FOSA, the Fiber optic sensing association awarded a 2022 Innovation Award to Hifi Engineering for its ‘Deep Fake’ pipeline verification and validation methodology. Deep Fake describes Hifi’s approach to testing its deep learning models for pipeline monitoring. ML-based pipeline leak detection algorithms are trained on real, in-the-field, tests of controlled leaks and other anomalies. The new approach adds to Hifi’s database of field tested leaks using a ‘patent pending’ simulation method to generate synthetic leaks of various configurations. Leak detection systems can be tested on such synthetic data rather than on a physical product release. The system is said to successfully detect and classify leaks, enabling cost-effective system auditing that ‘supports the long-term maintenance and reliability goals of pipeline operators’. Calgary-based Hifi’s fiber optic sensing systems are deployed on over 3,000,000 meters of pipeline assets globally. More from Hifi Engineering.
Another FOSA prize went to France’s Febus Optics and UK’s Lytt
for a jointly-developed well monitoring solution. The approach combines
Febus’ distributed fiber optic sensors with Lytt’s analytics platform.
More on the solution from Febus.
Hifi also merits a marketing prize for its use of the ‘Deep Fake’ trope. A first for industrial ML?
© Oil IT Journal - all rights reserved.