时频属性法薄互层预测

Thin inter-bed prediction with time and frequency attributes

1/4波长以内的薄互层厚度预测一直是勘探难题,本文主要探讨利用时频属性预测薄互层累计厚度的方法。通过薄互层楔状模型和两个叠置薄层模型,厘清薄互层累计厚度与时频属性的关系及其主要影响因素。模型研究结果表明:净毛比、层数和互层模式都对峰值振幅—毛厚度关系有影响,其中净毛比影响最大,其次是互层数,影响最小的是互层模式;净毛比控制峰值振幅、峰值频率与毛厚度关系,也控制调谐厚度范围;在调谐厚度内,峰值振幅(积分能谱)与毛厚度呈单调递增线性关系,峰值频率与毛厚度呈单调递减线性关系;积分能谱与净厚度呈线性关系,积分能谱有扩大调谐厚度范围功能,有利于计算薄互层净厚度。根据模型试算结果,在实际资料中运用人工智能方法,利用峰值振幅和峰值频率等联合计算薄互层净厚度。预测结果表明本文方法准确、可靠,实用性强。

It is very difficult to predict thin inter-beds with its thickness of the 1/4 wavelength.This paper mainly discusses a method for predicting the cumulative thickness of thin inter-beds with time and frequency attributes.Based on a thin inter-bed wedge model and two stacked thin inter-bed models,the relationships between cumulative thickness and time-frequency attributes of thin inter-bed and the main factors that influence the relationship are determined.The following understandings are obtained based on the model tests:(1) The net-togross ratio,the number of thin beds,and the interbed mode affect the relationship between the peak amplitude and the gross thickness,the net-to-gross ratio has the largest influence,then the number of thin beds,and the inter-bed mode;(2)The net-togross ratio controls the relationship between the peak amplitude(integral energy spectrum)and the gross thickness and the relationship between thepeak frequency and the gross thickness,it also controls tuning thickness range;(3)In the tuning thickness,the relationship between the peak amplitude and the gross thickness shows a single increasing linearity,and the relationship between the peak frequency and the gross thickness shows a single decreasing linearity;(4)The integral energy spectrum and the net thickness are linear,and the integral limit can expand the tuning thickness range,which is beneficial to calculate the thin bed net thickness.In practice,we calculate a thin bed net thickness with the artificial intelligence based on the peak amplitude and peak frequency,and the results show that the proposed method is reasonable and reliable.