Towards Automated Safety Vetting of PLC Code in Real World Plants - Mu Zhang

Explore a comprehensive approach to automated safety vetting of Programmable Logic Controller (PLC) code in industrial settings through this IEEE conference talk. Delve into the VetPLC system, which combines static program analysis and temporal invariant mining to generate timed event sequences for detecting hidden safety violations. Learn about the challenges faced by traditional vetting methods and how VetPLC overcomes them by creating timed event causality graphs and leveraging data from Industrial Control System testbeds. Examine the implementation and evaluation of VetPLC across 10 real-world scenarios, demonstrating its superiority over existing techniques. Gain insights into collaborative environments, event-driven programs, and the intricacies of PRC code analysis. Understand the methodology behind VetPLC's program analysis, data mining, and event sequence generation, concluding with comparative experiments and a summary of its effectiveness in enhancing industrial safety.