Machine Health | Case Study
Our client, who operate the UK’s fleet of nuclear power stations, came to us with an optimisation challenge. They needed to streamline the scope of their statutory outages, whilst reaming safe and compliant, to ensure the safety of their workforce in light of the Coronavirus pandemic.
As a consequence of the 2020 Coronavirus pandemic, our client needed to reduce the scope of their statutory outage to ensure safe social distancing during outage execution, while ensuring all equipment could run safely upon return to service. In doing this, they needed to review the maintenance schedules and maintenance history of every system across the plant. This would be a huge undertaking if carried out manually, with a typical statutory outage requiring more than 10,000 maintenance tasks and each task linked to a large, unstructured data set going back 40 years.
It was clear that the client would need a large-scale data analysis tool. Not only would this tool need to optimise the 10,000 maintenance tasks in one plant, it also needed to be compatible across multiple power stations for future outages. We sought to develop an optimisation tool-set and methodology that would apply to every nuclear power station in the fleet.
We took a two-pronged approach to reducing our customer’s statutory outage, developing two key tools. The first was the development of a natural language processing tool. This would gather all the maintenance history data for each individual task in the planned maintenance schedule. The tool then carried out an in-depth analysis of every machine’s maintenance history, helping to identify key tasks for deferral or removal from scope. This was verified in collaboration with the the station engineers.
Second, we created a data aggregation and analysis tool to enable detailed performance assessment of major plant systems. We carried out these assessments by assessing key system health parameters, allowing us to prevent any intrusive maintenance that could result in unnecessary work during the statutory outage.
The NLP tool enabled efficient extraction of key maintenance history information though processing large, unstructured data sets.
The tools were designed and developed to enable seamless integration with the client’s existing technology and business processes.
We developed a sustainable scope optimisation tool-set and methodology that could be replicated across all plants. This resulted in a ~20% reduction in maintenance work across the fleet enabling the delivery of a large statutory outage under social distancing conditions, as well as a significant cost saving of around £1 million per statutory outage. The client was very happy with the results, noting the significant reduction in resource demand and cost.
