Skip to main content

Detecting faults in Fin Fans from acoustic sensors

Team: Dr. Kristofer McKee, Dr. Amir Amin, Prof. Ian Howard, Jack Wiltshire(Cisco), Nathan Jombwe (Cisco)
CIC specialist: Shiv Meka

Fin Fans are heavily used to air-cool liquid temperatures in large gas plants, and refineries. They are manually inspected for various faults on a regular basis. Timely prediction, and detection of faults in these Fin Fans is pivotal for overall productivity of the rig. Working on a project for a large Australian petroleum exploration and production company, the team was able to design an embedded controller that demonstrates the detection of various fault types with very few sensors. In the process, an algorithm was also developed in-house to predict failures based on assessment of the frequency content of the signals at a fractional computational cost. While the embedded controller was structurally designed to be energy efficient, the team in conjunction with Cisco, was able to integrate low-power LoRAWAN module to extend the range of communication of the controller’s health metrics.