Numerical analysis of the performance of a PID control based real-time mechanical ventilation system to prevent smoke back-layering in tunnel fires

Abstract

Longitudinal mechanical ventilation systems are widely employed to prevent smoke back-layering in tunnel fires. Correlations for the critical velocity of ventilation in tunnels have already been extensively studied, for given fire and tunnel (cross-sectional) dimensions. This work describes the potential of a novel concept of an automatic ventilation system based on a PID (Proportional-Integral-Derivative) controller for preventing smoke back-layering under a priori unknown fire size. A few different tunnel dimensions and heat release rates (HRR) are addressed in order to illustrate how the ventilation velocity adapts automatically. Two methods to identify the location of the smoke front are proposed, based on temperature information. Simulation results show that the proposed approach can control the smoke well, preventing back-layering in real-time. It has excellent adaptability for the unknown fire size, which is considered an advantage compared to a system based on constant ventilation velocity. The smoke front becomes stable more quickly for higher HRR, while it takes longer to stabilize for increased tunnel height. This relates to the flame height: sufficiently high flames (compared to the tunnel height) induce a fast rise in temperature, which is the input parameter for the PID system. The tunnel width therefore has little impact. The values for the critical velocity and for the ventilation velocity at maximum smoke back-layering length (BL), as obtained with the proposed method, agree well with existing correlations. The influence of the set-point value for temperature is briefly discussed.