Calibration Software Uncertainty Implementation Philosophies and Practices Speaker/Author : Jorge Martins NCSLi 2011 Conference
Laboratory to Document and Approve • • • Calibration procedures Uncertainty analysis Uncertainty budgets Calibration procedures validation plan Calibration procedures validation results Traceability chain for each measurement method NCSLi 2011 Conference
Early Developments of Calibration Procedure Software • Lack of uncertainty standardization • Focus on the needs of a particular organization • No international uncertainty standards to follow NCSLi 2011 Conference
ISO/IEC 17025 and GUM 1978 ISO Guide 25 ed. 1 1982 ISO Guide 25 ed. 2 1990 ISO Guide 25 ed. 3 1999 ISO/IEC 17025 ed. 1 1995 GUM NCSLi 2011 Conference 2005 ISO/IEC 17025 ed. 2 2008 BIPM GUM
ILAC-P 14: 12/2010 Policy for uncertainty in cal Requires that to issue an accredited calibration certificate a laboratory must: • Estimate expanded uncertainty for each measurement • Estimation is compliant with GUM • Certificate shall report numerical value of uncertainty for each measurement • Implicit references to uncertainty are NOT acceptable NCSLi 2011 Conference
NCSLi 2011 Conference
Uncertainty Implementation in Software Grou p Measurement Uncertainty Type of SW Not supported • Open source • Procedure can be modified by the laboratory II Budget is supported • Open source • Procedure can be modified by the laboratory III User defined per point • Locked by developer • Procedure cannot be modified by the laboratory IV Developer defines budget • Locked by developer • Procedure cannot be modified by the laboratory I V • Locked by developer • Procedure cannot be modified by the Developer defines NCSLi 2011 Conference laboratory budget
Grou p I. • Most common in-house implementation of calibration software • Configure the instruments, sequence the readings and record the results • Measurement uncertainty not supported Grou p II. • Each uncertainty component can be defined • SW collects Type A at time of test • Effective DF calculated in real time • Expanded measurement uncertainty for each point (RT) • Original readings and Uncertainty components recorded NCSLi 2011 Conference
Grou p III. • Calibration procedure is pre-defined • User provides expanded measurement uncertainty • SW merges expanded uncertainty with measurement and saves data in DB Grou p IV. • Calibration procedure is pre-defined • Developer provides uncertainty budget • Developer is responsible for validating Calibration procedure and uncertainty budget • Data is saved on a DB • Usually developed by a third party or the OEM NCSLi 2011 Conference
Grou p V. • Calibration procedure is pre-defined • Developer provides uncertainty budget • Developer is responsible for validating calibration procedure and uncertainty budget • Data is saved on a DB • Usually developed by a third party or the OEM • Measurement uncertainty budget and calculation engine is accredited by independent SW laboratory NCSLi 2011 Conference
Document maintenance and validation responsibilities Group I Uncertainty External analysis maintaine (UA) d External Uncertainty maintaine budget (UB) d Repeatabilit y (type A) collection Not available Uncertainty analysis User resp. approval Group III Group IV Group V External maintained Not available to user Internal user available Original readings user available User resp. Unc. budget Not need SW calculations User resp. framework validation validated Calibration Procedure User resp. Internal External Internal Not available maintained user available to user Not available Original Internal readings user Not available to user User resp. Developer responsibility User resp. Not need SW framework validated Not need SW NCSLi 2011 User resp. Conference framework Dvlp. Resp. 3 rd party IT test accredited Not need SW framework
Laboratory Calibration Procedure implementation Efforts NCSLi 2011 Conference
Major Characteristics I User has complete control of measurement process Uncertainty validation is independent of measurement procedure Merged data (measurement and uncertainty) is available to issue calibration certificate Validation of uncertainty engine calculations Uncertainty budget is not directly synchronized with measurement procedure Type A data collection in real time Noisy units detection in real time User technically challenging to make changes on the measurement process Requires metrologist with SW knowledge User defined environmental conditions User defined Uncertainty analysis constraints and assumptions NCSLi 2011 Conference II IV V √ √ X X X √ √ X √ X √ √ X X √ 1) √ √ X 2) X 2) X X X √ √ X X X
Calibration SW Validation Efforts NCSLi 2011 Conference
Commercially Available Software When the software is: • Commercially available • Documented • Locked by the development house The laboratory procedure validation: • Trained technicians • Operational Capabilities • Resources NCSLi 2011 Conference
Calibration SW Development Process Published Calibration Procedure Measurement Uncertainty Analysis Procedure Revision? SW Code Implementation N SW Test OK? Y SW Validation OK? Y SW Release NCSLi 2011 Conference N
Calibration SW Development Deliverables SW House Calibration Final Report: • Reference to Documented calibration procedure implemented • Validation Report summary • Uncertainty analysis Summary Detailed SW specifications NCSLi 2011 Conference
NCSLi 2011 Conference
Conclusion • Significant difference in Laboratory: § Validation efforts § Calibration procedure control § Uncertainty analysis control Group Group I II IV V Estimation Laboratory Approval Efforts 100% NCSLi 2011 Conference 90% 55% 20%
Questions? NCSLi 2011 Conference
Скачать презентацию Calibration Software Uncertainty Implementation Philosophies and Practices Speaker Author
144ae9c73c04d35ea5723b38830b43b7.ppt