Trunnel Damage Identification And Health Evaluation Based On Ant Colony Clustering Algorithm

As a new method of researching tunnel and tunnel health damage identification,ant colony clustering algorithm has been widely used in slope,coal,wall rock and other fields for its superior robustness,excellent calculation and easily compatible with other algorithms.But the ant colony clustering algorithm has not been applied to the determination of tunnel health and tunnel damage identification.There is some subjectivity in using fuzzy comprehensive evaluation method and other traditional methods in tunnel health determination and damage identification.Therefore,the paper uses ant colony clustering algorithm to study the tunnel health identification and tunnel damage identification aiming at the exposed diseases characteristics in tunnel running period and the apparent damage of tunnels which is based on the data mining.The paper identified the damage extent of the tunnel which used the crack width as the research object.At the same time,the paper joined the genetic variation thought to improve the algorithm after deeply understanding the ant colony clustering algorithm.What is more,the paper applied the improved ant colony clustering algorithm to the damage identification in tunnel and the idea is achieved by using Matlab software.The model established 7 different levels of damage grade standard by simulating experiment to cluster three indicators data including the corrosion time,crack width and load,and clustering center of each index is obtained,and the standard of 1-5 scale is established.by analying the four groups of samples,so as to realize tunnel damage identification.After analyzing the factors of affecting tunnel health,the paper preliminary selected the 11 decision index from lining cracks,cavities behind the lining,lining material deterioration,lining delamination and spalling and leakage water and so on.By using the R-style cluster,rank and the RankSum test to screen the 11 determinant index so as to establish the tunnel health index system.Utilizing the improved ant colony clustering algorithm to train the model and build the 3 kinds of tunnel health grade to determine the health level of the new sample to be tested.The main conclusions are as follows:1)The damage level of tunnel test sections of 4 groups is basically the same as the actual damage grade,which verifies the feasibility of the improved ant colony clustering algorithm in tunnel damage identification.2)The health level of 23 tunnel sections is basically the same as that of the actual judgment.Among them,The accuracy rate of the tunnel health level of the sub-health section reached 100%,and the accuracy rate of the tunnel health level of the disease section and the health section was 91.3%,which verifies the feasibility of the improved ant colony clustering algorithm in the tunnel health decision application.This study clarifies the characteristics of ant colony clustering algorithm and improves the ant colony clustering algorithm.It provides a new idea and new method for the application and extension of ant colony clustering algorithm in tunnel damage identification and tunnel health judgment.This study has certain theoretical significance and practical value.