November 2023
With the Deliverable D5.1 being finalised for the work on Digitalisation, Automation, Industry 4.0, Baptiste Schubnel (CSEM) explains us why this work is being done:
“Why do we do work on digitalisation / have Industry 4.0 as goal in PILATUS, a PV line production upscaling project?” Digitalisation is about transforming processes and creating value with digital technologies. In a manufacturing environment, this transformation relies on Industry 4.0 concepts. In the PILATUS project, the goal is to use Industry 4.0 to enhance the yield and uptime of production lines, to optimise products across the supply chain, and to gain deeper insights using inputs from production lines and field performance data. Digital technologies enable large data collection throughout the PV value chain, from cell and module production to field performance. In parallel, novel machine learning algorithms enhance our ability to harness this data to improve manufacturing quality and yield.
For instance, pattern recognition algorithms can detect recurring defect patterns on production machines in real time, and, by alerting machines operators to clean or repair the machine, minimise waste and improve PV manufacturing quality. In PILATUS, we have a particular focus on a class of algorithms called “Causal Machine Learning” that can be used to search the fundamental causes behind faults or system behaviours from data. This insight enables operators and researchers to refine manufacturing processes and improve result quality.
“What will we (likely) achieve on the digitalisation topic in the PILATUS project?” A key result of PILATUS project is the setup of a dataspace for data collection and exchange throughout the PV value chain, aligned with the IDSA architecture standard. This progress is detailed in deliverable D5.1, and the specified infrastructure is now operational at CSEM and our project partners. It allows data exchange in multi-partner scenarios, while maintaining data sovereignty and setting clear usage policies among data sharers and data consumers. Algorithmically, we are currently developing with our partners innovative algorithms to enhance the yield of their manufacturing lines and to improve the reproducibility and outcomes of some of their manufacturing processes.