Description
This study presents techniques and cases of the calibration of 1D tunnel ventilation network models based on real-time observational data of operational road tunnels. The use of 1D network models in the design phase of a tunnel ventilation system is common, but less common is the use, maintenance, or calibration of the model once the system is in operation. Keeping a ventilation model of a tunnel ventilation system up to date has benefits in troubleshooting critical problems, running fire simulations and in assisting in any further design improvements or extensions to the system. Additionally, it allows the use of the model to optimise fan and fan power usage for actual traffic conditions.
Simulation carried out in the design stage of a system is done with the best available information, but nonetheless will likely produce results different from those that occur during operation. Incorporation of Lidar scan data of tunnel shapes, sensor readings, and actual traffic flows into the model improves its accuracy. This creates confidence in the model and makes possible its use in plans to change or optimise the system. The further introduction of electric or hydrogen vehicles, and the resulting changes in fleet compositions and emission profiles make future changes to ventilation requirements more likely. Therefore, the requirement for future design changes and studies during the life of the tunnel will increase the need for calibrated models.
This paper first presents current 1D ventilation simulation techniques, then reviews them in the context of calibrating a model based on real time sensor and traffic data, outlining the difficulties and challenges encountered during calibration. Also presented in this work are several cases where the techniques have been successfully applied to models of operational road tunnel ventilation systems, where the resulting model has been used to assist in reconfiguration of fan usage in the tunnel.
