Enter your email address below and the reference for this document will be sent to shortly from firstname.lastname@example.org.
Title: Analysis Of The Traceability Chain And Calculation Of Uncertainty Of A High-Pressure Test Facility For Flow Rate Measurements
Author: Thomas Kurschat, Ottokar Brandt,Wolfram Bremser, Werner Hasselbarth
Source: 1999 International Symposium on Fluid Flow Measurement
Year Published: 1999
Abstract: pigsar is an officially recognised high-pressure test facility designed for testing and calibrating gas meters used in fiscal metering at flow rates between 8 m/hr and 6,500 m/hr and pressures ranging from 14 - 50 bar. This paper presents the test facilitys calibration chain and describes an analysis of its uncertainty based on closed uncertainty propagation of all influential variables. This approach meets the requirements of the ISO Guide to the Expression of Uncertainty in Measurement. Owing to the relatively short calibration chain, the exclusion of installation-related uncertainties by consistently avoiding changing meter installation configurations during the calibration process and the use of uncertainty-reducing regression methods, the calibration uncertainty achieved for gas meters calibrated at the high-pressure pigsar test facility is not much above 0.1 %. Correlations between standards connected in parallel are taken into account. The results of the uncertainty analysis have been confirmed by the scatter found when recalibrating the pigsar working standards. Employing uncertainty budgets to make the sources of uncertainty and their effects on calibration uncertainty transparent not only makes it possible to specify individual uncertainties in detail but also to optimise the calibration and test methods of the test facility itself. Where deviations are found between calibration levels of different test facilities, the uncertainties determined for the test facilities involved can be used to decide whether the deviations observed are significant or whether - in view of the prevailing uncertainties - the results can be considered the same.