AbstractThe statistical characteristics of structural defects of shield and cast-in-situ subway tunnels are of great significance for risk assessment, preventive maintenance, and reduction of safety accidents. In this study, the apparent defect indexes, statistical characteristics, probability models, and index correlations of different lining types of subway tunnels are studied. First, quantitative and objective statistical indicators suitable for modern subway tunnel detection technology are proposed. Then, the statistical characteristics of defect data from the two different tunnel lining types are analyzed. Using the maximum likelihood estimation method, a probability analysis of the defect indexes is carried out, and the quantile test method is used to verify the probability analysis results. Finally, the correlation between defect indexes is discussed through correlation and regression analysis. The results indicate that the influence of different lining types on the defect characteristics of subway tunnels is different, and the corresponding statistical characteristics are significantly different. The probability distribution models of the two tunnel types are obviously different due to different quantitative bases of defect index severity, and appropriate probability distribution models for some indexes cannot determined. There is a certain correlation between the defect indexes of shield tunnels, but this correlation is weak in cast-in-situ tunnels. The safety evaluation index of tunnel structures can be selected according to the index correlation, and type and weight of the index of different lining types should be treated differently.