Distributed Sensors, Computation and AI for Automation, Protection and Maintenance of Power Grid

The advances in computers and communications in the ongoing Fourth industrial revolution have not been matched by corresponding advances in the power grid. Fifty-year-old problems of the grid remain, such as uncontrolled blackouts, lack of automation, and uncertainty about replacing aging equipment. In the Ukraine war, attacks on the power grid have drawn attention to limiting spreading blackouts. CTs and PTs for monitoring high voltage lines remain large and expensive structures. Human operators are required at substations, and much of their decisions are taken over the phone. New developments in sensors, communications, smart electronics are today allowing a smarter more automated grid. Aging transformers and generators nearing end-of-life can be better monitored and replaced. Load shedding can be smart and pre-planned, so that low-priority feeders, neighborhoods, households and equipment are disconnected first. A blackout from cascaded tripping of generators can be arrested by remote disconnection of loads equivalent to the lost generation. The new smart grid may be based on the structure of the internet, with distributed and non-centralized sensors and computations. The multiple inputs from sensors and various operation conditions are an excellent candidate for AI. Methods have been proposed here for prioritizing and shedding loads automatically and remotely before overstressed generators are tripped. A disconnection of a generator is simulated in the IEEE 36 bus, following shedding of an equivalent amount of load, so as to arrest the blackout. According to existing trends, improved smart grids will become a major growth area in the coming decades.