Abstract
The paper deals with the problem of bias randomization in evaluation of
the measuring instrument capability. The bias plays a significant role in
assessment of the measuring instrument quality. Because the measurement
uncertainty is a comfortable parameter for evaluation in metrology, the
bias may be treated as a component of the uncertainty associated with the
measuring instrument. The basic method for calculation of the uncertainty
in modern metrology is propagation of distributions. Any component of the
uncertainty budget should be expressed as a distribution. Usually, in the
case of a systematic effect being a bias, the rectangular distribution is
assumed. In the paper an alternative randomization method using the
Flatten-Gaussian distribution is proposed.
Go to article