Loss prediction of collapse hazard of a Tibetan tunnel with grey clustering algorithm

Collapse is a typical hazard in highway tunnel excavation process, in order to evaluate the collapse risk level of tunneling activity, the potential loss should be estimated of five severity levels. In this paper, an evaluation algorithm based on the grey clustering theory is proposed, the influential factors of segmental collapse hazard encompass the rock mass classification, distance from the portal, excavation method, ground water content, and embedment depth. The influential factors are input to a grey clustering model, five triangular whitening weight functions are deployed to calculate the relative importance of the influential factors. the loss severity prediction algorithm is applied to the Liuwu tunnel in Tibet autonomous region. The analytical result have shown that two of the ten segments have a fourth grade potential loss severity level, particular countermeasures should be taken to reduce the risk level to a tolerable degree.