top of page

Focus on Didactic Factories: E2-Lab experiment



The TERESA experiment, named E2-Lab: Self-evolving Monitoring Systems for Assembly Production Lines, aims to address the limitations in current diagnostic systems used in assembly production processes. The experiment is leaded by Jozef Stefan Institute (JSI) which is the technical coordination, then the University of Ljubljana's Laboratory of Control Systems and Cybernetics is conducting didactic exploitation of the platform, organizing seminars with students to develop and test new solutions, also Competence Centre for Advanced Control Technologies (KCSTV) is coordinating the experiment and providing administrative support, finally Domel, hosted by KCSTV, is a global manufacturer of electric motors and expert solutions provider that is hosting the demonstrator and providing insights into production specifics.  

Currently, the diagnostic systems on which the experiment is working, are either simplistic or highly specialized, making them impractical or insufficient for effectively detecting performance deviations and preventing process stoppages. Motivated by this challenge, the experiment seeks to develop and demonstrate approaches that can be widely applied to assembly production lines conducting repetitive operations. Building upon the principles explored in a SME driven experiment of the project, which focused on maximizing the availability, quality, and efficiency of molding machines, the TERESA experiment aims to generalize these concepts to other assembly production environments. 

The experiment's main concept revolves around providing a platform to support the development and demonstration of solutions based on semi-unsupervised data-based learning. This hybrid approach combines continuous data and sequential events to enable systems to self-learn from production operations, identify deviations, and provide concise insights into production states, operation deviations, fault occurrences, and fault source localization. 

Key objectives include preparing the experimental environment, including sensorics and platforms, developing demonstration algorithms, and conducting didactic exploitation to foster innovation and demonstrate practical applications in the production environment. Ultimately, the experiment aims to introduce new innovations that improve monitoring systems for assembly production lines, enhancing their efficiency and effectiveness. 



The experiment aims to create an industrial process environment for testing self-learning diagnostic systems in assembly production processes. These systems will use standard signals, ensuring easy installation without changes to existing control software or hardware. By recognizing patterns in process signals, the systems will create a digital representation (Digital twin) of the process, enabling simulation and virtualization. Edge devices will deploy the solution close to data sources, with results sent to a central system for visualization. The use-case involves final quality control at a DC drive assembly line, providing a range of process variables for analysis. This setup will allow the generation of operational data and simulation of various faulty conditions, facilitating the demonstration and comparison of different self-evolving monitoring systems. 

bottom of page