Edge AI Based Flare Monitoring to Reduce Global Warming - Livestream

This video-based flare monitoring system is developed for real-time measurement of the flare gas flow rate at the edge. Depending on the desired trade-off between speed and accuracy, either an object detection (EfficientDet Dx) or instance segmentation (Mask R-CNN) model is used for real-time detection of flare and smoke instances in the input video stream. Organic and synthetic data is used to achieve high precision and recall (greater than 0.98) for both flare and smoke. The detected rectangular bounding boxes or polygon masks are used to estimate the flame size, and predict the flare gas exit velocity or equivalently flow rate. The estimated flow rate is within +/- 10% of a reference flow meter. The Deep Learning models are “edgified” in order to shrink the size and improve the inference speed by ~3X on small footprint edge devices. For three stacks in the camera’s field of view, the error increases to about +/- 25% of full scale due to the decreased resolution or increased scaling factor (~2.5X). The system components (camera and computing device) can be configured to obtain the desired tradeoff between cost, accuracy and measurement speed. This software product "Flare Advanced" has been deployed. The prior product "Flare Basic" has been deployed multiple times. You can not manage what you do not measure, this product helps to measure global warming emissions in real-time, sending alerts via email or SMS.

Speakers: Greg Makowski and Himanshu Goyal