There’s a perception that not enough use is made of the data collected as part of offshore oil and gas operators’ subsea integrity management programs. What is collected is also limited in scope and quality. But, that could all be set to change. Data collection and use – from measurement and imaging tools, to the equipment used to carry them and the systems used to process the data collected – is entering a new era.
These issues were central to the joint Society of Underwater Technology, International Marine Contractors Association and The Hydrographic Society of Scotland seminar, The Leading Edge of Value-Based Subsea Inspection, held in Aberdeen late 2017.
It’s good timing. There’s a growing need for efficient inspection systems, to help operators understand the condition of their subsea infrastructure and therefore efficiently maintain it.
In the UK North Sea alone, for example, BP has 4500km of pipeline, 80 riser systems, 270 subsea trees, 92 manifolds, and a plethora of umbilicals, “which we need to understand,” Scott Higgins, BP, told the joint seminar. Inspections have been done the same way for 30 years, but, this is changing, he says, in terms of the inspection technology used, what information is gathered and when.
There’s an increasing focus on having an integrated planning process for inspections, he says, bringing together various departments and specialists, from subject matter experts and surveyors to environmental, riser and pipeline people, as well as structure and hulls people, etc., all working together to see what information is needed and when.
“This is about increasing the efficiency of data collection,” Higgins says. “Why we want it, what we need it for. Historically, we relied on a good deal of information from Work Class remote operated vehicles (ROVs). How many years of video we have would scare you. But, what do we actually need? What sensors do we want? What vehicles? Can we use integrated laser/imaging, field gradient cathodic protection inspection systems, alternative fast ROVs or autonomous underwater vehicles (AUV)?”
According to Global Marine Technology Trends 2030, a report by Lloyds and Qinietiq, by 2020, more of this work will be using unmanned platforms, including those deployed from shore, as well as resident vehicles, with more focus on what is data needed, faster interpretation and machine learning, says Scott. Indeed, just recently, Saudi Aramco announced a new AUV designed for offshore platform debris, pipeline and other surveys, which would be deployed from shore.
A range of other solutions were also discussed at the joint seminar. A large focus was on data gathering tools and live data processing.
Subsea laser scanning and photogrammetry techniques have been making major in-roads in the industry, with claims such techniques can offer high levels of measurement accuracy. The attraction of some of these systems is that they’re being offered without having to baseline surveys – i.e. put in markers or a system of beacons, to establish measurement reference points.
Within a short period of time, Cathx has become a household name in subsea imaging. The firm was founded by Adrian Boyle, its CEO, in 2009. It uses photogrammetry, but combines stills images with laser lines and is now developing machine vision systems to automate analysis of the collected data, e.g. automated eventing and measurement.
The firm built its own camera which takes 2 millisecond exposure images (i.e. 30 HD stills a second) in conjunction with pulsed lighting, so there are no blurs (i.e. from moving particles in water) on images and surveys can be done faster (i.e. instead of an ROV surveying a pipeline at 0.5knots it can be done at 5knots, with the ability to extract HD stills with 0.8-1.5mm resolution), even at <5m from the pipeline, says Boyle. The images are then built into 2D mosaics.
Cathx then uses other data (co-registering) to build and add accuracy to its models – i.e. laser line data, from which point clouds can be built, and time-stamp data. Because the position of the object is known relative to the ROV, navigation correction can be calculated.
An AUV data acquisition package would comprise strobe lights, UHD and HD high resolution cameras, taking pictures at seven frames per second for UHD and up to 30 for HD, with the laser and lighting synchronized, allowing for 4knot co-registered data acquisition speed, says Boyle. Using dual cameras, imaging the same scene, can further allow post-acquisition calibration for environmental factors (salinity etc.).
Dual source data acquisition is the basis for automation, but it requires real-time data integrity checks, and 3D laser data plus co-registered image data, says Boyle.
In 2018, the firm is going to be developing automated eventing using these technologies. This will allow large volumes of data to be reduced to events, quickly. This will, through range-based statistical analysis of 3D laser data, produce images and 3D data that can be reviewed efficiently because it’s based on the laser data (not memory-heavy photographic files). Once an event is picked up in the laser data, the corresponding photographic image can be found easily, because the data is co-registered using the time stamps.
Boyle says cross sectional geometrical analysis, freespans and circularity will be possible with 3D machine vision techniques, and are being tested in Q1 this year. This will include building a library of types of event that should be automatically detected, which will be combined to enable machine learning.
Comex presented its ORUS 3D system. It’s a subsea optical system for measuring and then creating high-resolution 3D models of subsea structures, without the need for an inertial navigation system. Bertrand Chemisky, Director of Innovation at Comex, says the system uses triangulation to create a 3D reconstruction.
The system comprises an integrated beam of tri-focal sensors, with four wide beam LED flash units, plus a data acquisition and processing unit, which fits on to an ROV for free-flying data acquisition and initial processing. Scanning works at a >40cm from structures, with the best resolution reconstruction from 1-2m from the object, says Chemisky. Thousands of images are taken, using strobe flashes, resulting in several millions of points in a square meter, the firm says.
The data goes through an initial processing real-time on the ROV to assess location and quality, before on site (on-board the support vessel) processing, to further quality check the data collected and create an initial scaled 3D model to cm accuracy. Final processing of the data, which collected as point cloud data, will then be carried out to reconstruct the site or object in a 3D model to mm accuracy.
Comex claims up to sub millimetric precision, with 0.1mm resolution images, depending on the survey, i.e. for a spool metrology survey, accuracy will drop over longer distances. Still, Chemisky says over 64m, with seven hours processing, just 2.4cm deviation was seen. For a 27m long survey, for a jumper metrology, survey data was processed over five hours to 1.8cm accuracy. An anchor inspection off Marseille, France, took 20 minutes processing to gain millimeter accuracy, thanks to being close to the chain.
Further algorithm optimization is ongoing, to improve reliability and validate repeatability, says Chemisky. The system is going through qualification with Bureau Veritas.
A small but fast-growing firm in subsea data collection space is ASV Global. It had built 95, up to 13m-long autonomous surface vehicles (ASVs), as at early November 2017.
ASV has been working on the Autonomous Surface and Sub-Surface Survey System (ASSS) project, in which long endurance ASVs provide locational and communication and control support for a side scan sonar survey AUV (in trials, the National Oceanography Centre’s Autosub long range AUV has been used), via optical communications systems, with both able to be deployed from shore.
As an extension to this project, the firm is also working on a concept for an Autonomous Pipeline Survey system, which would use the technology developed in the ASSSS program used for pipeline surveys, completely removing the need for manned ships to be involved in pipeline survey. But, they want to take this a step further and instead of just gathering side scan sonar data, create a system that could offer automated eventing, i.e. when it spots a fault or specific target, it sends an alert, real-time, which can then be acted on – instead of having to have someone assess the survey data. ASV is looking to set out a road map for what needs to be done to make this happen the gauge operator interest.
ASV is also involved in the ARISE (Autonomous Robotic intervention System for Extreme Environments) project, which is looking to put a remote operated work class ROV on an unmanned vessel. “To put an ROV in the water still needs a big ship. How do we take that from being on a big ship to a little boat with new people on board?” says Cowles. A number of tests have been done and it’s no easy task, he says. So, ASV is starting small, adding an observation class ROV to a C-Worker 7 (a 7m-long unmanned vessel). It will use it to trail things like inspecting a mooring chain, testing latency, communications and autonomy. “There are challenges around communication, latency, bandwidth, how to control both vehicles together, stopping the tether getting caught in a prop,” says Cowles. “The goal is to build a 12m boat with a work class ROV on board. That reality is still a couple of years away.”
More to come
This is an emerging field. It’s giving the industry tools that it could only have dreamed of in the 1970s, when photos were taken by divers on film and had to be developed on board. “Now photogrammetry can be done on the fly,” says Peter Blake, subsea systems manager, Chevron Energy Technology, at the
event. In the past, the industry has been good at gathering data but not so much gaining information, he says. The tools are being developed to change that.