Introduction
Digital twins are essential for optimizing performance, predicting outcomes, and improving decision-making across various fields. This study showcases initial efforts to digitalize the Tiger Motors Assembly Line, a manufacturing training facility at Auburn University. The facility offers hands-on training in Lean Manufacturing by assembling LEGO® SUV and Speedster cars, simulating real-world processes.
A multi-phase digitalization plan seeks to transform the Tiger Motors Assembly Line into a demonstration hub for digital manufacturing and Industry 4.0 technologies, focusing on small and medium manufacturers (SMMs). The current phase involves creating an agent-based digital twin model.
The Tiger Motor’s assembly line is divided into three work cells, each with five workstations. The cars are assembled sequentially, and each car needs to go through all fifteen workstations. The process of how cars travel through the assembly line is illustrated below.
Simulation model
For the development of the agent-based digital twin model, researchers used AnyLogic simulation modeling software.
All agents are created in the Main agent of the AnyLogic model, which is responsible for initiating the processes when the model is run. The Main agent includes the Communication agent, a population of 15 Workstation agents, and an initially empty population of Car agents.
The Main agent also contains an event that runs at the model start and is responsible for initiating the Communication and Workstation agents.
The Communication agent is responsible for receiving messages, breaking down the contents, and then passing the equivalent messages to the Car agent. The Communication agent is also responsible for querying the MySQL database, which contains information about the automation system.
The Car agent is responsible for tracking the location and state of each car assembled in the facility. To achieve real-time tracking of the cars within the facilities, the researchers implemented an agent-based statechart that describes the current state of each car.
The Workstation agent tracks the state of the assembly line’s fifteen workstations. Whenever a production run is finished, important statistics regarding key performance indicators are gathered using collected data. These key performance indicators include data about workstation utilization, service times, and downtime due to faults.
Read also: A White Paper on Digital Twin Development.
Results
The agent-based digital twin model can achieve bi-directional communication with the facility’s automation systems, and it utilizes Industry 4.0 technologies to create a digital replica of the production line.
By leveraging real-time data from IoT devices and automation systems, it tracks workstation statuses and product locations. The ABM approach allows integration with different simulation models of the facility, enhancing the digital twin’s capabilities, enabling real-time data-driven experiments, and directly influencing physical system behaviors based on experimental results.
The next steps in the project involve verifying and validating the model, developing visualization tools to help operators comprehend the system's production state, and utilizing the system during production runs to evaluate its impact on improving decision-making.
Future plans include integrating advanced elements like a SCARA robot and an automatic storage and retrieval system into the digital twin model. The agent-based digital twin model will also be combined with the existing multimethod simulation model.