CPC Corporation, Taiwan hosted the "2026 Occupational Health and Safety Seminar" on August 31, 2026, at the Hongnan Activity Center of its Refining Business Division. At the event, a joint paper titled "Application of Intelligent AI Imaging and Precision Detection Systems for Gas Leak Detection in Petrochemical Plants"—co-authored by TVCM Linyuan Plant, the Labor Inspection Office of the Kaohsiung City Government, and China Gulf Oil & Storage Corporation (Hua Yun Warehousing Co., Ltd.)—was selected for an oral presentation in the "Smart Industrial Safety Applications" session. The paper was co-authored by Guo, Qing-Ji (Kaohsiung Labor Inspection Office), Tsai, Ming-Kuang (TVCM), Hsieh, Hui-Nan (Hua Yun Warehousing), and Yu, Lin-Hsuan (TVCM), demonstrating concrete achievements in advancing smart industrial safety through public-private collaboration.
Petrochemical plants feature vast networks of equipment components and high-frequency tank truck loading/unloading operations, making continuous round-the-clock monitoring difficult through traditional manual inspections. Unnoticed gas leaks can significantly elevate the risks of fire, explosion, and other severe industrial accidents. Guided by its core strategy of "Safety First," TVCM partnered with government authorities and logistics partners to establish a three-tiered, defense-in-depth framework for smart monitoring and industrial safety protection, comprising:
- Level 1 — Plant-Wide Monitoring: Plant-wide VOC Diffusion Monitoring Platform
- Level 2 — Area Protection: 24/7 Disaster Prevention and Precision Detection System
- Level 3 — Operational Control: Smart Weighbridge and AI Vision Industrial Safety Recognition System
Through plant-wide monitoring, area protection, and on-site operational controls, this three-tiered approach enhances anomaly detection, real-time notification, and risk management capabilities. The paper also shares the implementation process, practical case studies, and quantified benefits of these three systems.
Among these, the 24/7 Disaster Prevention and Precision Detection System successfully issued an alert 3 minutes and 23 seconds faster than fixed gas detectors during an actual micro-leak incident. This allowed field personnel to take early emergency response measures, significantly strengthening accident early warning and risk management capabilities. Furthermore, according to paper estimates, preventing a major leak event avoids an estimated loss of NT$63 million caused by unplanned shutdowns, while mitigating the risks of environmental fines and perimeter odor complaints.
Going forward, TVCM will continue to deepen the application of AI and smart technologies in industrial safety management, enhance anomaly detection and emergency response capabilities, and share practical experience with industry peers as it works toward the goal of “Zero Accidents and Zero Tolerance.”