Interview

The What, How and Why of Digital Twin by Kongsberg Digital

Shane McArdle, Vice President of Production, and Lars Meløe, Regional Director, from Kongsberg Digital recently spoke to Go Digital Oil & Gas about the Digital Twin. 

We talked about the latest trends, practical value of digital twin and how to avoid the most common pitfalls in adopting it. 

  • Text:Go Digital Oil&Gas Middle East

That's a good point. Digital twin has become a confusing term in the industry. The broad industry understanding is that a digital twin is a data-based representation of a physical object, whether it is a single asset, like a pump, or a large-scale asset, like an entire oil & gas facility. In Kongsberg we call ours a dynamic digital twin. The Dynamic Digital Twin can be used during the operational stage, as well as the early life-cycle of field development. It is a virtual representation and a tool for understanding and impacting every part of your facility, using massive amounts of contextualized data and powerful analytics.

The amount of data needed to generate a potent dynamic digital twin is usually so comprehensive that historical and real-time data are not sufficient. Anyone working in the energy industry knows that there are a lot of data blind spots in a process facility. When you transition from a greenfield to brownfield, those data blind spots become more significant and the quality of the data becomes more and more uncertain. So, what Kongsberg Digital has done to remedy this is to embed physical models or simulators that provide a clean, synthetic data source available within our digital twin thus offsetting that data quality problem. Our long history of providing high fidelity simulators and now implementing them in the fabric of our dynamic digital twin gives it powerful analytical, predictive, prescriptive and Hybrid AI capabilities compared to a mere visualization tool. Operators using our dynamic digital twin solution gain full visibility into the behaviour and the multilayered interdependencies among the assets, the process, and the operations.

The digital twin is helping to transform the energy landscape and it is driving efficiency gains very quickly. The benefits are being realized way beyond the initial capital outlay that is needed to deploy these types of systems through predictive maintenance, process optimization and energy consumption, to name a few. True value though is very dependent on the business outcome or the business use cases that are defined, which varies from customer to customer. However, there are two core values that can be consistently created by using a digital twin.

The first one is collaboration. Regardless of the business case, once you create a framework for a digital twin and deploy it, you're already opening for collaboration across the organization, across the asset, across user groups, across different disciplinary teams because you are breaking up silos and sharing experience and knowledge across these teams. By pulling in all of the data from the different sources, presenting it in a way that is easy to understand and is easy to access, you automatically create this collaborative environment.

The second core value is experimentation. The physical models or simulators that are made available through the twin is the closest you are going to get to the reality of the process facility without actually interfering in a real production environment. So, you now have the capability to run 'what if' scenarios.

To provide an example of how this would benefit a particular user profile - let's say, a flow assurance engineer supporting operations. Imagine, there isn't a day that goes by when someone knocks on their door and says: what would happen if I close this valve or how much MEG can I inject before I start to kill a well? And the user doesn’t have all these answers at their fingertips. However, If they had an environment, virtual testbeds, they could just go in there, spin up a model and test what will happen if this valve is closed or opened. This allows for experimentation, and you can drive innovation within this environment.

For me, these are probably the two core values that you can achieve by deploying a digital twin, outside of concrete value drivers like energy efficiency, process optimization and so on. 

Three years ago, the industry was struggling to understand what the business value was for a digital twin. So, during the last three years, we have seen it mature to the point where we are now receiving requests for information (RFIs) specifically for digital twin. That hasn't happened in previous years. This is the first time we have seen this. Were these RFIs very clear? Not in every case. But it was an attempt by the industry to try and define what they are looking for. That is really positive.

If you think back a few years ago, there was something called the AI winter in the 90s. A lot of the advanced data analytics that we are doing today was around in the 90s. The AI pioneer, Marvin Minsky, said one of the reasons AI research stalled at that time was due to computers limited sensory input and the ability to interact with the real, physical world. The thing that now sets digital twins apart is interconnectivity, eg. you are connected to sensor data streaming from IOT enabled objects, you are connected to business and asset data, that is what sets it apart from other types of systems. It is all available right down to a single asset tag name. Eroding the barriers between the physical and virtual makes artificial intelligence a lot more viable, easy to use and more embedded into the heavy asset industry. So, I see in five years that AI machine learning is going to be an integral part. It will be much easier to deploy and will provide real value in the next five years.

Secondly, if a digital twin is set up correctly by a company, it can easily scale. Starting with a single asset, like a pump, you can quickly scale that out to a system like a compressor train. Then you can scale that out to an asset like an entire oil and gas facility. Then you can start adding more assets. So a single operator can have an overview of all its assets, in one single digital twin. And then you can start comparing how the operation efficiency was working for asset A compared to asset B. So, this potential scale is very important in the next 5 years.

You need to think about what you want to achieve, whether you can do that on your own, and finally, how to get your organization on board.

To start, what you are trying to achieve has to be very much business value-oriented. Lots of companies, even operators, have connected different sources of data. Great, now you are overwhelmed with all of this data. What do you actually do with it? You have to be very clear from the get-go what you are trying to do.

Second, you need to think about your own capability and who you need to bring onboard to realize your digital twin transition. What is your company good at, is it software or is it energy? How will the technology be maintained after a digital twin is set up? Having software experts with domain expertise from the energy sector building, deploying and maintaining the digital twin will allow operators to focus on their core expertise. Also, think about lock-in.

Kongsberg are taking an open ecosystem position and try to standardize as much as possible. It is very important that our customers can take something done on one asset and deploy it to another asset easily without being limited by vendor lock/in.

Kongsberg is a Digital Twin Partner both Go Digital Oil & Gas events in Abu Dhabi and Amsterdam
Kongsberg is a Digital Twin Partner both Go Digital Oil & Gas events in Abu Dhabi and Amsterdam 

Finally, companies using digital twins need to have a strategy for internal adoption. Think, how are we going to federate knowledge across an organization, how are we going to break down silos?  I have seen that some get the tech right, they build it at scale, they have a value case, but they fail on the adoption because they have not taken the users on the journey from the very start.

What we want to get across is this message: don't be frightened. Yes, it sounds like a big system integration project, but you can deploy a digital twin project very quickly, with a minimum level of complexity and get your organization to start using it. Then, when you have the foundation in place, you build out the complexity and the capability. Once you have one digital twin up running, you should think of how to roll it out to the next five assets? Keywords: Start small, think big, but scale fast.

Shane McArdle, Vice President of Production in Kongsberg Digital Shane.mcardle@kdi.kongsberg.com

Lars Meløe, Regional Director in Kongsberg Digital Lars.meloe@kdi.kongsberg.com