Intelligent Control Study Of Longitudinal Ventilation System With Air Exchange Shaft In Highway Tunnel

As a closed underground space,long highway tunnels need to be ventilated to dilute and discharge vehicle pollutant emissions under normal operation to drivers&passengers for safe-comfortable-healthy driving environment,and to control smoke movement and provide enough fresh air on fire mode to drivers&passengers and rescue team for favorable escape and rescue environment.In our country,the highway tunnels with length greater than 5000m are usually longitudinal ventilated through jet fans combined with air exchange by shaft,whose efficiency depends on the combined action both jet fans and air exchange fans.How to control the complicated ventilation system is an important issue for long highway tunnels construction and operation under normal and emergent operation.Based on Fenshuiling long tunnel of Zhang-Zhuo Expressway,applying fluid mechanics and aerodynamics theory,the characteristic of longitudinal ventilation system with air exchange shaft,the distribution of fire smoke temperature and concentration,the intelligent optimization control of ventilation system under normal and emergent operation were studied through model test,numerical simulation,field test,and artificial intelligence technology.A series of valuable achievements were got.(1)Based on fluid dynamics mechanics and model experiment affinity theory,the test model of longitudinal ventilated Fenshuiling tunnel with single-shaft was developed,and it’s reliability was verified by the basic performance test of model system.The influence test of ventilation system,the impaction of gradually shutting down fans under normal operation and the fire smoke movement of tunnel in case of fire were tested,which provided the theoretical basis for the next optimal control of ventilation system.(2)A particle swarm optimization and Gauss process regression(PSO-GPR)model for predicting the traffic volume of tunnel is proposed based on artificial intelligence and modern optimization theory.Seen from the traffic volume prediction result of one highway tunnel,the maximal relative error of this model is less than 5%and the average relative prediction error is 1.96%.The model can be applied to the forecast of short-term traffic volume,and provides real-time traffic data for the feed forward control of ventilation system during tunnel operation.(3)Based on the traffic volume forecast and the three-dimensional numerical simulation results of Fenshuiling tunnel during normal operation,in order to control the maximum CO concentration of 1.8m high axial longitudinal section of tunnel,the evolutionary support vector regression model used to optimize the ventilation system control during normal operation was established.Using the model,the control of ventilation system under different driving speeds and control targets can be optimized,which was verified by the field test data.(4)Based on the three-dimensional numerical simulation results of Fenshuiling tunnel in case of fire and the evolutionary Gaussian process regression algorithm,the constrained multi objective programming model of tunnel ventilation system optimization in case of fire was proposed,whose aimed to reduce the smoke temperature and concentration of 1.8m high 30m downstream the fire source when the fire smoke backflow was not occurred.(5)Based on the modern optimization theory,the PSO and BP neural network coupling model of the optimization of tunnel ventilation system under fire condition is established by using penalty function method.Combined with the fire numerical simulation,the model can realize the constrained multi-objective programming of the tunnel ventilation system,which is a new method to intelligent control the ventilation system in case of fire.