基于小波神经网络的薄互储层参数预测

THE PREDICTION OF THIN INTERBEDDED RESERVOIR PARAMETERS BASED ON WAVELET NEURAL NETWORKS

针对薄互储层沉积具有储层厚度薄且横向变化剧烈的特点,及油气藏所具有的复杂性,本文提出了一种新的薄互储层参数的预测方法——小波神经网络技术,小波神经网络是基于小波分析理论所构造的一种新的神经网络模型,它充分利用小波变换良好的局部化性质,并结合神经网络的自学习功能,因而具有较强的逼近能力,从而提高薄互储层参数的预测精度,并通过实例验证了此方法的正确性。

Considering the features of the thin interbedded reservoir that its thickness is thin and change violently in horizontal, and the reservoir is complex, in this paper a new prediction method based on wavelet neural network for thin interbedded reservoir parameters is presented. Wavelet neural network is a new neural network model that based on wavelet analysis theory. With the good partial property of wavelet and the self-learning capability of neural network, it possesses strong ability of approaching and it improves the predict accuracy of thin interbedded reservoir parameters. The experimental results prove that this algorithm is feasible.