Home » Articles » How Can AI Minimize Critical Drilling Rig Failures?

How Can AI Minimize Critical Drilling Rig Failures?

Offshore drilling is an expensive business. Downtime can make those costs balloon exponentially. According to research firm Kimberlite, the average offshore drilling rig experiences 27 days of unplanned downtime annually, resulting in about $38 million in losses. Companies are beginning to look to rig maintenance powered by artificial intelligence to drive down these costs. 

Small maintenance challenges have been solved through redundancy. For example, there are typically three mud pumps on a given rig. If one of them should fail, it’s inconvenient, but not catastrophic. This backup approach applies as well to generators, pumps, and many other types of equipment.

But what about top drive breakdowns that cannot typically be resolved through redundancy? Top drives are critical pieces of equipment that reduce the manual labor involved in drilling. Their million-dollar-plus price tags make them far too expensive to allow spares to be kept on-hand, which adds further delays for sourcing and procurement when the need arises. Failure here stops the entire drilling operation, resulting in potentially massive ongoing cost and lost revenue until repairs can be accomplished. In a worst-case scenario, top drive failure can be so catastrophic as to compromise worker safety

 

The cost of critical drilling rig asset failure

The daily maintenance and operation of a drilling rig can cost anywhere from $50,000 a day for an onshore rig to $1million per day or more for offshore platforms. For a high-end offshore rig, just two days of downtime can mean losses of $2 million or more, not factoring in labor and foregone revenue from lost production. All told, a top drive failure on an offshore drilling rig that requires a week to repair could end up costing nearly $8.5 million in total costs and lost revenue.

When oil and gas E&P companies suffer major drilling incidents, public faith in the company is often damaged. Company brand and image are historically difficult to value, but if a company suffers a failure of the sort described, it may well endure public relations consequences for years thereafter.

 

Improve predictive maintenance artificial intelligence

How can rig operators ensure they are for the failure of any critical E&P asset? The answer lies in predictive and prescriptive analytics, an AI-powered technology that allows companies to flag when maintenance or repairs are needed ahead of failures. Workers can then perform maintenance while the equipment is not in service, enhancing cost effectiveness and employee safety. 

The U.S. Department of Energy has found that predictive maintenance yields an annual average of $34 million in cost savings. According to the organization:

“Updating maintenance practices to more predictive efforts—driven by digital technologies and data-based optimization—can enable offshore production facilities to reduce their unplanned downtime and drive better operational efficiency.”

Traditional approaches to detecting and predicting failures rely on pre-programmed rules and physics-based models, all painstakingly worked out by human analysts. This process takes a great deal of time and labor, and the resulting model can be compromised if something changes about the routine operation of the rig equipment being monitored.

Predictive analytics employs machine learning algorithms to forecast impending drilling asset failures using unsupervised learning models that collect and analyze data from sensors on the rig, raising alerts of impending failures far in advance, and significantly decreasing asset downtime.

An AI-powered predictive analytics solution can reduce resource requirements dramatically, saving time and money, while also allowing the resulting model to automatically adapt to operating changes. Predictive maintenance solutions are built on Normal Behavior Modeling (NBM) technology; they continuously refine their operation, allowing  patterns to be learned and failures to be proactively identified, even for never-before-seen failure modes. This flexibility allows the model to identify unusual or so-called “edge” cases that a rules-based approach cannot.

Predictive maintenance improves failure identification and maintenance processes immensely. Relying on calendar-driven routine maintenance, after all, carries its own inherent risks. Working on equipment can cause additional problems to occur while trying to repair the original issue. In a perfect world, we would only perform maintenance when it’s actually required. Using AI-powered predictive maintenance makes it possible to know exactly when this work should take place.

 

Click HERE to learn more about Avathon’s AI-powered oil and gas solutions.

Share the Post:

Related Posts