Risks of a self-automated nuclear power station

24 May 2017

In May EDF announced it would begin to use artificial intelligence (AI) technology to make its nuclear power stations more efficient and help reduce the customer’s home energy consumption.

The power sector has traditionally been slow to evolve with such innovations like AI surrounding the industry. The very necessity of electricity for every household has allowed the power sector to evolve and adapt to change much slower than many other sectors. Until now changes in innovation, consumer expectations and regulatory pressures are challenging what power companies need to do in order to survive.

Many companies are now looking at how they can use digital innovation to enhance the customer experience but also help improve already very tight margins in their business.

David Ferguson, Head of Digital Innovation at EDF, announced at the Artificial Intelligence Summit in London that they would use AI for predictive maintenance, where it will take data from sensors placed around the plant to identify and outline the reason as to why a component has failed. Moreover, AI could monitor the various systems in a power station and provide real time advice telling plant managers what to do now and in the future. 

Such technological advances will bring significant change, helping to mitigate risk in some instances and increasing it in others. A particular risk to consider is cyber risk. As an emerging risk, cyber lacks volume of data on exposure and claims. In addition, many aspects of cyber risk have yet to play out – for example, if an automated nuclear power station is hacked and this causes an accident, liability is far from clear.

Insurers’ claims experience is slowly improving, but the market lacks experience and understanding of large catastrophic or systemic cyber losses – such as a hack against the power grid. The power industry is not immune to attacks. For example, the Stuxnet attack, which delivered a worm through a worker’s thumb drive not only was designed to extract the blueprint of the Natanz plant in Iran, but also tried to destroy the centrifuges.

Insurers are trying to develop modelling tools to better understand these types of cyber exposures. For example, as a first step, insurers are collaborating with catastrophe modelling firms to agree common core data requirements for cyber models.

When designing digital pilot projects within power companies it is important to discuss possible risk exposures with your insurance broker. Brokers need to work alongside clients to use scenario-based modelling to improve understanding of cyber exposures. The brokers need to give clients access to new cyber risk models that will enable clients to quantify their cyber exposures and model the effectiveness of various (re)insurance structures.

The broker of tomorrow will be much more on hand to help with risk prevention and mitigation, rather than just simply placing a policy.

For further information, please contact Gemma Claase, Partner on +44 20 7528 4129 or email gemma_claase@jltgroup.com

contact Gemma Claase
Power Team