PTQ Q3 2026 Issue

between operations, safety, and engineering functions. Decisions can be based on consistent criteria rather than subjective interpretation of detection data. Practical application Field studies show that less than 1% of components are actively leaking at any given time in refinery operations, yet these leaks can contribute disproportionately to total emissions, highlighting the inefficiency of detection-only approaches. In practice, this workflow can be applied fol- lowing leak detection to support rapid classification: • If composition is dominated by light ends, dispersion is likely rapid and flammability risk limited. • If heavier fractions are present, localised concentration and ignition risk increase. • If operating conditions promote vaporisation, additional analysis may be required. This enables operators to move from detection to deci - sion using consistent criteria rather than conservative assumptions. Toward integrated and real-time risk assessment The capabilities demonstrated in this study provide a foun - dation for these developments, where composition esti - mation and dispersion modelling can be extended beyond scenario analysis into integrated, real-time workflows. This study reflects a broader shift toward integrated, data-driven safety workflows in refining. Increasingly, monitoring data, dispersion modelling, and analytical frameworks are being connected to improve emissions characterisation and source identification.4 This integration reduces reliance on manual data transfer between tools and enables more consistent application of assumptions across analyses. Emerging implementations extend this capability further by incorporating real-time data streams from plant sensors, inspection systems, and meteorological inputs.4 When combining process twins with AI-enabled analytics, these systems can support near real-time estimation of leak composition and dispersion behaviour following detection, reducing the time required to classify hazard severity from hours to minutes. Conclusion Process fluid leaks will remain an inherent challenge in refinery operations. While detection capabilities continue As refining systems become more complex and operating conditions more dynamic, the need for integrated, context-driven risk assessment will continue to increase. Detection alone is no longer sufficient

to improve, the ability to interpret and act on detection data remains the primary limitation. This study demonstrates that composition is the domi - nant factor governing dispersion behaviour and flammabil- ity risk under consistent conditions. Leak location alone is insufficient to determine hazard severity. By integrating process simulation with dispersion mod - elling, operators can classify leak severity more accurately, reduce uncertainty in response decisions, and align safety, operations, and engineering teams around a consistent analytical framework. As refining systems become more complex and oper- ating conditions more dynamic, the need for integrated, context-driven risk assessment will continue to increase. Detection alone is no longer sufficient. The ability to inter- pret, classify, and respond to leaks in a consistent and timely manner is becoming a critical component of process safety and operational performance. This integrated approach provides a practical founda - tion for improving safety outcomes, reducing uncertainty, and enabling more consistent management of leaks across refinery operations.

Flaretot and Petro-SIM are marks of KBC (A Yokogawa Company).

References 1 MacKay, J. E., Smith, K. J., Monitoring and Containment of Fugitive Emissions from Valve Stems – Electrical Conductivity and Gas Adsorption Measurements on Metal Oxides, Department of Chemical and Biological Engineering University of British Columbia, Vancouver, BC, Canada, V6T 1Z4. 2 Sun, X-B., Liao, C.-H., Liu, Z.-Y., Zhang, Y.-B., Liang, X.-M., Chen, L.-M., Ye, D.-Q., Assessment of VOC mitigation from equipment leaks in petroleum refining in Guangdong, China: Insights from a long-term leak detection and repair (LDAR) implementation, Petroleum Science, https://doi.org/10.1016/j.petsci.2026.03.008. 3 Shi J., Li, J., Li, Y., Xie, Z., Zhang, L., Proactive decision-making agent for industrial leakage and explosion emergencies powered by Physics GNN and LLM, J ournal of Loss Prevention in the Process Industries , Vol. 102, August 2026, 105976. 4 Hu, Y., Zhou, J., Wang, H., Dong, P., Zeng, X., Du, K., Hong Lin, Ren, G., Precise monitoring and source analysis of fugitive GHG emissions: A case study of Nansha, Guangdong, Processes 2026, 14, (9), 1344. .https://doi.org/10.3390/pr14091344 Rodolfo Tellez-Schmill , Ph.D., is a Product Champion for Process Simulation with KBC (A Yokogawa Company). He has more than 25 years of experience in chemical engineering activities, including pro - cess engineering, quality control, project management, research and development, technical support, and training. He earned a doctor- ate in chemical engineering from the University of Calgary and is a Professional Engineer registered in the Province of Alberta, Canada. Michelle Wicmandy , DBA, is a Marketing Campaigns Manager at KBC (A Yokogawa Company), specialising in digital transformation and process simulation for refining and process industries. She translates technical capabilities into practical business value, supporting technol - ogies that improve reliability, profitability, and emissions performance. Wicmandy also serves as a Senior Sustainability Advisor with The ESG Institute, contributing to initiatives focused on responsible indus- trial transformation.

PTQ Q3 2026