Research On The Prediction Of Critical Velocity Of Highway Tunnel Fire Based On Elman Neural Network

As an underground space building that can cross obstacles such as rivers,seas,lakes and steep mountains,tunnels have been used more and more in highway construction,but the fire safety problems of tunnels are becoming more and more serious.critical velocity is an important parameter in the vertical ventilation and smoke exhaust mode of highway tunnel fires.Its value determines whether effective fire rescue activities can be carried out when tunnel fire occurs.Therefore,it is crucial to make rapid and accurate determination of critical velocity.At present,most domestic and foreign scholars' research on critical velocity is based on the influence of single factor,but there are many factors affecting the critical velocity of tunnel fire,and it is difficult to accurately describe it with specific mathematical expressions.Therefore,based on the traditional numerical simulation method,Elman neural network theory is introduced to predict the numerical variation of tunnel fire critical velocity under various influencing factors.In this paper,a full-size horseshoe tunnel model is established with reference to the highway tunnel design specification.The two-lane highway tunnel is taken as the research object.using traditional numerical simulation method,the simulation experiments of five factors affecting the critical velocity of tunnel fire are carried out in the fire dynamics software FDS,including fire area,fire location,fire heat release rate,tunnel slope and obstacles in tunnel.The importance of each factor to the critical wind speed is clarified and compare the simulation results with previous research results.Its reliability and effectiveness were verified.Secondly,the five factors of fire source area,fire source location,fire heat release rate,tunnel slope and obstacles in the tunnel are taken as the network input parameters,and the critical velocity is taken as the network output parameter,And the normalized processed numerical simulation data is used as the sample set,so the critical velocity prediction model of Elman neural network tunnel fire is established.Finally,the performance of the Elman neural network critical velocity prediction model established in this paper is tested.The test results show that the maximum relative error between the predicted and expected values of the critical velocity of the model is 0.033,which can meet the accuracy requirements of fire engineering in highway tunnel construction,and compared with the numerical simulation method of critical velocity.The time cost of critical velocity calculation is saved to a large extent,and a new method for calculating the critical velocity of tunnel fire under various influencing factors can be provided.