Abstract:

Rockburst is one of the main engineering geological problems greatly threatening the safety of construction. Prediction of rockburst is always an important project concerning the safety of workers and equipments in tunnels. The method of discriminant analysis is used to classify whether rockburst will happen in the underground rock projects and how much the intensity of rockburst is. Some main control factors of rockburst, such as the values of in-situ stresses σθ, uniaxial compressive strength σc and tensile strength σt of rock, the elastic energy index of rock Wet, are chosen in the analysis. Linear discriminant functions and criterion are obtained through training a large set of rockburst samples which come from a series of underground rock projects in domestic and abroad. Another samples are evaluated with the model. The evaluated results agree well with the practical records. Comparing the results of support vector machine method, fuzzy comprehensive evaluation method and artificial neural network method with discriminant analysis method, the evaluating results are same. Misjudgment ratio is low. Applying the classification model to rockburst prediction of the project of Kuocangshan Tunnel in Zhuyong Expressway, the result of classification is same as the results of fuzzy comprehensive evaluation method, artificial neural network method and numerical modeling method, just same as what happened in scene. It is shown that the classification model of discriminant analysis which with a low misjudgment ratio performed excellently with high prediction accuracy is an effective path to solve the problem of rockburst grade classification.

Key words: rockburst, classification, discriminant analysis, comparative analysis, misjudgment ratio