Researchers from the Massachusetts Institute of Technology (MIT) has developed a “zero missing rate” algorithm that can detect possible faulty components in a power grid.
Blackouts are major concerns for power suppliers especially during the summer months. One small component failure won’t likely knock out power for everyone on the grid, but two or more failures can lead a domino effect that may lead to a city-wide blackout. To ensure that the grid is up and running at all time, maintenance crews must have the right tools to detect possible component failures ahead of time, and the algorithm developed by the MIT researchers may provide just that.
According to the MIT researchers, in a test conducted using their algorithm and data from a mid-sized power grid, the researchers were able to find and weed out 99 percent of failures, and deeming them relatively safe, with the remaining 1 percent being the pairs of failures that would likely cause large blackouts if remained unchecked.
“We have this very significant acceleration in the computing time of the process,” says Konstantin Turitsyn, co-developer of the algorithm. “This algorithm can be used to update what are the events — in real time — that are the most dangerous.”
Turitsyn plans on testing his algorithm out on an even larger scale power grid with 100,000 components, and up to five billion pairs of failures.