As power grids fill renewable energy, power car charging stations and customer performance, it will become more complex and faster for their users who cannot keep up with it, a team of foreign investigators warns.
People will need help from smart machines - efficient computers that use artificial intelligence-making software systems, according to researchers at the French grid operator RTE, U.S.A. Electric Power Research Institute (EPRI) and other partners.
With the increase in low carbon selection, "grid becomes a major challenge in performance," said Jeremy Renshaw, EPRI's AI director. "Grid operators are already expanding at the end. Finding AI resources to help will be very difficult."
That ruling has been strengthened, says Renshaw, by an ongoing international competition called L2RPN ("Learning to Run a Power Network") that challenges AI developers to develop software that keeps the power grid in check.
Statistics are high: Keeping grid durability is already an edge job as operators fix weather hazards, cyber threats, and dozens of connected devices such as solar panels and smart devices. But in the energy-saving industry, AI will have to prove itself, Renshaw said. "Where will it be from five to 15 years before you see the widespread discovery," Renshaw said in an interview with E&E News.
At the moment, the race is open to creating AI programs by being able to direct future grid. The second round of the L2RPN competition last year attracted 300 entrants from around the world, including teams from the U.S., China, Russia, Colombia and Singapore. A new, expanded cycle continues.
An additional interest in the development of advanced computer decisions on the grid was evident at the recent EPRI conference which included the separation of US resources and computer software developers, EPRI said.
But entrusting computers and AI to making operational decisions that split seconds into grid emergencies is still a long way off, says Renshaw.
"There is a lot of hope that we will be ready to use that ability in five years. This is a very bad thing," he said.
"Loyalty to AI is a major problem," said Massachusetts Institute of Technology professor Koroush Shirvan, speaking at an EPRI conference in May.