Through a collaboration with the university spin out Eneryield, which is developing machine learning based methods for intelligent power quality analysis, Unipower is offering a new type of functionality to give insights in the work with evolving faults. The researchers Ebrahim Balouji and Karl Bäckström from Chalmers University of Technology are founders of Eneryield and behind the solution.
In collaboration with Eneryield, Unipower provides a machine learning based surveillance module that can predict faults in power grids. The new functionality utilizes historical patterns and trends of PQ-disturbances to estimate the time to the next severe fault. The early warning generated gives the grid operator the time needed to take necessary actions to avoid the fault. It is an important step towards a proactive approach to fault avoidance, and creates conditions for better grid stability and security of supply.
Eneryield’s technology is based on the latest research within machine learning and deep neural networks, creating new opportunities for analytics. The large amount of data that Unipower has access to enables data-driven and self learning methods that will set a new level for which information that can be collected from the power grid.
Unipower offers the solution as a new functionality that can be integrated in PQ Secure.