Non-destructive assay techniques and associated measurement uncertainties

To calculate the inventory differences for a facility or process, we should have random error and bias estimates for all the measurement systems in addition to the measurement results on each sample. For non-destructive assay (NDA) instruments, the assignment of these measurement errors is quite complex. This paper discusses the random and bias estimates of two NDA systems, one measuring homogeneous samples and the other measuring heterogeneous waste samples. Actual measurement data are used to illustrate the error estimates. The entire measurement system/material category combination must be examined to accurately assess measurement uncertainty for each measured item. Such an assessment might start with an error matrix for each instrument/material category combination as shown in this paper, in which many sources of error are considered. Only then can the random error and bias for measurements from the system be assigned.