Case study: How National Grid uses analytics to keeps the lights on

Scheduled maintenance

Equipment failure leads to electricity blackouts, so "we have to manage maintenance better", says Salisbury. 

National Grid is working with IBM (see Conditional predictive maintenance, below) to develop a strategic asset management programme that Salisbury says will help it understand the health of the electricity network in near-real time. "We will look at reliability across the network, allowing us to base maintenance on load or criticality," he adds.

The new system is not replacing National Grid's existing approach and systems for maintaining strategic assets, says Salisbury. "It is very specific to load-related conditions where you can measure the condition of equipment in real time."

The new strategic maintenance programme will combine historical data from certain types of assets, and add condition monitoring to check voltages and the operational temperature of kit. National Grid plans to use thermal imaging cameras to collect heat patterns of substations, which can be correlated to identify likely failures.

Analysis of photographic data will also be used to check pylons, says Salisbury. "Instead of inspecting steelwork on a pylon, we can now can fly a helicopter or drone to take high-definition pictures remotely, which we can use to monitor the pylon's condition over time."

A cultural shift

Salisbury adds: "The tool allows us to gather the data to make decisions, creating a dynamic maintenance programme."

This new approach to asset maintenance represents a major change for National Grid, he says. "From a cultural perspective, the workforce is used to fixing things at a certain time."

With 1,500 field workers, National Grid's traditional approach required planning and scheduling, but the new system introduces agility, where maintenance is carried out based on the data collected. 

"If you have a fixed timetable, it is easier to run a maintenance programme than if you need to be more agile," says Salisbury, who admits the business change aspect is enormous. "We are changing the way people have traditionally worked in the industry," he adds.

"It also changes the way we think about IT projects."

Because it is a business change, the project needs to run beyond the lifetime of a traditional IT project, says Salisbury. "We have to think beyond the time the system goes live. We have a whole load of business change that may take two years, so we will need to build a business change project when the IT project finishes." 

As such, Salisbury says National Grid will use Prince 2 around the business change.

Conditional, predictive maintenance

National Grid is one of the first customers of IBM Insights Foundation for Energy, a cloud-based analytics platform for energy and utility companies. IBM says the product helps utilities to remove the financial barriers to using analytics across energy generation, transmission and distribution while delivering improved situational awareness from large volumes of data.

The system enables National Grid to use condition-based preventive maintenance without having to replace its existing enterprise asset management system. And as part of a regulatory commitment to reduce costs, National Grid is benefiting from analytics-based asset management to reduce operating expenditure while maintaining grid reliability and availability.

"This is a new application which sits on top of asset management and gives us a brand new capability," says Simon Baker, head of solution delivery at National Grid.

Baker said the project’s impact on internal IT was "light", both from a people and IT infrastructure perspective. "We are buying a service from IBM and delivering a unique cloud system. This has enabled us to start very quickly without traditional IT [infrastructure] spent on getting business value."

Data is collected primarily from National Grid’s existing asset management system.