What is the virtual commissioning of the robot system? Let's think about a manufacturing scenario where a new assembly line is decided to be added to the factory shopfloor or modify previous ones. In this case, the idea behind the virtual commissioning is to test and validate new or modified system performance before the physical system is built or installed. Another approach is to build a system on the shopfloor(online) and test the system afterwards. If we build or modify online, there is a catch here: we need to shut down the current working station to implement and test the new design. Such a case will increase downtime of the shopfloor, increase risks and endanger the safety of operators working inside or outside the cell. It can increase the delivery time to market, leading to increased costs for the company.
Now, we understand what the common disadvantages of the online approach are. At the same time, the design team seeks to optimize this new implementation, so there would be a lot of "what if" scenarios. Here, a digital twin or simply digital replica of the system will be modelled in a virtual environment to avoid iterations and exploration on online platforms. There are factory simulation softwares from the industry to fulfil these needs, companies such as Dassault Systems with Delmia, Siemens with Tecnomatix, ABB with RobotStudio, Visual Components and so on.
I described the virtual commissioning of the robot system shortly in the above section. I have been teaching this subject for the last five years to my students at university. As a result, hereby, I simplify the virtual commissioning of the robot system into five steps. Let's look at them:
Step one. The virtual model of the robotic system
This is the main and the most important part of virtual commissioning. Every kind of step from here forward you need to do is depend on realistic virtual model. Any deviation in dimensions, positions, or miss-model would hugely impact further steps and lead to possible failure. This step requires a detailed 3D model of all components that will exist in the workstation, from specific components for manufacturing, then factory facilities such as conveyors, and machines, robots that you will utilize in the manufacturing process, and finally, other resources related to robots such as grippers and special finger design, welding torch, painting, etc. There could be two types of inputs for this step, first, the design team already came to the final proposal to some extent and provided you with most of these requirements or second, you are responsible for creating such a virtual model from scratch but keep in mind manufacturer can easily provide some of the models.
Step two. Robot programming
This step is crucial for ensuring that the robot system functions correctly and, more importantly, operates safely in the real world. In this step, you, as an engineer, create robot programs that control robot's movements, actions and interaction with its environment. For creating this program, the manufacturing process sequence needed to be defined and available for you to continue with this step. Later, you should reference this and actively utilize the best geometry features of components for robot movements. In this step also, you should simulate all signals regarding communication with the robot itself. Signals can include robot's gripper functionality, such as activating pneumatic pressure, magnetic gripper, etc. Another kind of signal can relate to communication between the robot and other machines as they operate in parallel or in a sequence with each other. Last but not least, all signal action is required for controllers such as PLC.
Step three: Debugging and optimization
Once the robot programs are created, it should be run in the virtual environment first and tested and monitored to function correctly. This step can include checking collisions in the system and avoiding facing robot's reachability and joint limitations. Next, you need to evaluate if your system is optimized or if you can improve it further, such as by reducing cycle times and improving the robot's program. Robot's program can be checked in a way that robot's speed and acceleration can be optimized or monitored to see if other robot's configurations could help improve the robot trajectory.
Step four: Post-processing
I could put this a bit more simply, and this step requires making all system programs, such as robots and controllers, ready to be uploaded on physical systems. Post-processing is the final step for mapping programs with specific robot vendors or controllers from the virtual world to the real ones. This depends on the simulation software you are using. Is the software capable of fulfilling all functionality of devices to a replica of the real one, or are they implemented in a more generic version, and do you need to translate some of the functionalities to real components?
Step five: Implementation of a physical system
The first and most important part is that the physical robot system should be assembled and installed according to the developed virtual commissioning design procedure. All hardware components, including robot's end effectors and sensors, should be installed correctly and securely. Next, the software programs finalized in step four should be installed on the physical robot system and tested. The physical robot system should be tested to ensure that it can communicate with other systems and equipment on the shop floor. This may involve testing communication protocols such as Ethernet, Modbus, or Profinet and verifying that the robot can interface with other automation systems such as conveyors or vision systems. And finally, the physical system should be tested to ensure that it performs the program as expected and performs intended tasks within specified cycle time and accuracy.
Virtual commissioning of robot systems through optimization simulations and utilizing the digital twin of the physical system can play a crucial role in integrating robots into manufacturing and production processes. In this blog post, I tried to familiarize you with five steps that are required to accomplish such seamless integration of the robot system. Indeed, through each step, there are more detailed tasks needed to be performed, such as robot's tool calibration, which requires inputs from the real robot and integration into the virtual replica. In the link below I attach the academic level of virtual commissioning of robot system videos in our own university. Let me know in the comments what your thoughts are on virtual commissioning for integrating robots into new factory lines. Have you ever used virtual commissioning before? If yes, what benefits or challenges did you experience from it?
Tampere University - Virtual Commissioning of the robot system projects
