Abstract This paper introduces a predictive Health, Safety, and Environment (HSE) model designed for proactive risk mitigation in PTTEP's operations. The primary objective is to enhance operational safety management by leveraging generative AI, machine learning, and real-time analytics. The model includes descriptive, predictive, and prescriptive analytics to identify incident causes, predict potential safety hazards, and provide recommendations to prevent incidents. This approach transforms HSE management from reactive to proactive, improving workplace safety. The model development follows a structured process, leveraging cloud-based data integration and AI-driven analytics. A comprehensive HSE data lake aggregates incident reports, near-miss events, and operational data using cloud-native data warehousing and ET L processes. Data is cleaned and preprocessed before training machine learning models using generative AI and machine learning frameworks. The predictive model identifies patterns linking past incidents to future risks and automates precursor alerts while prescriptive analytics generate effective recommendations. The model is planned to be deployed in real-time environment, ensuring continuous learning and adaptation. Currently in the proof-of-concept phase, the project has successfully developed a comprehensive HSE performance and incident analysis dashboard, utilizing root cause analysis and heat mapping to visualize risk factors in drilling and wellhead construction. Predictive analytics are being trained to identify high-risk scenarios and precursors to potential incidents, enabling proactive interventions. Prescriptive analytics will further refine risk mitigation strategies. Early results indicate a projected reduction in risk levels from 5E to 5D, signifying a decreased incident likelihood. The model is expected to reduce PTTEP’s Total Recordable Incidents (TRI) by 30% of the average cases in the past 3 years, translating to estimated cost savings of 627, 000 annually. Beyond financial impact, the initiative enhances safety culture and operational efficiency, reinforcing PTTEP’s commitment to HSE excellence. The current industry solutions focus on reactive incident analysis, lacking the predictive capabilities introduced in this innovative approach. The integration of descriptive, predictive, and prescriptive analytics marks a significant advancement in operational safety management. The model's utilization of AI-powered pattern recognition for continuous improvement further distinguishes it, enabling constantly refine predictive accuracy with new data. The model's versatility allows for its application across various operational settings within PTTEP and beyond.
Thienthanukit et al. (Mon,) studied this question.