Transforming decision making in supply chain logistics with digital twins

The supply chain logistics industry is undergoing a fundamental paradigm shift, driven by digital technology. Global supply chains rely on physical assets, like containers, warehouses and trucks—assets which have historically been difficult to bring online. But only by connecting those assets can we fully digitise supply chains.

But once those physical assets are brought online—then what? How can we use that information to make supply chains more resilient and flexible?

One of the key ways we’re addressing that question is through digital twin technology.

What are digital twins? The future of problem solving and decision making

As it says on the tin—digital twins are a digital representation (or “twin”) of a physical object or system. For example, a simple digital twin could be a digital representation of an airplane wing. That digital twin can be used to run simulations on how the airplane wing handles, its aerodynamism, when it might need repairs, etc.

More complex digital twins are digital representations of full systems—like a warehouse or a factory floor. Those digital twins can be used to analyse production processes and proactively address maintenance and repair.

Data between a digital twin and its physical counterpart can be transmitted in both directions. For example, a digital twin can receive data from the factory floor, a decision can be made based on the data analysis, then data can be transmitted back to the factory floor to trigger action.

Until now, digital twins have been primarily used to model and simulate specific and isolated processes

However, as of right now, digital shadows are more common than digital twins. The main difference between digital shadows and digital twins it that the data from digital shadows only flows one way: from the object to the shadow. The shadow is unable to send data back to the physical asset to implement changes.

Learn more about digital twins and how they’re redefining the future of supply chains in this video with Gavin Laybourne, CIO of APMT & Head of IoT & Automation Platform, and the inventor of data processing, Dr. & Prof. Wil van der Aalst.

Redefining supply chain logistics with digital twins

By analysing digital twin simulation data, we can identify bottlenecks, inefficiencies and potential risks in the supply chain. This leads to better planning and resource allocation, ultimately resulting in increased efficiency and reduced costs.

Beyond that, digital twins allow for:

• Real-time tracking and monitoring of assets: By placing sensors on or within a physical asset, we can feed data to the digital twin on the asset's location, condition, performance, temperature, damage, etc.
• Predictive maintenance and repair: Instead of reacting to machinery breaking down, we can use digital twins to proactively deal with maintenance. This allows us to repair assets on a more convenient and optimised timeline, rather than when it’s holding up a process, like unloading a vessel.
• Enhanced decision making and optimisation: Digital twins help us move from being reactive to proactive. When we have a digital overview of our physical systems, we can make more informed decisions about resource allocation, planning and risk management.
• Increased efficiency and reduced costs: Digital twins also help us optimise supply chains moving forward. For example, we can identify potential cost savings or efficiency gains through alternative routes or types of transportation.

How we’re using digital twins for visibility and predictability

Our ambition is to create a digital twin of Maersk’s integrated ecosystem. While we’re years away from that, we’re progressing by focusing on specific verticals of our business.

By creating vertical digital twins of vessels, warehouses, terminals, etc., we can eventually create the building blocks of a horizontal digital twin of our entire integrated ecosystem.

This shift towards creating a horizontal digital twin will help us move from digital twins used to visualize our supply chain to twins that predict and improve our customers’ supply chains.

One area we’re pioneering this technology in is our terminals. We’re currently rolling out a digital twin simulation product that predicts what will happen operationally in the near term.

We use container volume, vessel schedules, sensor data and terminal data to build this simulation, which allows us to see which vessels will arrive when and with what cargo.

We can then predict what resources we will need where—like cranes. That ultimately allows us to predict and adjust operations accordingly.

This digital twin simulation is focused on predictability. We need to precisely forecast how long certain actions will take. If we don’t know how long unloading a vessel will take, we have to build in buffers, which means equipment might not be in the right place at the right time—it’s slow and inefficient instead of tight and fast.

Some of the other challenges that we’re trying to overcome using digital twins include:

• Scalability: Many Maersk departments have their own digital twins, but they’re siloed with no overarching or interoperable architecture. We’re working towards a network of digital twins that connects different steps in the supply chain.
• Chatty IoT: IoT initiatives have produced a glut of big data, but the data often goes unused. We’re exploring how we can use that data to optimise processes through digital twins.
• Predictive ETAs: Digital twins can help us move to prescriptive actions and give real-time ETAs for delivery that are driven by artificial intelligence and machine learning, instead of giving general delivery estimates.

Five predictions on the future of digital twins in logistics

There are so many possibilities for how digital twins can be used in supply chain logistics. But narrowing it down, here are five key predictions for how digital twins will develop in the near future and especially contribute to the next generation of intelligent ports.

1. Further integration with artificial intelligence and machine learning

By leveraging ML/AI, digital twins will be even more powerful tools for optimisation and efficiency. Digital twins hold rich data sets that can then be analysed by machine learning algorithms to identify patterns and trends, leading to more accurate predictions and recommendations.

2. From digital shadows to digital twins

Digital shadows are digital representations of physical objects, but unlike digital twins, the data only flows one way: from the object to the shadow. The shadow is unable to send data back to the physical asset to implement changes. Digital shadows are much more common than true digital twins, but we will soon start to see many more ‘true’ digital twins within supply chain logistics.

3. Wider adoption up and interoperability

Right now, the use of digital twins (or digital shadows) is limited to more controlled ecosystems, like vessels and warehouses. However, those only make up part of a full supply chain. In the future, we will start to see digital twins up and down the supply chain, used by many different supply chain actors. Interoperability will be key to these actors integrating in a connected way. Organisations like Digital Container Shipping Association (DCSA) are already working to establish standards for a common technology foundation that enables global collaboration on connected supply chain networks.

4. Digital twins pioneering greener logistic

One of the most consequential ways we will see digital twins being used in the future is to pioneer greener logistics. We’re already using a digital twin simulation on some vessels to optimise power consumption from reefer containers. The possibilities for using digital twins for greener solutions are endless.

5. Continued need for human intelligence

Using digital twins, we can predict what will happen in the future within a system or asset. However, those predications are based on a range of normal events. But those predictions aren’t useful when there’s deviation from normal—such as the pandemic or the war in Ukraine. However, when we build digital twins on technology like process mining, which gives us an overview of historical processes, we can immediately detect anomalies. That allows us, as people, to decide how we want to override existing processes and address the unexpected. In the future, we will see more human intelligence and decision making combined with digital twins and predictive AI for the best results.

True gains only come with industry-wide action

Digital twin technology, especially when combined with AI/ML and predictive analytics, can drastically change supply chain logistics by improving efficiency and cost savings by optimising decision making and resource allocation.

To reap the full benefits of this technology, we, as an industry, must collectively overcome the challenge of interoperability and responsible data sharing in order to create fully connected, resilient and green supply chains.

Author:

Gavin Laybourne, Vice President and CIO, APM Terminals, brings over 30 years of experience leading digital transformation in various industries. He joined Maersk in 2019 and is instrumental in leading the tech led integration of container logistics with the digitisation of physical assets.