Pilot project successfully detects Fabric Maintenance issues
Our oil and gas supermajor client in the Gulf of Mexico asked us to carry out a pilot project for their offshore fabric maintenance (FM) campaign. The idea was to apply market leading Artificial Intelligence (AI) technology to automate the detection of corrosion on their offshore structures.
The client’s ageing assets, consisting of floating production units, needed a major FM campaign that normally takes thousands of man hours. So, they asked Inspection² for help.
Mission – to save time and resources
The traditional maintenance process involves an FM engineer spending a month studying around 6,000 images to manually mark up the corrosion. Each oil rig has an average of 20,000 images, meaning that would take over four months of human analysis per rig.
By implementing AI and contextual visualisation, our market leading AI algorithm analysed the same volume of images and identified anomalies in only two days, saving our client 98% of time and significantly reducing the FM backlog.
Fabric Maintenance – a critical issue in the oil and gas sectors
Studies in the Oil & Gas industry indicate that worn equipment, rusty tanks and other evidence of wear and tear are commonly attributed to unplanned shutdowns. If these issues aren’t quickly tackled, they can lead to costly repairs or, worse still, environmentally devastating oil spills, resulting in significant economic loss.
As the manual processing and reviewing of photographic data is inefficient, the FM backlog is piling up. Worse still, the current process cannot track and predict corrosion development over the years.
Whereas our AI processes can.
According to the NACE International — The Worldwide Corrosion Authority, the global cost of corrosion is estimated to be US$2.5 trillion. Implementing corrosion management practices could save between 15-35% – around US$875 billion annually.
The Inspection² Solution
As the current process doesn’t allow for tracking and prediction of corrosion development, our client required a learning-based image recognition and visualisation tool that:
- Automates the identification and grading of corrosion based on specified criteria
- Indicates whether the identified defects are on structure or pressure component
- Measures defects and estimates coating breakdown to better define FM work scope
Our client provided us with over 6,000 images, mainly from handheld cameras, including 360-degree images. A subset of 500 images were then used to train the AI algorithms within our Machine Learning and Computer Vision technologies.
Corrosion is challenging to capture, having no defined shape, texture or colour. Which meant that good detection results and high AI accuracy were not achieved overnight. But through working closely with the client’s expert engineering team, we trained the algorithm, validated results and refined the model through various stages of the AI development.
Multiple parameters were set up to tune the algorithm to deliver better, faster accuracy, such as image decomposition, anomaly size and distance, lighting considerations, etc. After each round of training, we measured AI performance and revised parameters accordingly.
The AI analysis results were then scored on different levels:
- AI categorises images into High, Medium and Low Corrosion Severity priority groups
- AI evaluates structural components and grades corrosion in each image
Saving in data analysis time
Points of corrosion detected
Saved in inspection data analysis
Average Corrosion Detection & Classification Success Rate
Substantial savings were also made from the reduction of unnecessary painting, offshore inspection visits, onshore data management and catch up on the critical findings.
The Verdict – AI corrosion algorithm can save 3,000 man-years in Fabric Maintenance
Our client was delighted. The project successfully demonstrated the ability of AI to process and analyse data, and then detect and assess corrosion far more quickly and consistently than any human could.
Pairing AI techniques with environment and maintenance history data provided a deeper understanding of corrosion development, coating failures and other asset anomalies. Ultimately, it proved to be a powerful predictive approach to maintenance in the Oil & Gas sector.
With an estimated 10,000 rigs in the Oil & Gas sector, Inspection² AI corrosion algorithm can save 3,000 man-years in Fabric Maintenance campaigns.
That’s a lot of saved time and money!
Being problem solving technicians and engineers, Inspection² invites you to challenge us with your engineering issues. Better still, our innovative AI solutions can be white labelled for your organisation.
So, do get in touch with your complex technical issues, from industrial inspection data to processes, analysis and more. We’ll find the ideal solution for your industry sector.