Is it an issue when sampling heads yield slightly different results in comparative measurements? Is an expensive voltmeter with a higher precision range required, or would a less expensive meter suffice? What point in the measurement process is the greatest source of uncertainty?
The answers to these and other questions can be found only by performance of a measurement uncertainty analysis. Since 1993, the Guide to the Expression of Uncertainty in Measurement (GUM) has become a recognized code in technical regulation for this purpose.
A measurement uncertainty analysis is also required for the validation of methods for the determination of airborne chemical agents. The calculations required for this purpose are generally resource-intensive and time-consuming. The IFA has responded by developing an application that reduces the calculation effort required of users for some of the validated methods: MUST, the Measurement Uncertainty Service Tool. Provided users observe the selected method and use the equipment specified for it, they need only enter the measured values in MUST and upload the method's validation data. The result is then obtained, together with the associated expanded measurement uncertainty according to GUM. In addition, MUST calculates the contribution made by the individual devices and method effects to the measurement uncertainty.
An introduction to the calculation of the expanded measurement uncertainty in the context of MUST can be found in the articles listed under Further information (in German), and in the slide set of a presentation on the subject (in English).
A complete description of MUST can be found in the instruction manual (PDF, 362 kB, non-accessible) . The user interface of the application is divided into tabs.
General information and details of the method can be entered on the first tab. The current version of the application supports the following methods:
The measured values are entered on the second tab. The application also requires two Excel files:
The concentrations/masses for which the measurement uncertainty is to be calculated can be entered manually, as can the duration of sampling, volume flow and digestion volume. After the calculation is performed, the results are displayed and are ready to be saved in the form of an Excel file for documentation purposes.
Tabs 3 to 5 display the data for recovery and calibration and model data, thereby indicating that they have been read in correctly.
MUST runs on Windows 10 (64-bit).
The application is written in MATLAB (version 2021a) and is supplied together with a MATLAB runtime environment (version 9.10) in a compressed archive. The application and runtime environment together are approximately 7.5 GB in size; even in compressed form, the archive is still around 3.6 GB.
Owing to the size of the archive, a stable Internet connection is required for the download. It is recommended to check at the end of the downloadg that the archive has been downloaded completely.
It must then be unpacked, as the application can be launched only from the unpacked files.
The archive was packed with 7-zip (a free application), which should also be used to unpack it; use of the Windows operating system functions to unpack the MUST archive may give rise to errors.
Once the archive has been unpacked, the link can be found on the topmost folder level. Double-click on this link to launch the software. The link points to the executable file. This in turn is located within the folder structure, and its relative path must not be changed. The top folder level also contains three sample Excel files containing calibration and recovery data for each method type (Thermodesorption_Calibration&Recovery.xlsx, Metals_Calibration&Recovery.xlsx, Extraction_Calibration&Recovery.xlsx). The data in these Excel files are intended for test purposes only, to demonstrate how the application works.
The top folder level contains a further Excel file (Modelldatenblatt.xlsx). This file contains further influencing variables and uncertainties that do not originate from the calibration and recovery tests. The file may be adapted to the local conditions (refer for this purpose to the manual). For example, it contains values for the repeatability uncertainty of volumetric instruments. These values can be replaced by data from your own calibration protocols.
MUST looks for Modelldatenblatt.xls under this name and only in this folder. Conversely, the file containing the calibration and recovery data can be located in any folder that can be accessed; the only requirement is that the name of the worksheet be retained unchanged (Zusammenfassung MU).
The relative path for the executable file is .\v910\bin\win64\ and must not be changed.
The IFA can make the source code of MUST available to users who have the MATLAB software. If interested, please contact must@dguv.de.
MUST has been developed with care and in accordance with current good practice. Nevertheless, no liability can be accepted for the correctness of the results, irrespective of the legal grounds
The IFA endeavours to ensure that its website remains free of viruses. No guarantee can however be given that MUST or the files linked to here are virus-free. Users are therefore advised to take appropriate precautions themselves and to install virus scanners before downloading MUST.
The executable MATLAB application and the Excel files supplied with it have been tested with Virustotal.com and classified as not suspicious.
The accessibility of MUST is declared in the accessibility declaration for the application.
MUST is freeware and may be used commercially and for educational and private purposes. MUST is available for autonomous use without restrictions. MUST may be passed to third parties free of charge. No entitlement to support from the IFA may be assumed. Any party to whom the source code has been made available may adapt it to their own requirements.
Should you have any questions or suggestions for improvement, please contact the IFA.
Bug Report: Version 1.1 contained a bug which has been fixed in version 1.2. Please download the latest version.
(Last revised on 20th February 2024)
Measurement Uncertainty for Measurement Methods of Air Monitoring (PDF, 1.8 MB, non-accessible) . Airmon Conference 2022