New Algorithm To Better Predict Power Grid Failure Points

No one likes that feeling when they are comfortably working, or playing, on their computer and the electricity at their location suddenly vanishes. Power outages can affect anywhere from small areas to occurring as major blackouts. Researchers at MIT aim to help.

Usually a power outage caused by a component failure or downed power line is contained in a relatively local area, such as a neighborhood or town. The power grids are pretty resilient at handling single point failure without the whole network going down. A multi-point failure, however, can take out massive areas as was the case in 2003 when 50 million people lost power in the north eastern US and Ontario, and in 2012 when the large blackout ever to occur took out power for 700 million people in India.

Researchers at MIT have come up with a new algorithm that analyzes all of the potential failure points in a network grid and determines which pairs of failure points have the potential to take down large portions of it. Using a moderate sized power grid model containing about 3000 components, the algorithm was able to quickly prune out 99% of failure points that were considered relatively safe, and brought to attention the 1% that were likely to cascade into large blackouts if not checked.

Given that the number of potential combinations that could cause grid failures could be in the millions or billions, and the ability of the power companies to get to the locations to check them all, there will always be a chance of something going wrong. Having more tools in the arsenal to help predict these problems, however, is definitely a step in the right direction.