Statistical Engineering Division
Seminar Series
A Software Tool to Calculate Uncertainty in Measurements
Marco Wolf
Eidgenössische MaterialPrüfungs und forschungs Anstalt
(Swiss Federal Institute for Material Research and Testing)
Statistical Engineering Division, NIST
NIST North Room 618
Wednesday, August 31, 2005, 10:30-11:30 AM
Abstract
Evaluating measurement uncertainty by hand can be a very complex task. The
"Guide for the Expression of Uncertainty in Measurement" specifies how
calculations should be done to evaluate uncertainty. A perfect world model
has to be used where the influences and their distributions are well known
and independent of dynamic parameters. Software exists to calculate the
measurement uncertainty for those static models. We develop new software
for simulating the behavior of measurement systems which takes into account
variation sources like time, person, etc. The software is also capable of
including correlation/dependencies in the model. It allows a
straight-forward definition of complex measurement systems and their
procedures/protocols in three steps. The software is able to calculate
expected measurement uncertainty using the Monte Carlo method and helps
analyze the influence of variation in settings.
Speaker Bio
NIST Contact:
Blaza Toman,
(301) 975-4516.
