Prediction of maximum ceiling temperature rise in inclined tunnel fire based on improved non-orthogonal ventilation plume model

• Reveals structure-ventilation coupling mechanisms affecting fire plume behavior and T max . • Explains effects of non-orthogonal ventilation and asymmetric entrainment. • Ignoring the non-orthogonal effect causes 15 – 50% systemic prediction error for T max . • Improves the traditional T max prediction model based on orthogonal ventilation. • Validates predictive model with errors within ±20% in high-slope tunnels. With the increasing construction of inclined tunnels, evaluating fire-induced thermal environments remains challenging, as most predictive models are developed for horizontal or mildly inclined tunnels. For tunnels with steeper slopes, additional physical mechanisms introduced by tunnel geometry are often overlooked, leading to systematic deviations in prediction accuracy. In this study, the Fire Dynamics Simulator (FDS) was used to investigate fire plume behavior and ceiling temperature distribution in inclined tunnels under natural ventilation. Results show that increasing tunnel slope causes the plume impingement point to shift upslope and significantly reduces the maximum ceiling temperature. This phenomenon is primarily attributed to the elongation of the plume’s vertical rise, which enhances cold-air entrainment and intensifies cooling effects. Further analysis reveals two dominant mechanisms responsible for this behavior: 1) the asymmetric entrainment effect, where upslope ground blockage weakens plume entrainment on the high-slope side and induces stronger lateral deflection; 2) the non-orthogonal ventilation effect, where induced flow accelerates vertical rise while reducing the deflection angle. Under high-slope conditions, these effects become more pronounced, resulting in a systematic deviation of 15–50% for classical models based on orthogonal ventilation. By introducing a modified dimensionless velocity and a slope-dependent entrainment correction factor, an improved model is developed. Validation using independent experimental datasets shows that the majority of predictions fall within a ±20% error range, demonstrating substantially enhanced accuracy. This work elucidates the key mechanisms of structure–ventilation coupling on plume behavior in inclined tunnels and provides a predictive model for the maximum ceiling temperature with improved accuracy and applicability at steep slopes.