The Forecast And Prevention Of Highway Tunnel Traffic Accidents

With the rapidly development of transportation, our country has the most complicated tunnels with the fastest developing speed in the world. However, the traffic safety situation of tunnel is not optimistic, tunnel has become a major distribution point of traffic accidents and accident black spots, and has a major accident hazard degree, aftertreatment dealing with hardly and causing secondary accidents easily etc., so improving tunnel safety management level in China is still an important topic which facing the tunnel transportation management.Based on presenting traffic accident characteristics and accident forms of tunnel, this paper analyses the influence factor systematically from direct cause, indirect causes and basic causes of tunnel accidents, expatiates the relationship between traffic accidents and factors for building up a forecasting model for tunnel traffic accidents.Using grey relational theory to analyze the influence causes, choosing seven higher correlation indicators (GDP, the tunnel length, passenger turnover, freight turnover, the number of civil passenger, freight vehicles and driver) as the final indexes of the forecast model.After analyzing the methods of traffic accident forecast and its characteristics, according to the characteristic that the traffic system is dynamic and time-variance and the advantage that ANN has in solving complicated and nonlinear problems, analysis the feasibility of BP Neural networks using to forecast of tunnel traffic accidents. And confirm using it for forecasting tunnel traffic accidents.This dissertation builds up a BP Neural networks model of tunnel safety prediction, and discusses some key technique and means applying the model, including the selection of input-output variables, the selection and pretreatment of swatches, conforming the number of the nodes in hidden layers, the selection of the initial weight and value, the selection of activation function, training arithmetic as well as parameter. At last, through the example of the tunnel safety prediction by the data of our country's tunnel accidents from 1995 to 2008, it validates that this method opens out the relation of the tunnel safety and its influence factors in some error bound, and could be applied to forecast tunnel traffic accidents.Finally, on the basis of above studied, this paper establishes the prevention system of tunnel traffic accidents, and puts forward preventive countermeasures from beforehand (reducing the possibility of accidents) and afterwards (alleviating the loss of accidents) for tunnel traffic safety.