A technology widespread in the world of industrial manufacturing is slowly being recognised as a useful tool in supply chain and port planning.
As a result, Gartner named “digital twin” as one of their Top 10 Strategic Technology Trends for 2018. After reading this article, you will have a better idea of how the digital twin solution differs from simulation software, how digital twinning is used in maritime supply chain planning and in ports, and what does it take to implement the technology.
In manufacturing, a digital twin is an exact virtual replica (mathematical model) of a real machine, including its physical parameters, design characteristics, as well as interactions with the physical world to which the product will ultimately be exposed.
Those interactions may include air flow, water flow, electric current, chemical reaction, etc. Digital twin has a real-time connection to sensors that continuously report the state and conditions they are monitoring. Algorithm(s) allows computers to evaluate what is happening in the real environment and continuously run “what if” scenarios of what could (prediction) or should (decision) happen next in the operations of the digitally replicated machine.
This is the key difference between simulation and digital twinning. Simulations only simulate what will happen to the machine based on what the human designer asks it to imitate. Simulations are limited by the imagination of the human deciding what events known to man can occur during operations of the machine, while digital twin is bound only by the limitations of the algorithm that continuously learns and evaluates scenarios beyond human comprehension.
Understanding the value of digital twinning specific to logistics and transportation processes requires one to let go of the assumption that supply chain and transportation models are 3-dimensional, like physical objects
Movement of goods and people happens over time, so the traditional 3D model needs to include the time dimension. In its entirety, a digital twin spans the whole spectrum of a business process, from the highest level of the mega supply processes and networked assets down to the lowest level of operational work instructions.
The behavioural evaluation of a complete mathematical model of such a complex organism would not be possible without artificial intelligence, involving fields as diverse as machine learning (including deep learning) and image and language processing.
Digital twin involves machine observing humans and operations, mathematically modelling motions and decisions, then computationally improving the twin in search of better processes and decisions. Every improved plan can be transferred back to the real machines and real humans executing the operations. In other words, there is no gap between the simulation and the machines. Thanks to all of that, modern digital twins achieve a level of predictability and self-improvement that is unique and more accurate than every previous model.
Digital twinning is critical to the development of autonomous behaviour of machines and their awareness of their work environment – a loading/unloading robot understanding what it means when a truck is ready for operation, a container crane robot understanding when the autonomous ship is ready to commence movement of containers, etc. Advances in the IoT space are critical to creating successful digital twins. Sensors/IoTs embedded in physical objects and attached to humans can collect massive amounts of data allowing the digital version to mimic and iteratively improve all interactions between humans and physical objects. This process will eventually lead to the faster replacement of human jobs with robotics, a subject of a separate and fairly fierce debate.
In the world of maritime supply chains, ports can be true beneficiaries of digital twinning, as we try to deploy autonomous pilots on autonomous ships operating with fewer human interventions.
The Port of Antwerp stated that they created a digital twin of their port environment to monitor a variety of operational events within the port. The computer was left to “imagine” any possible scenarios of behaviour, thus helping humans make better decisions affecting port operations.
On land, a large rail operator twinned their rail network and the ports they are serving. Based on real-time information received from the ports, the digital twin is used to re-plan movements of rolling stock and access to the rail network to remove any element of surprise in moving freight.
It is expected that widespread digital twinning in supply chain logistics space would lead to a 10% reduction in supply chain cost, and it wouldn’t be a one-time gain. While the costs of creating complex models are coming down, the ports and the maritime supply chain operators are gaining an important digital ally in their quest for a competitive position in the chain.