct functions using continuous variable data (including candidate engineering units of measure).
| |
40%
|
I can count failures or defective events using attribute data. I can tell if the function is not happening or is in some way falling short of proper performance (pass/fail performance assessment).
| |
20%
|
I have one or more ideas on how to count failures or defective function events using attribute data (pass/fail), but I don't know how to directly measure the function itself.
| |
0%
|
I do not know how to physically measure this function.
|
|
| QG17: Stability (Metric 2 of 7) |
|
Once the data is trustworthy through Measurement Systems Analysis, we can move on and gather data in a time-series context to assess its stability over time.
.jpg) | | Minitab I&MR Chart |
The chart above is an example of an Individuals and Moving Range (I&MR) Chart from Minitab statistical analysis software. Statistical Process Control charts are extremely useful during CPD&M stabilization activities in both technology and product development - not to mention their most common use in production process control. In development applications they are actually "Statistical Function or Characteristic Control" Charts. This kind of chart provides proof that your NUD functional responses are stable over time and when run successively can prove both stability and repeatability of measured performance. Without stability, the design is not mature enough to move on in the development process. If a stable sample of data is not obtainable, then a proper Capability Study cannot be run. Applying Modeling & Simulation as well as Designed Experimentation to search for stable areas of functional performance may be necessary. Stability is an essential building block for moving forward on any development project.
The table below illustrates a tiered structure of % Knowledge-In-Hand to rate the risk a NUD Function or Characteristic is carrying due to the current level of stability.
|
How to Assign % Knowledge-in-Hand for STABILITY
| |
100%
|
I use variable data SPC charts (I&MR, Xbar & R or Xbar & s) to track and document stability of functions, even if I adjust my Y=f(x) parametric relationships during iterative improvements and optimization of the design's functional performance. My function is under Statistical Process Control.
| |
80%
|
I can routinely characterize the stability of the measured function and produce Individual and Moving Range charts (I&MR) so I can see any trends, cycles or other patterns in the sample data due to noise (unwanted sources of variation beyond random effects in the measurement environment).
| |
60%
|
I have extended the time of stable performance and have enough sample data (>30) to characterize the type of distribution to which the data belong.
| |
40%
|
I have simulated the function and have defined stable performance. I have physically stabilized the function over a very broad range of variation for limited sample runs for limited periods of time. I need to conduct further development to shrink the range of natural variation of the stable performance.
| |
20%
|
I have one or more ideas on how to stabilize this function.
| |
0%
|
The function is unstable and not under statistical control.
|
|
| |
Is there a topic you'd like us to write about? Have a question? We appreciate your feedback and suggestions! Simply "reply-to" this email. Thank you!
Sincerely,
|
Carol Biesemeyer
Business Manager and Newsletter Editor Product Development Systems & Solutions Inc. |
