Abstract
The incumbent need of sustainable and cost-effective solutions for subsea inspections of live assets installed on the seabed along with the increasing complexity of next generation subsea processing (e.g., boosting, separation, water treatment, desulfation), are some of the drivers pushing the development of subsea robotics solutions. Several technological challenges have to be faced to achieve this goal, with one of the most demanding requirements being the capability to switch autonomously from a flowline horizontal inspection to a vertical riser. In general, the navigation capabilities must be able to seamlessly move from a 2D target to a 3D body with no human intervention in the loop. Many subsea drones currently being launched to the market are able to automatically track a pipeline, a flowline or a cable/umbilical (either exposed or partially buried), but a true challenge is represented by the vertical transition, i.e. when these assets turn into risers to connect the seafloor with the topside facilities. The required capability to manage this transition in pipeline configuration applies to simple catenaries as well as to lazy-wave catenary patterns. A subsea drone with comprehensive capabilities should then be able to track the riser throughout the water column following pre-set inspection shapes namely a one-side (180° domain) or a two-sides (360° domain) GVI or CVI mission. Data coming from stereo-cameras and laser scanners are digitalized, merged and processed by the navigation algorithms supported by AI based computing tools in order to ensure the pre-determined inspection pattern is complied with. Transition from a DVL based tracking to a pure video-referenced navigation has been one of the most demanding features to be dealt with. Ultimately, the drone should be able to hold the position automatically when the algorithms indicate the detection of anomalies like corrosion, torn parts, scratching and the like thus increasing the resolution of data acquisition relevant to the identified anomaly. High-resolution image sets are taken during the inspection and exported into a Survey Report for Client's record and use within its own Life-of-Field master plan. The riser tracking features and, in particular, the management of the transition from the seabed flowline to the riser linking the production system to topside is extremely challenging since it implies a smooth change in how the subsea drone navigation system guides the vehicle itself. This paper presents how information coming from different sources is managed and processed by the subsea drone AI to safely guide it while performing an autonomous subsea riser inspection. Results obtained from a deepwater pilot project offshore Brazil by Saipem with its FlatFish, developed in a JIP with Shell Brasil, PETROBRAS following ANP Levy regulations, in partnership with SENAI CIMATEC, are presented too.
