CONTROL OF SAMPLING ERRORS

Year
1976
Author(s)
F.P. Roberts - Pacific Northwest National Laboratory
R.J. Brouns - Pacific Northwest National Laboratory
Abstract
Measurement control is increasingly recognized as a necessary element in special nuclear materials control and accounting. To be fully effective, a formal measurement control program must take into account all sources of errors, including those arising from sampling of materials for chemical analyses. The errors from sampling and sample handling are controlled by the use of formalized sampling plans and procedures that are based on statistical and experimental tests. The tests provide the data needed for determining the number of increments and the amount of material per increment, mixing and blending requirements, choice of sampling method, and evaluation of sample stability. Most importantly, testing provides the means for estimating the magnitude of the systematic error of sampling or, preferably, reducing it to an insignificant level.