Digitalization, digital twins and cyber-physical systems…what’s in it for us?
In a recent article, I wrote a bit about technology waves and the difficulty we experience when we have to decide on jumping onto a wave or just letting it pass by. Actually, the same applies to all the buzz-words you face when working in the tech-industry. Technology phrases and expressions can often be tough to interpret and perhaps you have to Google a lot around to find the exact meaning. Another challenge is that these buzz-words have a tendency to come in continuous waves. In todays blog post I will talk a bit about the ongoing digitalization revolution, “digital twins” and “cyber-physical systems”. Did you already forget how to spell those words…? Please keep on reading! You will soon realize that the digital twin technology will make (or is already) making a huge impact in your life. It applies to railways, airlines, factories, ships, offshore structures, wind turbines, cars, homes…..virtually any physical asset!
As you may have noticed, digital twins and cyber-physical systems are both playing important roles in the DNV GL strategy and our way ahead as a data smart company. I recommend to take a look at digital twin video posted on our web-site today to get a brief understanding about the concept.
Actually, the “digital twin” term has been around for several years already. There are different versions about who introduced the term first, such as NASA and a digital replica of a space shuttle with all the data they had available back in the 1970s. In DNV GL we started to define very detailed product models of ships in the early 1990s as part of our Nauticus Production System. The idea at that time was to have a structured way of representing an asset like a ship and “map” incidents, properties and actions to it. Isn´t that somewhat like a digital twin although we have a different opinion about it today?
In DNV GL we use the following definitions of digitalization, digital twins and cyber-physical systems:
Digitalization is about ensure that the components of wider integrated energy and transportation systems work together as a whole. Digitalization is also about cyber security and using data in smarter ways to gain insight for better decisions.
The digital twin is a virtual image of your asset, maintained throughout the lifecycle and easily accessible at any time. One platform brings all the experts together, bringing powerful analysis, insight and diagnostics.
A cyber-physical system comprises physical components that can be monitored, controlled, and optimized by smart sensors, software and actuators.
What does all this mean in the context of ships and offshore structures?
Let us first take a look at some of the opportunities that lies ahead for digital twins and cyber-physical systems. As you might know, there was a recent potential major disaster (see this BBC-article) in the North Sea after a large, unmanned barge went adrift in stormy high seas and came close to colliding with offshore oil platforms in December 2015. Even though it turned out that the barge missed the oil platforms with only 1 kilometer, the oil company was forced to shut down the production and evacuate the platform. The most current information and calculations may not be readily available for such crucial decisions. Could we improve our customers decision process in case we had a virtual image of the asset maintained throughout the asset lifecycle and easily accessible at any time? Could it be combined with sensor data measuring both the structural integrity of the asset as well as environmental data in order to provide a more predictive way of deciding upon actions?
Opportunities with digital twin technologies
I think some of the opportunities with digital twins nowadays compared to some decades ago are as follows:
- The momentum on Internet of Things (IoT) is turning advanced sensor equipment into a commodity rather than a proprietary and rather expensive solution
- Everything is connected….or will be connected. According to a recent Gartner-report, we will see 6.4 billion connected “things” and a sharp increase of 30% compared to 2015. Most ships, systems, and components will be linked to the Internet, making them accessible from almost any location.
- We are well into the 4th industrial revolution and the connectivity makes it possible to collect and analyse data and provide the derived insights immediately
- Various companies like Microsoft, Google, Amazon and so on have an impressive momentum towards hyper-scale computing, so you will soon have “unlimited” storage and compute power for a limited cost. All the plumbing work with regards to setting up local clusters, networks and storage accounts is history. You have IaaS or Infrastructure as a Service instead.
- Artificial intelligence, cognitive services, big data analysis, machine learning, predictive maintenance and so on also becomes (or is) readily available on multiple platforms
- Big data, automation of knowledge work, advanced analytics, internet of things and connectivity creates a vast amount of opportunities and potential new business and service models
Short “story” or scenario which might help you to understand a digital twin concept from a structural engineering point of view
- A oil company decides to develop an oil field and ask some contractors to design and manufacture a platform, e.g. a jacket platform or something similar
- During the design phase various engineering teams will have different “views” upon the physical asset
- A concept model briefly laying out the design of the asset
- Some “super elements” on the conceptual model in case you have distributed design teams or just want to “organize” a huge model
- CAD (Computer Assisted Design) models containing all the details
- CAE (Computer Assisted Engineering) models suitable for simulations and so on
- Finite Element Analysis models where you subdivide a large problem into smaller, simpler, parts, called finite elements (it is an “idealized” version of the CAD-model)
- Capacity model needed for the fatigue calculations and code-checks
- ….and more
- During the operational life you will need a sort of history on your model (replacements, repairs, strengthening etc.)
- An asset integrity model is needed to improve reliability and safety whilst reducing unplanned maintenance and repair costs
- From a risk management point of view you might want a QRA (Quantitative Risk Assessment) model
- In the days of IoT, you probably need yet another model where you are mapping all the live sensor data to virtual objects in your digital twin
The big challenge is that all the items above are related to each other. Even though we have a vast amount of models, we are still thinking about the same physical asset. The only difference is that we have different views on the model depending on the tasks we are about to perform. This means that something like “beam #1” in your real-life model, will map to “beam #1” in the concept model and perhaps “beam #1_2” in your finite element model. And to make it even more complicated, “beam #1” might be split into “beam #1a” and “beam #1b” in your capacity model for fatigue calculations.
Whenever you inspect your asset (manually or by instrumentation), you will in some cases raise an action if you find something critical. Let´s say you have to strengthen or replace a beam due to corrosion or damages. You will have to “notify” the concept model that a change is needed, and the finite element model needs to be re-meshed and re-analyzed.
Or let´s say you have instrumented your asset heavily with strain gauges, pressure devices, temperature devices and so on. Rather than doing a manual inspection, you can have machines doing that for you. The machines will load tons of data into your digital twin. Once a certain threshold limit is reached for a single sensor, an “alarm” will occur and you will be able to take action in due time. This sounds fairly simple, right? But sometimes there will not be a single sensor triggering such an alarm. It might be a combination of various signals and we can use technologies like big data analytics and machine learning to predict structural failures before they actually occur. Sensor data, monitoring and analytics made possible through the digital twin can enable profitable, safe and sustainable operations based on broad and deep domain knowledge.
These days we are already working heavily in DNV GL – Software towards new services and solutions for digital twins. We call it “eco-systems” and we plan to deliver significant added value and new service offerings for our customers. Stay tuned!