Reduction of Sampling Inspection Using Variables Data in the Presence of Batch-to-Batch Variability

Author

David Trindade, PhD, Fellow and Chief Officer of Best Practices, Bloom Energy

Many companies perform attribute sampling in which a specified number of units is inspected and each unit is classified as either conforming or nonconforming based on single- or double-sided specification limits. An acceptance number determines the lot disposition for acceptance or rejection. Attribute plans are simple in concept since only counts of defective units are involved. In contrast, variables sampling plans require the measurement of a quality characteristic and use of criteria related to the specification limits. Such plans offer many benefits including: the same or better discriminating ability as attributes inspection; smaller sample sizes; reduction in time and measurements costs; and greater efficiency in the use of company resources. However, several critical assumptions involving the distribution of measurements in variables sampling need to be verified, and additional considerations apply in the presence of batch-to-batch variability. In this presentation, we describe a successful implementation of variables sampling in a manufacturing environment. We demonstrate many capabilities in JMP for checking the validity of assumptions and analyzing batch-to-batch data. Variables sampling has resulted in a dramatic reduction of inspection requirements and increased productivity.

Author

David Trindade, PhD, Fellow and Chief Officer of Best Practices, Bloom Energy

Discovery Summit 2014 Resources

Discovery Summit 2014 is over, but it's not too late to participate in the conversation!