It Proposes to Transform the Landscape of Autonomous Manufacturing Through the FAUSTO Project, Boosting Process Quality and Efficiency with a New Generation of Digital Solutions
ldeko will pave the way towards a more sustainable, efficient, and safer industry through the deployment of autonomous systems capable of real-time optimization of manufacturing processes in terms of sustainability and productivity, by introducing the most disruptive AI based on deep learning. A true challenge addressed through its participation in the Fausto project, funded by the CDTI Transmisiones program, alongside companies Danobat, Ekin, Savvy DataSystems, Sensofar, ITP Aero, and Fersa, as well as the universities of Zaragoza, UPV/EHU, and the University of Murcia.
According to the technology center, the project tackles both the application of deep learning algorithms and the development of the necessary components to generate large labeled datasets. The initiative will therefore develop mechatronic elements, computer vision, and novel optical devices installed on the machine to generate appropriate datasets to feed deep learning (DL) algorithms.
As a leader among the research organizations in the AEI consortium, ldeko assumes the technical coordination and monitoring of the project, while also participating in various research activities around the development of physical and virtual sensors for near-cutting-point measurement in grinding and broaching processes, in order to obtain meaningful data to feed AI models; physical models for the generation of high-quality synthetic datasets to train deep learning AI models; AI applications trained with hybrid data for grinding process optimization and improved surface finishes; and autonomous grinding systems through the generation of a process fingerprint.
Use Cases
The introduction of AI based on Deep Learning focuses on two industrial pilot cases identified as suitable for manufacturing improvement: automotive bearings (Fersa Bearings) and turbine disks for aircraft engines and aero-derivatives (ITP-Aero).
In the aerospace case, the Fausto project will enable the use of sensors integrated into the broaching machine to assess the wear level of each broach tooth, avoiding catastrophic failures. In grinding, advances in the advanced signal processing of acoustic emission sensors will allow for more precise detection of contact between the part and the grinding wheel.
The process fingerprint must be able to automatically identify process deviations. As ldeko reports, the use of AI will optimize production and diagnose defects more accurately and rapidly, leading to autonomous decision-making that improves the reliability and adaptability of production.
The second pilot is focused on Fersa’s plant in Zaragoza. This line consists of two flows, each corresponding to one of the bearing components: the inner ring (IN) and the outer ring (OUT). Each of these components undergoes various grinding and superfinishing processes, where the necessary functional surfaces are machined, including both inner and outer diameters and faces.
Physical and virtual sensors can monitor the grinding processes on the line to detect any kind of deviation or anomaly. A labeling station will be developed to measure the dimensional, topographical quality, and thermal damage of all produced parts. The integration of precise inspection systems will support full control and production autonomy. Additionally, digital twins will be used to predict the final quality of parts after processing, in terms of topography, dimensions, and surface integrity.
All developments, along with adaptable traceability systems, will generate a data stream that feeds AI algorithms, supported by synthetic data generated by simulation models, to ensure compliance with tolerances in the pursuit of autonomous processes.
