Speaking at the 2020 LBCG Pipeline leak detection congress in Houston earlier this year, Barrett Walker presented Cheniere Energy’s drone-based approach to pipeline surveillance. A drone-mounted Velodyne LIDAR detector provides orthomosaic mapping with ground control points. DroneDeploy software extracts features such as buildings and equipment. Such surveying is rapidly becoming a viable solution for fugitive gas emissions as the software improves quantitative measurement. There is a wide range of sensors available, ICI OGI Inspector, www.infraredcameras.com, Mirage HC Optical Gas Imaging (OGI) camera, tunable diode laser absorption spectrometer (TDLAS). Moreover, proposed changes to the EPA air emission measurement center’s Method 21 are making these types of sensors ‘more viable and cost effective’. Some can even detect leakage from buried pipe. Infrared thermo-photography shows leaks vividly! Walker strongly advocates putting drones to work across your business as ‘unmanned systems impact every department!’ But a word of warning, you need to be aware of the evolving regulations of the dronosphere and work with law enforcement to know the process and contacts in the event of an incident. One useful resource is the API Guide to Developing a UAS Program in the oil and gas industry.
Chris Minto reported on 2019 trials of OptaSense’s optical fiber-based DAS technology carried out in conjunction with Brazil’s Centro de Tecnologia em Dutos (CTDUT) pipeline testing facility in Rio de Janeiro. The facility’s closed-loop pipeline allows engineers to test leak detection systems in a realistic situation. The trials have validated OptaSense’s leak detection technology which was shown to detect leaks in under 0.1 seconds 0 10s. Sensitivity was good too, at 15 litres/minute. Minto concluded that validation of external leak detection systems is only possible with representative tests at full scale and that the CTDUT site is an excellent location with a full-bore line and an existing fibre optic cable. Watch the movie.
Eric Bergeron presented FlyScan’s airborne UV spectroscopy which fills the gap between cheap, low sensitivity SCADA or eyeballed air patrols, and high end, expensive fiber/pigs/DAS or satellites. Bergeron cited a probability of failure vs. cost of failure analysis from pipeline risk consultants WKM Consulting. This showed, from an analysis of PHMSA reportable incidents, that the cost of failure amounts to some $5,000 per leaked barrel. FlyScan’s long term vision is to provide a right of way air patrol leveraging artificial intelligence, imaging and laser technology to fulfil inspection requirements as per 49 CFR Part 192, 195. Bergeron referred to Kent Muhulbauer’s classic, ‘Pipeline risk assessment, the definitive approach’.
Heath Spidle from the Southwest Research Institute in San Antonio showed how machine learning has been applied to fiber optic and Lidar data. ML has been used on data from MWIR OGI cameras to ‘autonomously and reliably’ detect methane with low false alarm rates. The approach is now being adapted for aerial (drone based) detection. ML has also been successful in detecting crude oil sheen on multiple surfaces including water. Currently, ‘detecting leaks from liquid pipelines [using distributed temperature sensing] continues to pose a significant challenge’ and ‘small leaks are of particular concern’. SWRI has investigated several machine learning and deep learning techniques and has come up with a SWRI ML algorithm that is good at large and small leak detection with zero false positives. Spidle concluded that ‘ML can find the hard to find leaks.’
Jay Almlie manages iPIPE, the intelligent pipeline integrity program at the US Energy & Environmental Research Center (EERC). The program scouts for new technology to be co-funded by iPIPE. iPIPE then organizes trials and demonstrations on live pipelines. AN iPIPE success is the Satelytics projects, an ongoing trial of leak detection algorithms working with hyperspectral satellite data.
John Hull (HiFi Engineering) presented the results of field trials of distributed fiber optic sensing* (DFOS) on a pipeline conducted for ExxonMobil. DFOS is an emerging technology for real-time pipeline monitoring. In particular, a new class of DFOS, ‘high fidelity sensing’ has been specifically designed to directly sense low volume leaks. In a 90-day DFOS field trial for ExxonMobil, on a West Texas pipeline, water and nitrogen were used to simulate liquid and gas leaks. In a blind test, the system detected 118 out of 134 simulated leaks, with zero false positives. The system was sensitive down to 200 psi, at low flow rates. Ongoing algorithm improvements and sensor proximity are expected to further improve leak detection accuracy
More from the conference home page.
* DFOS, distributed fiber optic sensing, also called distributed acoustic sensing (DAS) involves observing backscattered light in a fiber optic cable. By accurate time-gating of the returned light, continuous measurements of sound, temperature and stresses along the fiber can be made.
© Oil IT Journal - all rights reserved.