Although the current electrical system has protections to isolate typical defects and inform crews, it is unable to detect these situations until they already occur. Or other times, the disturbance is not large enough to trigger an alert.
“Conscious is the saint of the management and reliability of defects,” said Hamed Valizadehhaghi, principal advisor to grid operations. “This AI system can precisely predict defects at an early stage and identify their location at a precision rate of more than 80% in the immediate vicinity, sometimes to specific equipment.”
The tool that started as a simple design on a whiteboard has slowly matured in an advanced solution not only to identify when a defect will occur, but its location and the type of defect.
Thanks to automatic learning, the tool is constantly learning and develops a catalog of defect types. He establishes knowledge to differentiate the disturbances he follows, objects striking a thread, an underground problem or a transformer on the verge of failure, conscious can make the difference.
“He examines the relationships of what should be normal currents and tensions, and whenever there is a deviation from these conditions, he feels it and will tell you” hey, something is wrong “,” said Trehan. “As we form it, he learns to save certain signatures as specific problems and deposits it in a library.”