Measurement data management with openMDM
Day-to-day practice in the test departments of companies shows: Modern measurement and test procedures create enormous amounts of data. It is increasingly becoming a problem to manage them efficiently. In many companies, approaches to cope with the data flood that have been pieced together over many years still prevail. Three distinct approaches can be identified here: management of measurement data in file-based directory structures, management in individually developed database solutions and the use of proprietary solutions from providers in the field of test engineering. These approaches have one thing in common: They are limited by departmental boundaries.
Let testers speak the same language
The automobile industry as an example: Here the product development requires a lot of testing effort. It is therefore not surprising that many of the efforts to make test results more easily interchangeable are rooted in this industry. ASAM (Association for Standardisation of Automation and Measuring Systems) is a working group of leading automotive manufacturers that, in ODS (Open Data Services), has defined a format for the storage of measurement data and their metadata. Thereby, a cross-application, vendor-independent standard was established.
The good thing about it: ODS is generic enough to also map any specifications apart from the automobile industry. It is therefore perfectly suitable to establish a standardized format for the results of different test procedures and thus to create a standardized test data management system that applies company-wide. Test data from different test systems are documented with descriptive information to be able to always correctly interpret and compare them, even from a different location or at different times. These metadata are used to evaluate the professional, organizational and technical context of the data. The context includes, for instance, the description of the test specimen – the “UnitUnderTest“ in the language of the ASAM ODS – the test sequence, the test structure, the simulation parameters and the organizational and order-related data. This meta information is an important prerequisite to be able to search and navigate later on in the data stock, e.g. to be able to assign it to different product and test data.
Experience shows that setting up a company-wide uniform test and measurement data management system is a management decision. The decision only rarely comes from the individual test departments because, at first sight, it involves additional work to allocate metadata to measurement data and to save them in the ODS format. However, test documentation can be facilitated with a suitable software platform. openMDM is software that has been tried and tested in many cases. Its origins lie in a measurement data management system by carmaker Audi. Since 2008 it has been further developed into an extremely versatile and scalable enterprise solution. Since then, the number of users has increased continuously – even outside the automotive industry. It has become evident: Above all, technical rather than branch-specific requirements are the key factors for the implementation of a professional measurement data management system.
openMDM is a modular system which shows the entire process from test planning to archiving. Though it normally works with Oracle databases, the use of other databases is possible. Besides ASAM ODS standard it also uses CORBA and OSGI. Interfaces for different measurement and analysis systems such as Matlab, Famos, Diadem, Artemis, TestLab and many others already exist. Cross-system evaluation and calculation processes for measurement data can be automated in the software via workflow definitons.
These are stored in a structured way in the ASAM ODS database, so that already defined partial processes can easily be reused in other workflows, or assembled into completely new workflows. openMDM is scalable upon demand. As experience teaches, according to the principle "think big, start small“, it is recommended to start with an individual application-specific solution and afterwards expand the solution department by department.
What is important: The use of a comprehensive data format and of a measurement data management system that is designed for company-wide usage should not result in an extensive effort by engineers within the individual test departments. The primary objective of measurement data management stays to support daily work at the test bench. Such a process support begins, for example, during test planning, and includes test commissioning, test data storage and test evaluation. If thereby test planning is systematized via openMDM, the planning data laid out can be used immediately as meta information for measurement data recorded later. Using freely definable templates, you can define which attributes are used to describe test orders, test specimens, measuring equipment and so on, and the test steps used to make up a test. Distinction can be made between meta information, which remains constant during the entire test, and information that changes from one test step to the next. Time-consuming manual entries and input errors or gaps associated with this can be minimized by pre-allocating in template toggle fields. For instance, routines for the automatic generation of unique test designations facilitate management.
The data flood in one individual test department can be the reason to start a measurement data management project. If afterwards, the topic is dealt with systematically and with foresight, an information base emerges that can be linked to other information, and thus be used by different departments such as research, development, production and after sales. Thanks to flexible role and rights models, openMDM can also be used in projects with external development and test service providers. Test and measurement data are too valuable to remain shut away in isolated solutions. They should be available company-wide and during the entire product life cycle.
About the author:
Dr. Hans-Jörg Kremer is general manager of Peak Solution GmbH
All images: Peak Solution.