Embedded algorithms design faster and more accurate industrial and health-care equipment
Today, much of the equipment used to perform routine tasks or to manufacture common products, are complex systems. These consist of complex computer-controlled mechanisms called cyber-physical systems. Delivering the improvements that customers expect requires technological advancement of the whole chain of systems. This requires sophisticated methods to identify detailed behaviour down to the level of physics. This continuous quest nowadays gets assistance from tools that autonomously assess and analyse industrial machines, looking for potential improvements.
Building blocks control model-based systems engineering
The EU-funded I-MECH project developed a framework to employ advanced control solutions in industrial settings. The chosen approach is known as model-based systems engineering. The project developed 11 building blocks that (among other functions) monitor or control industrial processes to find incremental improvements. One example concerns manufacturing errors in electronic hardware. “Engineers can introduce advanced algorithms that learn what type of disturbance is causing erroneous performance,” explains project coordinator Arend-Jan Beltman. “The algorithm identifies the repetitive nature of the fault and then starts compensating for it via its control logic.”
Three layers support interoperability
To simplify the complexity of cyber-physical systems, the I-MECH team identified three layers in industrial manufacture where such algorithms could be applicable. Some of I-MECH’s building blocks focus on a single layer, whereas others contribute in all three. The lowest, called the Instrumentation Layer, or Layer 1, interacts with the system physically. Building blocks at this level act as sensors or actuators. The team developed several rapid sensing devices, some of which are wireless. Following, the next level is an industrial communications bus, which consolidates all Layer 1 inputs. At this layer, the algorithms control accurate machine motion. The third, or systems, layer functions as a container for algorithms that interact with system functions and operators for factory management systems. Here, the built-in intelligence automatically calibrates the system and predicts maintenance needs. The layered approach allows for clear interfaces between engineers having different backgrounds. Engineers at Layer 3 work with a model of Layer 2, while Layer 2 engineers use a model of Layer 1. This supports interoperability.
Successful pilot applications
The team applied its building blocks to five pilot applications, which use machinery developed by project partners. The applications include a generic substrate carrier, which is the conveyor component of large-format inkjet printers, and a 12” wafer stage of semiconductor manufacture. The remainder cover a teabag machine, a computer numerical control (CNC milling machine) and a healthcare robot that moves an X-rays system around patients that lie on a table. In each case, the systems received upgrades identified by the building blocks. Eventually, all building blocks, and an entire toolchain, will be available for industrial customers, who will be able to select just the building blocks they need for their specific application. “During September 2020,” notes Beltman, “I-MECH lead partner Sioux Technologies submitted the final proposal for a successor project. That will add artificial intelligence to the existing set of building blocks, for manufacturing systems where I-MECH left off.” Specifically, the new project will develop a fourth layer that facilitates orchestration of multiple systems in the same factory. Smart industrial systems that design and refine newer systems will lead to efficiencies and rapid evolution in manufacturing.
Keywords
I-MECH, building blocks, industrial, manufacturing, manufacturing systems, cyber-physical systems, model-based systems engineering
