Académique Documents
Professionnel Documents
Culture Documents
1. Introduction
This help file describes how to use the uncertainty spreadsheet and equation editor for the
evaluation of uncertainties in measurement results.
This help is complementary to the help in the uncertainty menu.
The methods used comply with ISO 17025 and with ISO GUM and also provide for
uncertainty calculations using Monte-Carlo simulations both from the spreadsheet or
directly from the measuring equation.
The uncertainty tools are located in 3 windows. To get to the main widow from the
Evaluators Board (main window) click Uncertainty. An uncertainty spreadsheet opens.
The second tool is the uncertainty equation editor. To open it click Description from the
uncertainty window. The editor is at the bottom of this window. These two tools are
sufficient when doing routine uncertainty calculations.
A third tool opens when clicking Estimator in the main uncertainty window and can be
used for some specialized work. Refer to the help in this window for how to use it.
In the following the uncertainty spreadsheet and equation editor are described.
2. How the evaluator calculates uncertainties
The correct theoretical way to start uncertainty calculation is to write the measurement
equation. This equation has the form
Y= F(X1, X2,Xn)
[1]
Where X1 etc are the components contributing to measurement outcome Y, and F is the
function that defines their contribution. From the equation F one could in principle
derived how the components uncertainties affect the uncertainty in Y. This is not always
easy and approximations are used. The simplest approximation assumes that the
uncertainty in Y can be obtained from the root some of squares of the uncertainties in the
Xis. The ISO GUM supports this approach and it results in reasonable outcome in many
cases.
If there are correlations, the ISO GUM tells how to calculate the standard uncertainty in
Y.
[2]
duplicate a page. If you do not provide a number or last is indicated, the target for these
operations is the last page.
You can flip font size by clicking the A/A button.
You can copy the contents of a cell by dragging it to another cell.
For more editing functions right click on a table.
Drag up/down the Drag to size button to show larger portions of some tables.
You may add rows or columns to some of the tables, just right click to find the right
command.
All tables can be copy copy-pasted to a spreadsheet like MS Excel in both directions.
3.1 Components Table
This is the starting point for building an uncertainty budget (even if you decide to use the
measurement equation). When the program runs per ISO GUM (click on the blue
triangle or the menu calc(ISO)) it calculates the table row by row and from left to right.
After each row is calculated there must be a value for c*u in each row, either calculated
or edited. If you do not provide the required data to calculate c*u, you should provide the
value for c*u. If you provide partial information and c*u cannot be calculated the
program will tell you that there is a problem in that row.
You may chose for each component a distribution from the supported distribution dropdown: first chose a distribution and then right-click on the required row in the
Components Table.
You may chose a component description from the Suggested Components drop down
that is a reminder of possible components. Right-click on the required row as above.
Click Description in the menu to fill-in additional info about the measurement.
At this time ignore the Equation Editor.
To calculate a field in this table write
=expression
in that field.
For example =2*3
Also, a calculation can include the mean value of the component. For example:
=1/mean(3)
where 3 is the row number and mean(3) is the value of the mean column in row 3.
To add a comment to a field, right-click on it.
The left-most column of this table shows an index of the format Ui, where i is the
component number. This is used to reference that row in other tables and calculations as
will be explained later.
You can choose a supported distribution from the drop down distribution list. Do to so,
chose a distribution and then right click on the component row Ui and chose Insert
Distribution from the popup menu.
You can choose a suggested uncertainty component from the drop down of suggested
components list. Do to so, chose an item from the list and then right click on the
component row Ui and chose Insert Suggested Component from the popup menu.
3.1.1 Supported distributions
As of today, the supported distributions are
Data
Student
Normal
Rectangle
Triangle
U-Sine
Two-peak
The distribution names must be type exactly (upper/lower case) as stated or chosen from
the drop down list as described above.
The Data distribution is the empirical distribution of the actual measurements. For more
on this see the paragraph on the Measured Data Table below.
The U-Sine is the U distribution.
The Two-peak distribution is a two-delta function distribution (two narrow spikes).
When choosing a distribution, an icon next to the distribution drop down list is displayed
showing the form of the probability function is shown in a little window.
3.2 Measured Data (repeatability) Table
The Measured Data Table is used when repeated measurements are to be incorporated
into the uncertainty budget. To add the contribution of the repeated measurements, first
decide which component in the Components Table it will be. To do so, in the first row
of the Measured Data Table write the number of Ui. That is, if you chose to make the
measurement component 3, write 3 in the first row. When you do so, in the header the
title Line U3 will appear. Starting from the 2nd row, enter all the measured data as a
column. When you click calc (any ISO, Monte-Carlo or one of the triangle buttons) row
3 in the Components Table is filled (note that if not all required fields are filled in the
program will complain). The default calculation was done assuming that the repeated
measurements are normally distributed (it will add student distribution and degrees of
freedom accordingly). This can be overridden by filling a different distribution, e.g. Data
distribution in which case the calculation will be different.
Also, the calculated component in (row U3 of the above example) is the standard
deviation of the mean of the used data. You can choose a different statistic and even use a
calculation on the data before it is used. This is useful when the data is given in nonlinear
units, like dB, which should be converted before calculating the standard deviation.
The available other than standard deviation of the mean (sdm) statistics are
sdm standard deviation of the mean
ssd sample standard deviation
rms root mean square
Add the above tokens after the row number, e.g. 2 ssd
Will use the data in row 2 and calculate the uncertainty as the standard deviation.
To apply a formula to the data before the statistic calculation follow this example:
2 ssd=()/2
where the formula is to the right of = and the empty () stand for the data.
This example will first divide the data by 2, calculate ssd and put the result into row 2.
The expression
5=10^(()/10)
will convert from dB to ratio, calculate the standard deviation of the mean and use the
result in row 5.
To do more complicated calculations on the data, use instead the Data window on the
main board menu. When done, import the data in that table to the Measured Data Table
and just edit the row numbers.
When the chosen distribution for the measured data is Data, the standard deviation (and
other statistics) are derived from the actual data. When computing uncertainties using
Monte-Carlo simulation (see below) with the Data distribution, the simulated values are
the actual measured data points.
If there is only one measured data point, do not use the Measured Data Table. Instead edit
a row for this component and provide the uncertainty from prior or other knowledge. The
measured value can be written in the Mean Value field in the Component Table.
To enter a known correlation coefficient in the table, write in the first two columns the
component row numbers and in the 3rd the correlation coefficient value. For example :
Ui Uj rij(correlation)
1 2
0.5
To compute the correlation coefficient of two data sets, e.g. two columns in the Measured
Data Table, which correspond to two components (e.g.1 and 2) write:
Ui Uj rij(correlation)
1 2
x
x
Another possibility is to Select all from the edit menu of the main board (after
previewing to the board), copy-paste the text to a word processor and print from there.
All tables can be copy copy-pasted to a spreadsheet like MS Excel in both directions.
4. Complex Uncertainty Components
Complex components may be used in the calculations. For each such a component you
must devote two rows, one for its real part and one for the imaginary part. In each row of
the Components Table only the c sensitivities and therefore also C*U can be complex. All
other numbers must be real. This restriction is used to facilitate the computation of
uncertainty directly from the measurement equation.
To indicate that a c value (a derivative) is complex use j next to its imaginary part, e.g.
3+2j.
If the program encounters complex or imaginary numbers, the results are calculated
twice, for the real and imaginary parts, as described above in the Result Table section.
5. Monte-Carlo Simulation from the uncertainty tables
Monte-Carlo simulation is a calculation based on generating possible outcomes for each
of the components in the Components Table in many iterations. This process is repeated
many times and a distribution of the outcomes is obtained. All the uncertainties are
calculated directly from this distribution.
When using the tables for a Monte-Carlo Simulation, The actual distributions given in the
table are used. However the assumption is that the measurement equation is just the sum
of all components. With this respect, this type of Monte-Carlo is a compromise between
A simulation directly from the measurement equation and the ISO GUM simple
approach.
6. Monte-Carlo Simulation directly from the measurement equation
The measurement equation (see equation 1 above) relates the components to the final
measurement result. A mote-Carlo simulation inserts possible values for the components
and calculates the outcome of the equation many times, for many possibilities of each
component. A probability distribution for the outcome is generated from these trials and
the uncertainty is calculated from that distribution.
To write a measurement equation represent the components in the equation using u(i),
where i is the component row number in the Components Table.