Input measurand:
,
i=1..N
Output measurand:
Estimate of
:
Standard uncertainty of
:
Estimate of Y:
Combined standard uncertainty, GUM order 1, uncorrelated input quantities
Estimate of standard uncertainty of Y=
combined
standard uncertainty:
With
:
sensitivity coefficient
Combined variance:
Combined standard uncertainty, GUM order 1, correlated input quantities
where
is
the estimated
correlation
coefficient
between
et
:
where
is
the estimated covariance associated with
and
.
Combined standard uncertainty, GUM second order, uncorrelated input quantities
-Useful in case of significant non linearity. Every distribution is supposed normal and only most important term is retained.
-Estimate of standard uncertainty of Y:
With
Remark:
Each measurand can eventually be affected by many error source:
,
where the error
has
zero average and
standard uncertainty
The standard uncertainty
foris