Fault interpretation using a support vector machine: A study based on 3D seismic mapping of the Zhaozhuang coal mine in the Qinshui Basin, China

Abstract

The identification of small faults is of great significance to safe and efficient coal mining. In this study, a support vector machine (SVM) algorithm was applied to fault recognition problems using multi-attributes acquired from the Zhaozhuang coal mine in Shanxi Province, China. Using a fault model with varying throws, we analyzed the response characteristics of eight seismic attributes in the coalfield. Correlation and cluster analyses were used to select five attributes—variance, curvature of orientation, dip deviation, dip continuity, and maximum amplitude—as training data. The fault model revealed that small faults respond weakly to these attributes. Consequently, we developed an improved method to enhance the response of small-throw faults. A particle swarm algorithm was used to find the optimal parameters for an SVM model to identify faults in the Zhaozhuang mining region. Subsequently, experimental results showed that the SVM model produced accurate fault recognition results.