CASE STUDY 10
Safety AI
PETROCHEMICAL & ENERGY
Managing product yield and critical failures on a vibrating screen
Operator of one of the world's largest petrochemical complexes · polymer pellet sorting · 2024 – present · operating infrastructure
CHALLENGE
- The client sought to minimise downtime and maximise yield on the vibrating screen used for polymer pellet sorting in the EVA plant.
- Manual monitoring of the existing CCTV feed was unreliable and intermittent, and made worse by the vibration the equipment itself generates.
- Tears and clogs in the output tubes, and changes in product yield, were detected late — costing several million dollars a year.
SOLUTION
- Vivity Safety AI deployed for 24/7 automated monitoring on the existing camera infrastructure.
- Proprietary vision technology dynamically allocates the region of interest using a robust scene classifier and compensates for extreme vibration.
- Real-time analysis with instant notification, preventing downtime, reducing clean-up cost and limiting product loss.
RESULTS
- More than 99.9% accuracy in failure detection, with a false alarm rate below 0.07%.
- Immediate response to detected issues rather than discovery at the next manual check.
- Yield control moved from periodic sampling to continuous observation.