Quickly harvest the value from data

Technologies such as artificial intelligence are advancing at high speed. However, access to consolidated, quality data remains a challenge.

Almost from day one, OCAS was conscious about the value hidden in our R&D data. Our in-house developed database tool has meanwhile transformed into a mature lab integrated management system (LIMS) database providing our researchers and technicians significant efficiency gains.

From lab management to data capture to deep learning

With this important step, OCAS integrated content: “what needs to be done on which sample”. Each sample is subject to a ‘standard activity’, enabling a detailed description of the job in a straightforward manner. An activity lists the resources needed such as authorized operators, calibrated equipment, up-to-date recipe: international standard and/or standard operating procedure (SOP), chemicals which Safety Data Sheet (SDS), raw materials, etc. After execution, each standard activity generates specific data that is stored for further use.

For example, the casting activitiy manages the preparation of casting blocks of any composition. Based on a selection of available raw materials (pure elements, alloys), the system computes the respective amounts to be added to get the desired chemical analysis. After casting, the amounts effectively used are substracted from the stock. Next, a “cutting” activity manages the partition of the casting block in various to-be-rolled blocks. A sample is also extracted for determining the chemical analysis for which a series of activities (one per analytical technique) have been designed. Further on, numerous other activities are carried out: rolling takes place followed by heat treatments, mechanical testing, surface treatments, microstructure analyses, etc.

With the latest update of our LIMS, OCAS linked its lab efficiency to an even higher level of quality assurance, thanks to these digital standard operating procedures. The reporting module has been improved with monitoring tools and dashboard views. Validation flows are also integrated. In addition to the benefits of this modular in-house developed application, the major advantage is the fact that it is highly reliable, very close to our needs and easy to customise to an ever-changing environment.

Quality awareness has become even more important. Our processes are now described in much more detail and our overall management system has been adapted to be fully compliant with accreditation. This helps us to push our performance to a higher level: continually improving data quality and laboratory effectiveness, results in lean processes to the benefit of our customers. Improved availability and analysability of data makes it easier to estimate the technical feasibility of projects leading to a higher success rate.

Quality and timing are important for our customers. Being fully certified according to ISO9001 & ISO17025 for its quality and OHSAS18001 for its policy on occupational health and safety, we feel that OCAS further enhances its service level.

Some quick wins

  • Fast reporting: as the database contains all necessary information: method, samples, operators, validated results, etc. The report is generated automatically.
  • Data transfer from equipment to database through Excel connectors avoiding typing errors
  • Data consistency is checked with business rules such as ‘a chemical composition must equal 100%’ or ‘the average value shall be between min and max’
  • Operational excellence: dynamic statistical process control, easy determination of uncertainty of measurement, easier planning (people & equipment), etc.
  • Process control: life expectancy of consumables (e.g. crucibles), detection of systematic deviations (e.g. chemical composition of castings)
  • Database is ready for datamining and easily linked with deep learning package

 

 

“The result database of our LIMS is the keystone for further developments including big data analysis.  A lot of information can  be extracted from e.g. microstructure images using deep learning.”

Roger Hubert, Chief Scientific Officer, OCAS