Читать книгу Paint Analysis - Roger Dietrich - Страница 32

The previously discussed random sampling serves as a tool to find out if a certain test portion of a population exhibits a feature or failure and quantify the risk of the appearance of certain features. The targeted sampling is a method which is used if a feature is already known and/or visible and portions offering this feature are specifically selected.

If you look e.g. at a population of coated polymer parts like shown in Figure II.11 and the task is to find out if there is certain risk that this population may contain an (unknown) percentage of failed parts with poor paint adhesion (that you do not see without testing) you would choose for a random sampling of parts that will be tested by a cross-cut test to find out if the population is o.k. taking account of a certain confidence level (see Example: Zero defect sampling in Chapter 3.2. Part II).

If you see the defects (coating peels off or exhibits bubbles) you pick out the failed parts by a targeted sampling and explicitly investigate these failures.