Quality Assurance in Decentralized I/M Networks: Analyzer Measurement Error and Field Performance

Judgments of HC/CO analyzer accuracy and precision are an implicit part of an inspection/maintenance quality assurance program. New York State's ongoing experiences in developing a program to define and measure the variables affecting field accuracy and precision are reviewed. Distinctions between design specifications, calibration curve conformance, instrument capability, and natural tolerance limits are discussed.

Field performance results for calibration curve conformance at the time of calibration are presented along with the results of three field performance surveys. The field survey data indicate a positive trend in system quality improvement. Improvements indicated by the most recent survey are confined to high range HC and CO and should be validated on a wider cross section of analyzers.

Daily variations from 11 tests on 5 instruments over several monthly test periods in a laboratory setting are analyzed as a way of estimating instrument capability.

Finally, New York State's approach in using acceptance sampling principles in determining instrument network quality is contrasted with the approach of making judgments on a single audit measurement of a single instrument in an instrument performance audit sample.