Sellafield are seeking to innovate in the way they analyse Health, Safety and Environmental data to improve insight, trend analysis and early prediction of safety issues.

The data which needs to be analysed includes safety observations, assurance activities, assurance action tracking information, accident reports and unsafe condition reports.

Applications are invited for technological solutions to meet this challenge.

The deadline for applications is Friday 24th January at 12 noon.

The Situation

Sellafield are exploring the use of Machine Learning (ML) to help analyse health, safety and environmental data to improve prediction of risk.

Sellafield’s internal document management and reporting systems contain condition reports, observations, assurance activities and related action tracking, and a wealth of structured and unstructured data from which insight can be drawn.

Current Practice

Data is trended against categories, and dashboards are prepared at a business unit level, with individual reports reviewed at a facility or building level.

Current examples of trending categories include ‘near misses’ and ‘slips, trips and falls’.

There are many safety reports (>100k), safety observations and assurance activities across a wide number of plants (>100) dating back some 15 years, important insights are expected to be contained within free-text.

The data is stored within Sellafield’s internal document management systems and interrogated manually. This is a laborious and time consuming process.

Challenge Aims

Sellafield would like to undertake a Proof of Concept project to determine how Machine Learning could be used to interrogate their safety data.

It is proposed that data relating to a sub-group of facilities on the Sellafield site could be cleansed then provided for analysis.

This data set would contain in the order of 20,000 safety reports generated over the last 15 years. The data will be available as comma separated variable (csv) files and therefore machine readable.

Sellafield are seeking a demonstration of AI technology and a report highlighting insights gained from the data. They expect this activity to take up to 3 months.

Benefits to Sellafield Ltd

  • Improved analysis of safety reports and prediction of issues.
  • Reduced time spent reviewing data, enabling skilled Health, Safety and Environmental professionals to focus on value added mitigation activities.
  • Ability to obtain real time insights and quality checks of safety reports to collect information as it is generated.
  • Better search capability and knowledge feed, which is accessible to non-Health, Safety and Environmental professionals.
  • Opportunity to integrate into Sellafield’s GIS (Geographic Information System) tools to provide another view of the data.


  • Sellafield can only provide a subset of cleansed data for the trial.
  • Sellafield need to ensure that the data is censored to ensure safety and security.
  • General Data Protection Regulation issues must be addressed; cleansing of data will remove names and locations replacing them with a unique code kept by Sellafield.

Functional Requirements

Essential Capability

The solution must be able to:

  • Process and analyse natural language data.
  • Provide real time trends and insights.
  • Integrate with SharePoint and offer improved search and query functions.
  • Integrate data and insight into Sellafield’s geographic mapping tool.
  • Deliver insight which as a minimum complies with best practice in Safety Reporting as defined by the Health and Safety Executive.
  • Provide interrogable decisions.
  • Run safely and securely on Sellafield’s IT network.
  • Facilitate a stepwise transition to AI, allowing increasing levels of autonomy as Sellafield’s confidence in the technology grows.

Desirable Capability

It would be desirable if the solution could:

  • Address other data, such as asset management condition and programme stress.
  • Be scalable to site and across the Nuclear Decommissioning Authority’s estate.
  • Provide a real-time quality check of safety reports.
  • Connect with external safety related knowledge bases to provide commentary on trends.


Funding Available

Proof of Concept funding is available through the Game Changers Innovation Programme for new technologies which may aid Sellafield in their mission and which demonstrate commercial potential for the innovator.