QUALITY CONTROL PROGRAM OF THE CEESI VENTURA CALIBRATION FACILITY
Thomas Kegel
Abstract:
The CEESI Iowa Quality Control program is made up of three parts. The first part addresses management issues as described in the ISO 9000 series of standards. It consists of the documented procedures and policies that govern the day to day operations. The second part addresses technical issues as described in the ISO 17025 standard. Examples include uncertainty analyses, calibration records and software verification. The Measurement Assurance Program (MAP) is the third part of the CEESI Iowa Quality Control program. This paper discusses the components of the CEESI Iowa MAP. The main component of the MAP is the use of check standards. Each test section includes an ultrasonic meter check standard that is present during each calibration. Over time historical data have been accumulated on each check standard and typical performance has been quantified. At the conclusion of a customer calibration the consistent performance of the check standard provides assurance that the entire calibration process is also operating consistently. The formal tool used to monitor the check standards is the control chart. While the check standards monitor the entire calibration, several programs are in place to monitor the individual components. The turbine substitution test identifies the performance of individual turbine meters relative to each other. The critical pressure and temperature measurements are made with redundant instruments. Control charts are used to monitor the transducer pair differences. Finally, the weekly calibration results for the gas chromatograph are monitored using control charts. There are two overall objectives in the CEESI Iowa MAP, the first is to assure that the measurement process is operating consistently. The resulting benefits have been understood in the manufacturing community for many years, it is only more recently that the measurement community has adopted similar techniques. The second objective of the MAP is to provide data for the uncertainty analysis, this results in several benefits. The first results from obtaining uncertainty estimates based on data rather than other methods. Data based estimates improve credibility and the likelihood of inadvertently excluding any components is reduced. The second benefit is the general improvement in uncertainty resulting from using historical calibration data instead of manufacturer’s specifications. A manufacturer may includes operating constraints not present in a particular application. A set of specifications may reflect the performance for an entire production batch, the performance of a single unit may be better. Finally, the use of historical data allows for the continuous maintenance of traceability, an important part of measurement uncertainty.
