径向基函数神经网络在砂体厚度预测中的应用

目前,人工智能神经网络在地震储层参数的预测方面具有广泛的应用,最常用的为BP神经网络,但是效果并不是十分理想。径向基函数神经网络(RBFN)是一种前馈神经网络,其在函数逼近、模式识别方面都优于BP网络,已经在岩性识别、孔渗预测方面取得了较好的应用效果。本文首次将此方法运用于预测砂体厚度,利用地震属性信息和神经网络的学习,基于实际数据计算,最后计算出相应的砂体厚度值,并与实测值进行误差分析。实例分析表明,利用径向基函数神经网络进行砂体厚度预测具有一定的可行性和实用价值。

Application of Radial Basis Function Neural Network in Thickness Prediction of Sand Body

At present, artificial intelligence neural network has widely applied in prediction of seismic reservoir parameters, the most commonly used one is BP neural network, but its effect is not very ideal. Radial basis function neural network （RBFN） is a kind of feedforward neural network; it is superior to the BP network in the aspects of function approximation and pattern recognition, so it has already gained good application effect in lithology identification, permeability and porosity prediction. In this paper, this method was first applied to predict sand body thickness. Through using seismic attribute information and neural network learning, based on the actual data, finally the corresponding sand body thickness value was calculated, and error analysis was carried out by comparing with the measured values. Example analysis shows that, using radial basis function neural network to predict thickness of sand body has certain feasibility and practical value.