After a scientific panel selected a TKH Security Solutions white paper on automatic flare stack monitoring for presentation, the company demonstrated its latest innovation in automatic flare stack monitoring at this year’s Annual Technical Conference and Exhibition (ATCE).
TKH’s automatic flare detection and monitoring system with LWIR sensor technology is computer vision-based, plug-and-play and designed as a built-in part of the camera. The system is the first automatic flare stack monitoring system with flare size tracking and automatic event signaling.
The model has been optimised to perform on an embedded platform. The reliability of the proposed algorithms has been extensively tested for live flare stack monitoring in the Port of Rotterdam Botlek area, the Netherlands. Preliminary results have shown a reliable system free of false alarms.
A permanent setup for year-round testing is required to complete more comprehensive reliability tests.
In several cases, it may be beneficial to add a visible light camera, e.g. to detect black smoke. Depending on the presence of such black smoke, a dual camera setup could assist in regulating the amount of steam fed to the flare stack to increase combustion efficiency. The development of algorithms taking of dual cameras input is being considered. For applications such as black smoke detection, machine-learning techniques may also become available to learn which features from which camera are important.
With different zone shapes, the developed framework can also be utilised for other applications such as hotspot, fire or machine overheat detection. A dual camera setup could for example serve as an automatic forest fire detection system.