Workshop - Traceable fault diagnosis on electrical diagrams and PLC code via AI

When a plant stops, the time spent finding the cause of the fault often weighs more than the repair itself. The technician must consult electrical diagrams from different suppliers, trace back the PLC logic, and cross-reference information scattered across documents, software, and the experience of a few senior colleagues. This is valuable knowledge, difficult to transfer and increasingly scarce.
The workshop presents a different approach: transforming multi-vendor electrical diagrams and PLC code into a queryable knowledge graph, on which a reasoning engine reconstructs the path from symptom to cause. Unlike systems based solely on generic language models, every answer is traceable and verifiable against the original diagrams — an essential requirement in maintenance, where a wrong indication costs machine downtime and safety.
Through concrete cases, we will see how this approach can reduce diagnostic times, support technicians in troubleshooting, preserve company know-how, and turn technical documentation into an operational tool rather than an archive.