Wisky - stock.adobe.com
Seadrill is bringing together safety, operations and financial data from various systems and turning these into actionable insight to enable its works to make the right decisions.
The company has an internet of things (IoT) approach, deploying data analytics collected from drill sensor monitors to create digital twins of its equipment that operate up to 3km down on the sea bed. It runs one of the most modern fleets in the offshore drilling industry, and uses industrial control systems with sensors collecting millions of data points every day.
AI, IoT, big data and advanced analytics make up the Seadrill digital platform, says Kaveh Pourteymour, vice-president and CIO at Seadrill. “The way we look at AI is that it is an element of digitisation,” he adds.
The company operates a hybrid cloud environment of which 65% is a private cloud and 35% runs on the public Azure cloud. Data sets from drilling rigs are processed locally, at the edge, using high performance computing on board the drilling ship. The data is also pushed up to the Azure cloud.
According to Pourteymour, Seadrill can see exactly which resources are going into its inventory for planned and corrective maintenance, giving the company the ability to analyse its spending and reduce costs.
“We are using Azure Machine Learning for our automated inventory optimisation. This is a capability that’s now gone live on six of our rigs, allowing us to reduce our costs by managing our stock more effectively,” he says.
Describing the journey of data from the equipment on the drilling platforms to the cloud, he says: “We have edge analytics on our rigs, which allows to interpret the data collected from the ICS [industrial control systems] which operate our critical equipment.”
Through an internal project called Plato, Pourteymour says Seadrill is bringing together IoT, edge analytics and cloud analytics. Data from the drilling rig is processed in real time using edge computing, and is also pushed up to Azure for trend analysis and fleet management.
While an element of high performance computing is done locally, at the edge, Pourteymour says: “We use the power of the Azure cloud to process larger data sets.”
Datasets from Seadrill rigs are sent via a satellite communications link up to Azure, where Seadrill is able to run trend analysis across its fleet of drilling rigs.
“Our customers want predictable results and want to understand how a quality well can be delivered. The algorithms allows us to detect anomalies in our equipment. Over time, using machine learning, the algorithm becomes smarter and can help us improve how we drill,” says Pourteymour.
The idea of a data-driven digital twin, updated in real time, is one of the aspect of Seadrill’s digital transformation strategy. “Our rigs have a lot of sensors installed and generate lots of data. One of the element of what we do is take real time data from the equipment,” says Pourteymour.
By unlocking the data, Seadrill has been able to create algorithms to build digital twins, which exist for the blow-out preventer (a piece of machinery which sits on the seabed and controls the well) and the top drive (the equipment which spins the pipe to drill the well).
“The idea is that we can proactively monitor the condition of the equipment and maintain it as and when it is needed at the most optimal and safest time,” says Pourteymour.
Once information is gathered from the equipment, Pourteymour says there is an element of edge analytics, but data is also pushed up to the Azure cloud in real time to enable Seadrill to compare equipment across its fleet. To analyse this data in real time, Seadrill is making use of Microsoft Azure’s inbuilt features and algorithms.
“If we detect an anomaly on one piece of equipment, we can determine if the issue will occur anywhere else on the fleet,” says Pourteymour.
For Pourteymour, being able to make the right decisions gives Seadrill a competitive edge. “We use Power BI and Microsoft’s Data Warehouse technologies to bring together safety, operations and financial data from various systems and turn these into actionable insight,” he says.
The other aspect of Plato is that it can be used monitor drill drivers, building a data model that can be used to help other drill drivers improve their drilling. “Every driller could be driving the equipment in a different way,” says Pourteymour. “The machine learns how you drill and can recommend how the driller should drive the machine to optimise drilling.”
Machine learning can also be used to augment the drill drivers with the right information to enable them to improve their drilling technique, Pourteymour adds.
Read more about industrial IoT
- IoT, artificial intelligence and other new technologies create a number of digital systems for the agriculture industry, providing better visibility to the farmer..
- GE Digital calls for the need to build industrial machine learning systems that are cognizant of the effects of a good and bad answer.