关键词: 变形预报, 嵌入维, 混沌特征, 井筒, SVM, 变形预报, 嵌入维, 混沌特征

Abstract: The forecasting and pre-warning for shaft disaster is an important guarantee for safety production of coal mine. Based on the analysis of deformation monitoring data of a Yanzhou mine shaft well，it is proposed to adopt maximum Lyapunov index and correlation dimension to describe the system's chaos characteristics. The embedding dimension is implied to determine the input dimension of SVM model in order to establish the relationship of deformation and time，the inside and outside precision is introduced to evaluate the model's accuracy and reliability. Compared to the prediction results of the BP neural network，it is found that the SVM model has good fitting capability as well as good predicting ability. This study has an important guiding significance to disaster monitoring and treatment of the main shaft.

Key words: Shaft, SVM, Deformation prediction, Embedding dimension, Chaos characteristic, Shaft, SVM, Deformation prediction, Embedding dimension, Chaos characteristic