We are working with the processing industry to help develop tools that will help us find the factors that lead to Loss of Containment (LoC) incidents. These happen when hazardous substances (such as gas, fuel and chemicals) escape from storage, sometimes leading to catastrophes such as:
- The devastating explosion at the port of Beirut in 2020, killing 218 people when a cargo of ammonium nitrate being stored in a hangar exploded
- The explosion at Buncefield oil refinery in 2005, when a fuel storage tank was over-filled. Thankfully no-one was killed, but the incident caused £1bn of damage to the economy
- The disastrous leak at a pesticide plant in Bhopal, India in 1984, which led to thousands of deaths.
We need to learn more about the causes of LoC events so we can help to prevent them and save lives. We’re looking into the use of natural language processing (NLP), which can search through free text to find clues to what happened in a LoC incident.
Before a catastrophic LoC occurs there have usually been ‘precursor’ events which, if picked up, might have provided early warning. These could be things like an over-filled container that didn’t trigger an alarm, or a fault that might have been picked up during an earlier maintenance routine.
When we know what these precursor events are, we can work together to find ways of stopping them leading to bigger failures, saving lives and protecting the economy. However, the complexity of the processing industries – everything from chemical processing to nuclear energy – means that digital tools aren’t enough.
Discovering Safety is asking partners in processing industries to help us develop a tool that can interrogate data, informed with human expertise and decades of real-world industrial experience.
The first stages of the project have included:
- Looking at HSE’s existing data to find clues to the causes of LoC
- Looking to our partners in the process industry worldwide to help build on this data
- Using data analysis techniques that work on unstructured data, which makes the process of providing us with data much more straightforward.
Join our members area to read more about how we are using data to achieve the aims and objectives of this project
We’re looking for industry partners to share data with us – thanks to our work with Ohalo we can do this in a fully anonymised and secure environment.
If you’d like to be part of this drive for more proactive safety, contact email@example.com.
Collaborate with us to gain unique insights, build new relationships, take the opportunity to contribute ideas and shape the industry, join a like-minded community, hear about new data analytic techniques and learn how they can positively impact your business, and use data to improve health and safety.
Data-sharing: we aim to develop a rich and diverse data pool with varied sources. Secure platforms are being developed in the early stages of the programme to enable safe, secure and anonymised data-sharing. You can contribute your data to the programme.
Industry endorsement: we want to solve real problems, which is why industry involvement and endorsement is key. The first phase of the programme involved extensive industry consultation to shape the projects. It is critical that industry remains involved and continues to influence and shape our work.
New ideas: innovation is essential to Discovering Safety and we welcome new ideas for projects, topics or tools.