Digital Energy Journal- DNV GL - incorporating probability in digital twins

Join | Existing Member Login

You are Home » News » View Article

DNV GL - incorporating probability in digital twins

Thursday, April 2, 2020

DNV GL has announced a concept of a 'probabilistic digital twin', which can take the probability of risk factors into account, when modelling live operations.

The risk models which oil and gas companies make are usually done as static 'desk' studies, having a snapshot picture of risk. But in the real world, risk is continually changing - so risk models would be more useful if they were continually changing too.

Meanwhile, oil companies are making more use of 'digital twins' - digital models of physical objects - which provide digital information about how the physical object is behaving, using real time data from the physical asset. But these digital twins are often designed as though everything is working perfectly.

To resolve both of these problems, DNV GL is developing a concept for a 'probabilistic digital twin', which uses a digital twin combined with real time data from the physical asset, together with continually updated risk models.

This would enable operators to answer questions based on the changing risk levels. For example, they could determine if the failure of a certain component, were it to happen, would put the risk of the overall system above a certain threshold, given the current operating conditions.

The higher resolution you have of your changing risks as part of your ongoing operations, the less precaution you need to take, says Frank Børre Pedersen, vice president of oil and gas technology and research at DNV GL Group.

This is analogous to how banks are allowed to keep less cash reserves if they can demonstrate they are being more precise in how they manage their credit risks, he says. They know they can operate at a thinner margin without compromising safety.

A probabilistic digital twin might include models for how steel degrades over time, or reliability of equipment declines.

Risks models can also be developed for cyber physical systems, working out ways that a system could fail or be hacked.

The commercial model could be that DNV GL provides expert knowledge on the various failure nodes of the equipment which could be used by a client to populate a risk model on its equipment or systems.

Alternatively, DNV GL could develop the risk models itself, creating digital twin 'platforms' which can be used as a basis for building a model for your own equipment.

It is possible that risk models could be shared. It may make sense for multiple wells to have different digital twins, because they have different configurations, but the basic risks to them are the same, so they can share a risk model.

Associated Companies
» DNV GL

comments powered by Disqus

CREATE A FREE MEMBERSHIP

To attend our free events, receive our newsletter, and receive the free colour Digital Energy Journal.