Conformity assessment: Product & measurement assessments

Product (or entity) conformity assessment

The aim of conformity assessment of any type of entity is to assess conformance of actual values of a characteristic with respect to specification limits [Montgomery 1996, Joglekar 2003].  Dispersion in characteristic entity values will be due to actual variability in the manufacturing process when product is made. Subsequent entity variability will be due to wear and tear during product lifetime.

The various tools of statistics can be used in different ways in conformity assessment: they can be used to describe actual variability as well as enable modelling of probable variability. Such statistical modelling is useful in cases where actual variability is not known; where prior knowledge needs to be included; and when predictions are made in advance in order to plan for measurements in the best way.

In evaluating production variations of quantity η = X in the ‘entity (or product) space’, measurements made multilaterally might be on repeated items in a production process or by taking a sample of the population of items subject to conformity assessment. The corresponding probability distribution function PDF, gentity (x) , will have a form determined ideally (in the absence of measurement or sampling uncertainty) by the intrinsic quantity variations of prime interest in conformity assessment.

Specification limits are often set in conformity assessment on actual values of a characteristic of a type of entity.

Irrespective of which entities are subject to conformity assessment , it is important to specify the assessment target as clearly as possible: ‘Global’ conformity denotes the assessment of populations of typical entities, while ‘specific’ conformity assessment refers to inspection of single items or individuals [Rossi and Crenna 2006].

Measurement conformity assessment

Measurement quality variations, expressed with a measurement uncertainty PDF, gtest(y) of quantity η = Y in the ‘measurement space’, may partially mask observations of actual entity dispersion. As such, measurement variability is just one, and hopefully a relatively minor, source of uncertainty which needs to be accounted for when making decisions of conformity. A general challenge is to find methods which reliably can be used to separate these different variability components. Uncertainties associated with entity variability and stimulus variability will often be of dominant concern. Overall decisions of conformity will need to account for each of these different kinds of uncertainty in order to assess commensurate risks correctly.

When using and interpreting measurement results it is usually essential to distinguish between random and systematic measurement errors. Often comparisons of various kinds need to be made, e g

  • between the measurement result and a specified limiting value;
  • between a test or group of tests in order to decide if differences (effects) are present;
  • within groups in order to judge heterogeniety

With comparisons it is not necessary to take care of error components which are common and constant (systematic or locally systematic) and therefore for different kinds of comparisons different uncertainties have to be estimated and given. To identify and estimate the relevant uncertainty is usually one of the most critical steps in a decision.

Specification limits are often set in measurement conformity assessment on actual values of a characteristic of a type of measurement.

Difference between measured error and measurement error

A principal distinction must be made between:

  • Measurement error, and
  • Measured error

An error in a ‘quantity’ of an item (property of aphenomenon, body, or substance: VIM 1.1), such as the Thickness of manufactured spacer or Indication of an instrument, often refers to a quality characteristic (‘property intended to be assessed’) of such items when assessing their conformity to item requirements (specification limits, MPE of item) which are set prior to measurement.

Actual measurements made of item error subject to conformity assessment will produce results with which estimates can be made of the actual value of that quality characteristic of the item – this will be the “measured error” (not “measurement error”). In such measurements there will also be measurement errors which lead to estimates of the item error which differ from the ‘true’ item error.  Those measurement errors, if not corrected for, will lead to measurement uncertainty in the measured error which can lead to risks of in-correct decisions of conformity.

Confusion might easily arise (as in the earlier VIM definition of ‘maximum permissible error’) between measured error and measurement error, particularly (as in legal metrology) where a measurement instrument is the item subject to conformity assessment. The error in indication of an instrument is not a measurement error but a measured error in conformity assessment (but of course an instrument showing an erroneous indication will lead to measurement error when used in measurements).



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: