When one of the European largest energy utilities wanted to improve the reliability of their grid by eliminating partial discharges in their transformers, they turned to artificial inteligence and machine learning specialist Neuron Soundware.
Neuron Soundware uses acoustic emissions and other physical parameters of machines, such as temperature, pressure, electric current, magnetic field, digital images and many more, to provide unparalleled diagnostics to control machines and processes.
The detection and localization of partial discharges in high voltage insulation systems is important to avoid blackouts in power networks. The early detection and localization of partial discharges in power transformers leads to much lower maintenance costs and to an increase in the reliability of power networks.
Partial discharges in power transformers emit both acoustic and electromagnetic waves. The acoustic and electromagnetic waves are in the ultrasonic (20 kHz – 1 MHz) and ultra-high frequency (UHF) ranges (300 MHz – 3 GHz), respectively.
The Neuron Soundware solution is based on utilizing acoustic sensors mounted on the outside walls of the power transformer, making acoustic detection a non-invasive technique. An artificial inteligence algorithm ensures that an acoustic signal is not contaminated by external acoustic environmental noise, such as vibrations in the transformer itself. Neuron Soundware’s acoustic-based method is not affected by electromagnetic interference in the measurement environment.
(Under the terms of the Non-Disclosure Agreement with the client and sensitivity of the technology involved, Neuron Soundware is not allowed to disclose the name of the client or describe the solution in greater detail. We will provide more information as soon as we are permitted to do so).