The terms simulation and digital twin are often used interchangeably but there is a difference. This difference can be quite confusing as both terms are used to mimic a real-world system within a virtual environment.
Simulation
A simulation is “The process of designing a model of a real-world system and conducting experiments with this model for the purpose either of understanding the behavior of the system or of evaluating various strategies within the limits imposed by a criterion or set of criteria for the operation of the system” (Shannon, 1975).
A simulation model simulates the basic behavior, but it does not necessarily incorporate all the details of the real-world system. Instead, it is a standalone replication of the real-world, completely independent from that real-world system. All the necessary controls to drive the system are modelled internally within the simulation model. A flight simulator is the most well-known example. The pilot in training feels that he is flying an airplane, but the simulation is fully independent and not connected to any system of a real plane or flight control.
Digital Twin
A digital twin is a rapidly growing technique in product design. However, in process design, and especially in logistics engineering, it is currently a bit of a buzz word that lacks a clear definition. In order to provide some guidelines, there are at least two key elements required to set up a digital twin:
A digital twin can be seen as an emulation of a real system. It is a digital representation of a real-world system that is so close to reality that strategic, tactical or operational decision making can be based on it. Because the digital twin is not a stand-alone simulation, but has an interface to an external system, it most often runs in real-time. This is required in order to prevent time synchronization issues between systems or to prevent that all systems are overwhelmed with data, messages, etc.
In some applications, the digital twin is a real time digital representation of the reality which runs synchronous to the real system. These real time applications typically cannot stand on their own and are often referred to as ‘Digital Shadows’.
Software
Depending on the main purpose of the digital twin a different type of software is required:
Discrete event simulation (DES) jumps from point to point in time where the state of the operational system is changing. A strong advantage of using DES for a digital twin is that most of the DES tools are capable of modelling automated and non-automated systems which allows for more use cases. DES systems are especially powerful when used in combination with higher level control such as ERP, WMS or WCS. The system is modelled as a series of successive (future) events and the simulation jumps from one event to the other ignoring the time in between the two events. Each event will trigger changes in the data which is turn will create or alter (future) events scheduled within the ‘event controller’.
This gives the possibility to progress through time really fast which is helpful when trying to rapidly test multiple scenarios. The draw back for using DES on a digital twin is that both, the twin and the interface system, typically have to ‘listen’ to the each other. This can cause time synchronization problems as it will be impossible to know when (what point in time) a connected service sends a message to the digital twin and what point in time the Event Controller of the simulation will execute the next event (possibly triggered by the message). Then, the message is most likely sent at any unknown time between event i and event i+1. This creates a gap in time between the message actual being sent and actually being received by the digital twin. This problem can partly be solved by using a real time factor. The event controller will jump through time with ‘dummy events’ to fill the time between two actual events. The chosen size of the time interval between two dummy events is the maximum error that could occur between the actual time the message is send and the time it is received by the digital twin.
To help combat this time synchronization issue, Emulation software is more often used for making digital twins with automated systems and MHE equipment. It can model MHE equipment with a high level of detail but a limited amount of effort. The 3D environment is not only a visual representation but it contains the real mechanical components. That allows the digital twin to simulate in much greater detail and include complex calculations such as collision detection. These applications typically simulate in continuous real time only, which prevents the issue of time synchronization. Emulation software is more often used in combination with low level controls such as PLC, where the I/O is explicitly modelled within the emulation model and controlled by the PLC.
A newer technique within digital twinning is to use software that initiates from the game industry. Physics engines or game engines such as Unity allows the digital twin to go into even further levels of detail in modelling behavior and include variables that have an effect on the system out of the control of the system such as gravity or natural forces. It allows for instance to define how fast a curved conveyor can move until totes start to tip over or break down.
Whatever solution is being used to create a digital twin, it must be an open solution that can easily be connected to external IT system without much effort and use the external application API. Care must be taken to check if the software can be used to do both, send messages and listen to incoming messages. Most frequent interfaces include:
Use cases
Digital Twins allow the user to support functional designing, testing and validation of the external control system(s) or to evaluate how the IT landscape interacts and is capable of handling all information that is sent back and forth while monitoring operational systems in real time. Examples of use cases include:
A Digital Twin in logistics is still a relative new technique, but as you have seen, it brings a great deal of possible use cases. It helps companies to plan, design, predict and maintain their operational systems in greater level of detail and with high accuracy, have shorter downtime during commissioning and speed up design projects.
Shannon, Robert E., “System Simulation: The Art and Science”, Prentice Hall, 1975
—Auke Nieuwenhuis, St. Onge Company
