An Autonomous Unmanned Aerial System Inspection Platform for High-Efficiency 3D Pipeline/Route Modeling/Change-Detection and Gas Leak Detection-Localization

Overview

Main Objective

To enhance the quality and efficiency of UAS pipeline and route inspections, and to evaluate and enhance the performance of OGI pipeline leak detection-localization

Public Abstract

There has been growing interest in the use of unmanned aerial systems (UAS) in oil/gas pipeline/tank inspections. While UAS can substantially extend pipeline operators' view range, the quality and efficiency of UAS data collection are significantly constrained by line-of-sight difficulties and the 3D shape of the inspected bodies. Meanwhile, UAS based thermography has also been used in leak and insulation damage detections with the aid of infrared cameras. While studies have indicated that the standoff distance between the infrared camera and the damaged pipe or tank spots significantly affects the performance of leak detectability, it is a demanding task to manually control and fly UAS at consistent close distances along the 3D surfaces of the pipelines or tanks to be inspected. Furthermore, there is a lack of technology to automate the gas leak detection based on the infrared images. In this project, an autonomous UAS inspection platform is proposed to address these line-of-sight and complex 3D surface-follow issues. The developed platform can be used for inspecting both oil and gas pipelines and tanks. The platform has four primary modules, each designed for a key system function: (1) PIDMIM for inspection data management, integration and visualization; (2) CPP for autonomous 3D UAS inspectionpath planning and control; (3) PCIQ for identification and quantification of 3D profile changes of pipeline and/or route; (4) PLDM for automated gas leak detection and localization. The proposed platform can be used in numerous areas of interest relevant to the Office of Pipeline Safety such as identifying geological impacts, right-of-way conditions, rivercrossing changes, land movement, third-party excavations, and pipeline mapping verification. The autonomous leak detection and localization function of the platform is carried out through a developed deep-learning-based image processing using optical gas imaging (OGI) as the input image source. The detected leak spots observed by the onboard infrared and visual cameras will be automatically registered on the 3D pipe model to pinpoint the potential leak locations on the pipes or tanks. UAS-based OGI can also be used to evaluate the impact of oil spills as necessary. The proposed autonomous UAS inspection and leak detection/localization platform is anticipated to significantly enhance pipeline safety, and to transform pipeline inspection and maintenance operations.

Potential Impact on Safety: The developed platform can be used for inspecting both oil and gas pipelines and tanks.

Summary and Conclusions

The project developed an end-to-end unmanned aerial systems (UAS) pipeline visual inspection framework to identify and document pipeline's visual defects, geo-hazards, and gas leakage, encroachments on right of ways. The framework contains three key components: 1) the novel UAS coverage path planning (CPP) algorithms, 2) the infrared thermograph (IRT)-based gas leakage detection method, and 3) the 3D model-based inspection data management application. Two novel UAS coverage path planning (CPP) algorithms were developed to allow accurate, low-cost and efficient photogrammetry 3D reconstruction of complex structures of pipelines and the associated structures. The developed CCP algorithms also allow close-range UAS IRT imaging of the pipelines. The capability of the close-range IRT detection of pipeline gas leakage can only be achieved through the developed 3D CPP that controls the distance of the IRT camera to the inspected pipelines and detect the leaking spots within the IRT cameras' best effective range. The test results from both lab and field experiments indicated that accurate 3D structures in pipeline facilities can be reconstructed using the developed algorithms. The performance of the 3D reconstruction accuracy is comparable to traditional laser scanning method but takes much less time and has much less access constraints compared to the time-consuming stationary terrestrial laser scanning methods. The developed inspection method can be applied to detect encroachment, river scour sediment or soil movement if the change is greater than 5 centimeters. It can be also used to conduct tank settlement evaluations, such as factors of out-of-plane, body-tilting, edge settlement, and other body deformations when the body changes of geometry dimensions greater than 2 centimeters. IRT-based gas leakage detection tests were conducted in outdoor settings with different wind speed and gas leaking rates. The detecting results demonstrated that IRT video clips from high-performance IRT cameras can be used to detect gas leakages under even adversary environmental conditions such as strong wind condition, when the video clips were recorded within cameras' effective range, which can only be achieved through the autonomous approach when scanning long distance pipelines in relatively short time. With the UAS mounted infrared camera, the developed infrared-based leakage detection method is well-suited for detecting gas leakage and local dispersion from the tank. Given the developed method is based on the temperature change and contrast against ambient environment, it should be able to apply to any type of gas leakages involving temperature contrast. Since FLIR A8303sc camera can record full-frame images (1280x720) at 60 fps, it is a good fit for using this detection method from a fast-moving platform like a rotary UAS that was tested in this project.

Relevant Files & Links