论文部分内容阅读
煤层与岩层的顶、底板间形成的地震反射强度的变化将引起煤层厚度预测结果的变化.首先提取了目的层的反射波的地震属性,然后深入研究、分析了这些属性特性,最后应用BP人工神经网络方法以及多项式回归分析的方法进行研究,并在实际地震资料中应用这些方法,结果表明:在对矿区的煤厚预测中,BP人工神经网络模型误差最小,多元二次回归次之,多元一次线性回归模型误差最大,证明了用多属联合分析技术进行煤厚预测是一种卓有成效的方法.
The change of seismic reflection intensity between the roof and floor of coal seam and rock formation will lead to the change of prediction result of coal seam thickness.Seismic attributes of the reflected wave of the target layer are firstly extracted and then studied and analyzed in detail. Finally, BP artificial Neural network method and polynomial regression analysis, and apply these methods to the actual seismic data. The results show that BP artificial neural network model has the smallest error, the second is the multivariate regression, the second is the multivariate regression The linear regression model has the largest error, which proves that it is an effective method to predict the coal thickness by using multiple joint analysis techniques.