We are all presented with new technologies or the repurposing of a technology on a daily basis. As engineers, manufacturers, or decision makers, we must decide if this technology will enhance what we do and if it is worth the investment. This is the first in series of articles that look closer at the individual pieces of a digital twin world and look at it from a perspective of history, how its fits in the big picture, the details of the technology, the challenges it presents, and a practical example of the technology as a reference.
What is Virtual Commissioning?
Virtual Commissioning, also known as PLC Emulation, is the creation of a virtual environment that allows the user to test controls programs for automated facilities before the facility is built or renovated. Traditional applications of the technology include creating a virtual manufacturing simulation to connect to a real or virtual PLC so the ladder logic of the PLC can be fine tuned and validated before the launch of a program.
The diagram below shows a typical set up for PLC emulation (virtual commissioning). The hardware for the factory is represented by the simulation environment, the PLC (either real or virtual) runs the PLC code to be verified, and a server program acts as the conduit between the two environments across a common network of some sort. In this case, the simulation is the OPC client, the OPC server acts as the conduit, and the PLC runs the code that is being verified. Ethernet is the network providing the connection between the various programs. There are many variations to this theme. Sometimes the server and client operate as one program. Sometimes there are multiple PLCs or they are run "virtually" as well to allow the entire setup to be run without having physical hardware or a physical PLC. Lastly, OPC is not the only communication protocol. Each PLC has a more native language that can be used to communicate faster. OPC is simply a generic way to discuss these connections.
My first experience with Virtual Commissioning was in 1993 working with Deneb Robotics. We were able to hook up a series of hardware switches to a 3D computer simulation of a series of conveyors and control the route that a part took through the simulation by toggling switches. This setup required a few expensive UNIX workstations, a small PLC program, and a simple simulation. It was a cool demonstration of technology but not terribly useful for reducing the amount of time it took to launch a complex controls project. In those days we were exchanging entire “words” of information and then parsing out the bits we needed from memory to make decisions. We used OPC to provide the connection between the various technologies to make this happen
As time progressed, specialized simulation programs emerged that were designed to make these connections and simulations easier to build. Today there are several companies that tout solutions in this space that are standalone ways to model systems quickly. As with many of the technologies will discuss, the tradeoff comes down to the specialized interface of a standalone product that is specifically targeted at a problem versus the collaborative synergy of a larger suite of tools that save time by connecting users and data in a consistent manner.
Why do Virtual Commissioning?
So how does virtual commissioning fit into the Digital Twin story? As we create new ways to manufacture and distribute goods, we need to be faster in renovating our facilities. The automotive market has been at the leading edge of this concept for a few decades. They are often presented with a short shutdown in the summer and winter of about two weeks each where they can make model year and major renovations to support the next model year of their vehicles. They do not want to wait for conveyor or other equipment to be set up on their floor to test any custom PLC ladder logic required to run the system. They also do not want to pay for the expense of their vendors “running off” the equipment in a standalone fashion (different than their environment).
This lengthens the project timeline and only finds the most basic issues in the ladder logic. Instead, they opt for a virtual version of the next generation of the manufacturing operations. The controls engineers are required to bring their programs to a lab and test them well ahead of the planned shutdown. This allows for more aggressive scheduling during the shutdown and more possible changes in the same time period. The cost savings from this approach can be substantial but the true benefit is accelerating an implementation and reducing the time to market.
What are the challenges with Virtual Commissioning? As with any endeavor to change, you are often met with resistance from the status quo. Controls engineers do not want someone overseeing their methods and dictating their timing for a project. It is also true that Virtual Commissioning is excellent at enforcing programming standards that are often unappreciated by those who are responsible for writing the code.
Another issue is the availability of data. In a standalone mode, the simulation modeler can have a difficult time staying current with the mechanical team’s drawings of the physical layout and the controls view of memory allocation, I/O design, and ladder logic development. This is where a collaborative environments shine. They house all the relevant data in a common database and the modeler can be sure that the data is current. A single source of truth, so to speak.
The next challenge is timing. When to build and analyze. Systemwide simulations are often done in the very early stages of a project to determine the amount of equipment and best configurations for a given space in a manufacturing facility. This simulation would generally be excellent for the future testing of PLC code but often the controller, IO, or even the control philosophy is not known at these early stages. All too often, these initial factory models are not used in emulation because the logic that will eventually reside in the PLC is coded into the factory model in a different proprietary language. This can also be addressed by a common database for all types of analysis and modeling. The effort spent in the early stages can be reused when it comes time to develop a virtual version for commissioning.
The last word of warning is speed related. There have been many advances in technology over the years that allow for direct, native communication with PLCs but high speed and high IO traffic can cause delays in the turnaround times for the information flow to and from the simulation and the PLC. Testing of a system is warranted whenever the system is large, or the response time is critical.
The Practical View
To tie together the conversation, an example of a past project is a good way to wrap up these discussions. We discussed the traditional automotive example for commissioning a new system on a tight timeline. There is also the case where changes are being made continuously to a system that runs almost 24x7. Approximately 15 years ago, Forward Vision worked with a large computer manufacturer on a system where they received orders for computers over the internet and phone. These orders fed all the way through their systems to the assembly workstations on the floor. Their challenge was the ground floor of this digital stack. They wanted to test all of the ordering and accounting systems that all tied together down to the actual creation of the computer on the floor but the facility ran 164 out of 168 hours each week. They only had a four-hour window for loading new programs and testing them.
Very little time existed for debugging. We built them a virtual environment (simulation) and set up a lab of 8 Physical PLCs. They were able to use that testbed to improve their system and add new features without impacting production or the experience of their users. It is a great example of using virtual commissioning as a continuous improvement tool and it is a creative solution to a problem that was slowing innovation and hurting the quality of the end product.
Virtual commissioning has been around a long time. Getting quality data in a timely fashion is a key to accurate modeling and the data often changes quickly. This supports the case for incorporating it into a large Digital Twin effort but be aware that people will want to stay with their current methods and the speed of the final solution should be tested before a large project is conducted.