A team of researchers in Colombia have successfully developed a solution for better interaction between operator, computer and machine, enabling increased flexibility in manufacturing systems.

A century ago, Ford painted all their cars in one color: Black. Black paint lasted the longest and cost the least, so that was what consumers got. As a consumer today, it can be hard to imagine such a world without a wide range of choices for every product.

The demands nowadays drive companies to offer a variety of product choices and the opportunities to personalize or customize products, moving production from mass-production to mass-customization. This is a major challenge for the product manufacturing industry that struggle to find ways to make small-scale, adaptive production economically viable.

Research team took on the challenge

In Bogotá, Colombia, a team of researchers at the Mechanical Engineering Department of the University of Los Andes study these challenges facing the industry today:

“As the industry has to produce flexible products, plants need to be very flexibly and allow for reconfiguration”, says Assistant Professor Giacomo Barbieri leading the research team.

To support the industry in this transition, the researchers wanted to figure out a solution for a production system to self-configure and self-adapt to product and environmental changes. Normally, the operator of a machine producing customized products would have to both test new control logic and manually deploy the code in the PLC for every customization.

“Our idea was to cut out all these manual steps that are not acceptable in a flexible production environment and instead let the machine automatically self-adapt the control logic to the production order and test it”, Giacomo Barbieri says.

Seamless interaction between human, computer and machine

To achieve this, the research team developed a solution with real time interaction between the human operator, the production system and a computer testing the control logic through virtual commissioning.

The process starts as the operator inputs the production order, after which the system automatically self-adapts its control logic to fulfil the production order and virtually tests the new logic. The operator then checks if the virtual test fulfills the product specifications and either deploys the new code and triggers the production, or manually adapts the code if the virtual production did not fulfil the product specifications.

“We wanted to make human-machine collaboration more seamless, so the operator only needs to supervise the production and decide whether or not to deploy the code,” Giacomo Barbieri says.

But that doesn’t mean that a human operator will be obsolete in the future – the case is quite the opposite, if you ask Giacomo Barbieri:

“Operators are faced with even higher demands, as production gets more automated, and they are still hugely important to make decisions and anticipate process malfunctions,” he says

Experior essential in case study

To validate the solution, the research team applied the concept to a case study setup. A Universal Robot simulating a cutting operation constituted the production system, the software UR Sim simulated the robot movements and Experior emulated the robot and its interaction with the production environment to test the control logic through virtual commissioning. You can see illustrations of the setup on the above and a video below.

David Andrés Gutiérrez Silva was part of the team developing the solution. He was first introduced to Experior during his master’s degree at the university and now works as a C# developer for Xcelgo, so he knows the advantages of Experior in such a setup:

“Without Experior, we wouldn’t be able to simulate the manufacturing environment. Now the operator is able to virtually oversee the production, identify mistakes and correct them in the control logic routine,” says he says, and Giacomo Barbieri agrees:

“Experior was a big advantage in our setup, as it enabled us to reproduce the effect of the production,” he says.

Next step is industry implementation

The solution is still a prototype concept, and the team continue their research at the university:

“This is the first step in developing this new methodology. The conditions in our current setup are ideal, and the coding process is easy, so in future investigations, we want to include more rotations etc.,” says David Andrés Gutiérrez Silva.

The most important next step is however testing the setup at an actual manufacturing company:

“We know that the concept works and would be valuable at a company with many changing production orders, but we need to team up with such a company to figure out how to implement the solution in a real production line,” Giacomo Barbieri says.

Towards a true Digital Twin

In the long term, Giacomo Barbieri has ambitions to transform the setup into a true Digital Twin with bi-directional communication between the simulation and the physical world:

“With censors and bi-directional communication, we can adapt production uncertainties from the physical system to the virtual model and foresee mistakes,” he says.

The current setup can then function as a natural fore step to the implementation of Digital Twins:

“This concept implements a new workflow at the company that can act as a steppingstone to using Digital Twins. Once you realize the benefits of virtual commissioning, Digital Twins are a natural next step,” Giacomo Barbieri says.