GRANTA MI Supports University Research

Software aids study of road construction materials, captures results for future use

University of CambridgeUniversity of Cambridge researchers have applied the GRANTA MI materials information management system to study the bituminous materials (asphalt) used in roads. As well as providing insight into one of the world's most widely used construction materials, the project showed how research groups can benefit significantly from tools—and a systematic approach—to capture materials data, analyses, and results.

The need for materials data management

 
"The GRANTA MI material data management system was found to provide all of the necessary data types and data structures needed for managing this test data. Its automated input and output facilities enabled the data importing and analysis tasks to be made fast, reliable and consistent, ensuring efficient generation of high-quality research data."
   

Materials and engineering research groups in universities, government labs, or industry aim to develop a deep understanding of a subject area over the years, as each research project builds on prior data, experience, and analyses. In practice, the process is rarely ideal! Results from previous projects are typically not retained in a systematic manner that enables easy re-use. Research projects are recorded through publications, reports, and theses, so that their broad conclusions are passed on. But details such as raw data, models, and experimental parameters may be left scattered on CDs and file systems or in researchers' notebooks. This means results are often difficult to verify and repeat, because the data and analyses that support them cannot be traced. Consequently, experiments often need to be duplicated by subsequent researchers. Many research groups are also missing the opportunity to build, over time, a coherent body of data that could itself be a valuable research resource.

The GRANTA MI materials information management system has been developed in response to similar problems in industry: companies want 'traceability' so that they can, for example, examine the pedigree of data used in design; they want to waste less time repeating experiments or searching for reliable data; and they want to re-use and develop prior knowledge effectively.

The asphalt project

A research team in the Transportation Research Group at the University of Cambridge had these wider issues in mind when they came to study the mechanical properties of asphalt. The team built an integrated database, using the GRANTA MI platform. This was used to capture all of their raw test data, its associated pedigree information, and publications on permanent deformation of bitumen and asphalts.

In the most recent phase of the research, the objective was to investigate crack propagation through a road surface under repeated wheel and/or environmentally-induced loads. This is an important issue in considering how to improve road performance. Their approach was to video cracks as they developed under controlled mechanical testing conditions, measuring and recording the output from the testing machine alongside the resulting images. An automated image processing routine determined the crack profile and measured the crack length from each video image. It then uploaded both the image and crack length data into the database and 'rolled up' data for crack length versus time.

This created large amounts of data—including numeric data, graphical data showing the variation of properties with time, and the video images. GRANTA MI allows users to capture, store, and analyze such a range —from specialist materials information to generic media files and documents. It also provides powerful tools to capture, explore, and maintain the relationships between items of data, which are often vital in a research project.

Crack image

Figure 1. A sample digitized image from the experimental work. Videos of crack propagation were converted to a series of such images, which could then be analyzed to measure properties such as crack length. The raw images and the analysis results were all stored in the GRANTA MI database alongside measurements collected by the testing machines.

GRANTA MI also ensured that the research team's data and results are now available for use by any of their colleagues in the Transportation Research Group and by future researchers.

Results

The team summarized the practical aspects of its work as follows:

  1. A materials data management system was implemented to collect, organize and analyze an extensive set of experimental data on bituminous materials. It covered a wide range of test types including tension, compression, triaxial tests, adhesive joint fracture, and three-point bend tests. It included all necessary pedigree information about the material, specimen, test standards and equipment and testing conditions.
  2. The structure of the database enabled full traceability of all aspects of the data. This ensured that the data can be re-analyzed and re-used in the future with confidence.
  3. An automated process was developed to manage two separate data streams generated by three-point bend tests. Video images of the test were used to compute crack profiles and crack length statistics and this information was synchronized in time with force, deflection and crack mouth opening displacement data collected by the testing machine. All of this data was stored in a test data record, from which it can be extracted in order to compute fracture parameters such as fracture toughness, J-integral and C*-integral.
  4. The GRANTA MI material data management system was found to provide all of the necessary data types and data structures needed for managing this test data. Its automated input and output facilities enabled the data importing and analysis tasks to be made fast, reliable and consistent, ensuring efficient generation of high-quality research data. Its flexible reporting tools enable rapid retrieval of information for inspection and further analysis by external software systems.

GRANTA MI screenshot—click for larger image

Figure 2. A typical crack record in the GRANTA MI database. Numerical, graphical, and image data is all captured. The record can be linked to related records describing, for example, test machine output, the pedigree of the sample being tested, or rolled up data for the complete test. Such capture and linking can be automated, making it practical to deliver full traceability through all of the project data

 

Conclusion

GRANTA MI can support practical research projects and, in so doing, can help research groups and departments to capture data, information, and knowledge for future use. Wider implementation of such systems could:

  • help researchers to build more effectively on previous work and to publish their data in a structured digital format
  • maximize return from the investments made by project sponsors
  • help to eliminate duplication of work, and
  • build richer data resources for researchers to mine.

More information

Portillo, O, Johnson, G, Lam, W. and Cebon, D. ‘Automated analysis and data handling for asphalt fracture experiments’ Technical Report CUED/C-MATS/TR.262, ISSN 0309—6505, October 2007.

Transportation Research Group website: www-mech.eng.cam.ac.uk/trg/