Text preview for : Uncertainty_Propagation Uncertainty Propagation for Measurements with Multiple Output Quantities c20 part of
| Home
Uncertainty Propagation for
Measurements with Multiple Output Quantities
Michael Dobbert
[email protected]
Bart Schrijver
[email protected]
Keysight Technologies
1400 Fountaingrove Parkway
Santa Rosa, CA, 95403
Abstract: The ISO Guide to the Expression of Uncertainty in Measurement (GUM) [1] limits
the description of the law of propagation of uncertainty to real input quantities and a single real
output quantity. The GUM provides little guidance for uncertainty analysis of measurements
with multiple output quantities, such as complex valued S-Parameter measurements that have
both real and imaginary components. Complex measurement quantities are common in RF and
microwave measurements. Likewise, measurements with multiple output quantities exist in
many disciplines. Supplement 2 [2] to the GUM extends the law of propagation of uncertainty to
an arbitrary number of output quantities, which is a more general solution. This paper discusses
this more general solution clearly and concisely using matrix notation. It demonstrates that the
GUM expressions for uncertainty propagation are a specific case of this more general solution.
This method is then applied to a practical measurement uncertainty example involving complex
quantities.
1. Introduction
The GUM [1] assumes that a measurement system is modeled as a function of multiple real input
quantities and a single real output quantity. This is represented as
( ). (1)
In this case, the measurand, , is a scalar quantity as are each . There exist,
however, measurement problems where the measurand must be represented by more than one
quantity. To demonstrate this, consider the following example from electrical metrology.
A common task in electrical metrology is the measurement of sine waves. Sine waves, of
course, are represented by the sine function
( ) ( ), (2)
where