Overview
ABSTRACT
The term “sum” used in this article means a sum or a mean of low level measurements, i.e. close to detection limits. The methodology proposed here overcomes the biases due to the classical use of non-significant value substitution methods. It is based on two fundamental principles: (i) the sum of measurement values must be determined from each of the individual measurement values, including non-significant or even negative ones, and (ii) in sum operations, decision thresholds are combined in the same manner as uncertainties. Three examples illustrate the proposed approach throughout the article.
Read this article from a comprehensive knowledge base, updated and supplemented with articles reviewed by scientific committees.
Read the articleAUTHORS
-
Marielle CROZET: Research engineer at CEA - CEA, DEN, DRCP, SERA, LED
-
Cédric RIVIER: CETAMA Technical Secretary at CEA - CEA, DEN, DRCP, CETAMA
-
Stéphane PUYDARRIEUX: Process engineer, applied statistics expert - AREVA NC La Hague
-
Alain VIVIER: Senior expert in dosimetry and statistics - INSTN, ETSR, Saclay
-
Vincent BRUEL: Operations Research Engineer - AREVA, BUE, SET, IPE, Tricastin site
-
Guillaume MANIFICAT: Head of the department for monitoring and studying radioactivity in the environment at IRSN - PRP-ENV, SESURE
-
Marcel MOKILI: Research engineer - SUBATECH – UMR 6457: École des Mines de Nantes, IN2P3/CNRS, University of Nantes
-
Bernard THAUREL: Research engineer - IRSN – PDS, DEND, SATE
INTRODUCTION
A wide range of fields are concerned by cumulative analysis: for example, liquid or gaseous discharges into the environment, waste management, material balances, analysis of impurities in a finished product or reference material, environmental monitoring...
This article deals solely with the accumulation of low-level measurement values, and applies in particular to chemical and radiological analyses.
Most of the aggregation methods currently in use involve the substitution of insignificant measurement values, thus generating biases, sometimes very significant, in the aggregation result. These biases are most often positive and correspond to virtual quantities of material analyzed, thus artificially distorting the balances established by laboratories as part of their regulatory requirements, for example. These biases can be negative if non-significant measurement values are replaced by zero. Changing these values not only introduces bias into the cumulative result, but also distorts the assessment of its uncertainty.
This article has been written to clarify the rules for calculating measurement accumulations. After defining the basic terms required, it sets out the recommended method for expressing the accumulation model, uncertainties and associated decision thresholds. The various cumulation methods currently in use are then presented. Three examples are used to illustrate the significant differences in cumulative values obtained by using one or other of these methods.
Exclusive to subscribers. 97% yet to be discovered!
Already subscribed? Log in!
KEYWORDS
decision threshold | quantification limit | detection limit | chemical analysis |
Cumulative measures
Definitions
Article included in this offer
"Laboratory quality and safety procedures"
(
140 articles
)
Updated and enriched with articles validated by our scientific committees
A set of exclusive tools to complement the resources
Bibliography
Exclusive to subscribers. 97% yet to be discovered!
Already subscribed? Log in!
