The Coal Seam Thickness Prediction Method Research Based On Seismic Attributes' Multiple Regression Analysis

The research on seismic attributes began in the last century, and it has been widely used in coal, oil-gas exploration. The seismic attributes refer to physical quantities, which are derived from seismic data and related with the geometry, kinematics, dynamics and statistics characteristic of seismic wave. Seismic attribute technology is a technology to extract, store, visualize, analyse, verificate and evaluate seismic attribute by applying the study, algorithm development and integrated software system.There are over dozens of seismic attribute parameters extracted from the seismic data, but in the coal seam thickness prediction the use of more parameters doesn't mean better, because invalid parameters will increase the workload and consume limited resources, even lead to the curse of dimensionality. In order to use these attribute parameters effectively in prediction, they must be optimized.This paper researches on a method of predicting coal seam thickness through the application of seismic attribute technology. First, it extracted the information of seismic attributes and coal seal thickness around the drilling, and optimized and selected the seismic attributes by analysing the relevance between seismic attribute and coal seam thickness. Then taken the selected ones as the basic parameters for coal thickness prediction model, combined with the known drill data, it obtained the multiple regression prediction model between every seismic attribute and coal seam thickness using multiple regression analysis theory, and carried out significant regression test. The paper implemented the coal seam thickness prediction by applying the model, and achieved good result, which proved the feasibility of seismic attribute technology to predict coal seam thickness.