Research & Innovation Services

Processing Unstructured Data in Naval Engineering Records

KTA Partnership Development Award

Innovative design and Manufacturing Research Centre (IdMRC), Dept of Mechanical Engineering and Babcock International Group Plc.
Prof Chris McMahon and Prof Steve Culley 
“Working with Babcock has allowed the academic team to engage with practising engineers who share their vision of intelligent computer-aided exploitation of engineering records.”

Professor Chris McMahon, Department of Mechanical Engineering, University of Bath
“Working with Chris McMahon allowed us to take ideas and exploit them for real. Out of that has come some very useful lessons learned, which we can pass on to others. For example, whilst the methodology and technology we used has solid foundations, there is more we can learn from the data we have, which can feed back into improving the services we offer.”

Chris Rowley, Head of Information and Knowledge Management, Babcock Marine & Technology

Challenge

The maintenance costs for naval assets, such as surface ships and submarines, run to hundreds of millions pounds Sterling each year. A 1% improvement in efficiency can swiftly translate into more than £1 million.

Chris Rowley, of Babcock Marine & Technology, had noticed that ways of classifying and interrogating data, set up early in an engineering project, often constrain the thinking that follows. This can apply many years after a naval engineering asset, for example, has been designed, constructed, deployed and subsequently modified.

Is there a way to revisit data already gathered to reclassify and reinvestigate it to find hidden connections? Doing so could prompt new, productive questioning, which ultimately would help in servicing, repairing and redesigning complex engineering assets.

Solution

Chris Rowley approached Professor Chris McMahon and his IdMRC team at the University of Bath. They have leading edge expertise in faceted classification – assigning multiple classifications to an object so that a database of such objects can be interrogated flexibly in multiple ways. For example, car sales websites often use faceted classification to enable customers to search for cars using numerous criteria. In this project, a sophisticated classification scheme was applied to components in naval submarines and surface ships, drawing upon both structured and unstructured data.

Benefits and outcomes

  • A faceted classification scheme was created for text descriptions of naval vessel engineering concessions (deviations from design specifications, process or schedule) and applied to Babcock’s database systems. 
  • By analysing and classifying free text used to describe concessions, further breadth and depth was added to the classification scheme.
  • Concession records can now be browsed and searched to reveal previously hidden patterns and to prompt and answer new questions.

The productive collaboration between the University of Bath and Babcock is continuing, with follow-up studies on the risks associated with engineering components and systems fitted on naval vessels.

Project team

Professor Chris McMahon, Project Leader and Engineering Director of the Innovative design and Manufacturing Research Centre, Department of Mechanical Engineering
Prof Steve Culley, Co-Investigator, Department of Mechanical Engineering
Dr Uday Thangarajah, KT Fellow, Department of Mechanical Engineering
Chris Rowley, Head of Information and Knowledge Management, Babcock Marine & Technology

Funded by the University of Bath’s EPSRC Knowledge Transfer Account.