DIH4AI: I-PRAG-5 Simulation and digital twinning for industrial systems
Advanced automated design of simulation models and their use of soft-sensors
Advanced automated design of simulation models and their use of soft-sensors
An increasing level of connectivity in Industry 4.0 cyber-physical production systems enables new opportunities for data logging, analysis, and processing. Server-based or cloud-based solutions are capable and suitable for soft-sensing unmeasured variables from the shopfloor. Some of the process variables are difficult, costly, or even impossible to measure directly, therefore, it is better to calculate them from other measured variables. For this calculation, mathematical-physical relationships have to be expressed or learned. Such a calculation is done with simulation models. In this technological experiment, the simulation software is not required to be installed onsite, close to industrial components on industrial PCs (IPCs), but it can be deployed on an industrial site server (on-premise) or into the cloud. This approach saves computational resources (equipment) in the area of operational technology (OT), decreases the number of needed licenses, as simulations for more than one variable can be calculated within one simulation environment in parallel, and makes maintenance processes more efficient. The experiment is being implemented with the partner Lego, which has a packing facility near Prague. The simulation is focused on energetic behavior of packing machines. Implementing and deploying this experiment in practice can bring significant improvements in production line efficiency, which may be mainly caused by reducing costs compared to introducing and purchasing new hardware solutions (e.g., sensors, PCs onsite, etc.). The effort needed is then reduced compared to traditional approaches as the effort/workload for its implementation is systematic, it eliminates ad-hoc engineering work, which is frequently not as efficient and error-prone as it is in the case of systematic work. Among the expected impacts is thus a reduction of time-consuming effort/work that is currently made in industrial production facilities for implementing traditional approaches. It reduces the costs needed for purchasing licenses for instances of simulation software by joining the simulations in an industrial facility under one umbrella.