PetroVR Fire: a virtual reality training system for catastrophic petrochemical firefighting with enhanced collision detection

Firefighters are essential for mitigating hazards in petrochemical facilities, where rapid response and critical decisions to oil and chemical fires is critical. Traditional training methods suffer from high costs, safety risks, andlimited realism. To overcome these challenges, we present PetroVR Fire—an immersive Virtual Reality (VR) trainingsystem that reconstructs petrochemical plants using 3ds Max for 3D modeling and Unity for real-time interaction, deployed on PICO 4 headsets. A key innovation is the SAT-GJK algorithm, which fuses the Separating Axis Theorem(SAT) and Gilbert–Johnson–Keerthi (GJK) methods to enable precise, real-time collision detection in VR. This ensuresaccurate object placement, prevents interpenetration, and supports seamless user interaction. Experimental results showthat SAT-GJK improves system responsiveness and reduces latency, enhancing the fidelity and usability of thesimulation. PetroVR Fire offers a scalable, safe, and cost-effective alternative to live drills, significantly improvingfirefighters’ situational awareness and emergency decision-making in complex scenarios. While still in early-stageevaluation, this approach marks a pivotal step toward replacing resource-intensive physical exercises with immersive, safety-critical VR training. It lays the groundwork for future advancements in industrial emergency preparedness throughsimulation-based learning.