Health Assessment Of Operating Expressway Tunnel Based On GA-BP Neural Network

Applying scientific and accurate methods to evaluate the operation of highway tunnels can accurately evaluate the degree of disease,ensure the safety of tunnel operation,avoid traffic accidents,make timely targeted maintenance of tunnels with limited resources,and prolong the service life of tunnels.At present,there are many methods for evaluating the structural health of operating highway tunnels,most of which are based on subjective experience and qualitative analysis,resulting in a strong subjectivity of the evaluation results,timely adoption of impact measures and later maintenance work.BP neural network makes use of a large number of data samples to train the network model,which makes the health evaluation results of operating tunnels more accurate.Combining with the main contents of tunnel special inspection,the typical diseases of tunnel are analyzed,and the relevant diseases of tunnel in accordance with the principles are selected as evaluation indexes,and the health condition of operating highway tunnel is evaluated according to the detection data of each evaluation index.This paper designs the health condition evaluation set of the tunnel running expressway,formulates the quantitative standard of health grade,and establishes five health grades.The principle of BP neural network is analyzed in detail.Aiming at the defect that BP neural network is easy to fall into local optimum in training process,this paper proposes to combine the advantages of genetic algorithm and neural network,and select genetic algorithm to optimize the initial weights and thresholds of BP neural network.Through the training of historical disease data and expert evaluation results of Baoding section tunnel of Zhangzhuo Expressway,using the learning ability and non-linear processing ability of BP neural network,based on MATLAB software,the evaluation model of BP neural network optimized by genetic algorithm is established.Using the established model to test,not only can the tunnel health evaluation value be accurately and quickly obtained,but also the error accuracy is greatly improved compared with that before optimization.