Subsea pipeline leak inspection by autonomous underwater vehicle

Abstract

Faced with the difficulty in finding and locating leakage points of submarine pipelines timely as well as the high cost of pipeline routine inspection, a scheme using autonomous underwater vehicle (AUV) equipped with multibeam echo sounder (MBES) and forward looking sonar (FLS) for automatic inspection of submarine pipelines was developed. In the process of inspection, the AUV maintains autonomous navigation along the pre-set pipeline route at a fixed height above the pipeline, while the MBES collects water column images. After extracting the outline characteristics of gas-filled bubbles and making a leakage risk judgment, it rises to the sea surface to issue an alarm to the shore-based command center via satellite. The results of sea trial verify the effectiveness of sub-sea pipeline leak detection and real-time obstacle avoidance of the vehicle proposed in this study. Additionally, a variable buoyancy system (VBS) is adopted, which enhances the navigation efficiency of the vehicle. On the other hand, considering the complex operation environment of the pipeline inspection, the online collision avoidance is indispensable. The traditional image segmentation algorithm is not ideal for the high noise and bottom reverberation in operation sea area. Based on the traditional algorithm, an improved Otsu algorithm is proposed to improve the operation speed and denoise effect. As a result, an improved Otsu algorithm is proposed to accurately identify obstacles. Moreover, Kalman filtering is also introduced to estimate dynamic obstacles.