时空域与τ-p域预测反褶积压制多次波方法对比研究

Comparative Study on Multiple Suppression Method in Temporal Spatial Domain and τ-p Domain Predictive Deconvolution

预测反褶积是地震数据处理中广泛使用的方法,兼具参数设置简单和运行速度快等特点,正因如此,很多处理人员往往对两种方法影响参数、计算效率、保幅性等问题未进行深入研究,最终导致处理效果不够理想。本文通过对比时空域预测反褶积和τ-p域预测反褶积方法,阐述了两种方法的基本原理,分析了两种方法的主要影响参数,并分析对比两种方法的频谱变化、保幅效果等。研究证明,两种方法都具有相对保幅性,时空域预测反褶积主要影响参数为预测步长和算子长度。而τ-p域预测反褶积由于公式近似数值离散,会在τ-p正、反变换过程中带来假频,同时该方法深受τ-p域最大道数、采样间隔及最大频率等参数影响,在处理时需要更多实验反复论证参数有效性。实际数据结果表明,时空域预测反褶积计算效率是τ-p域预测反褶积的1.5倍,相对保幅性更好;而τ-p域预测反褶积在去除多次波效果、拓宽频带、提高低频、高频信息方面要优于时空域预测反褶积。本文研究方法对从业者方法选择和参数设置极具借鉴价值,对实际资料处理中提高去多次波效果、拓宽频带和保幅处理方面具有实际应用价值。

Predictive deconvolution,a widely used method in seismic data processing,has the characteristics of simple parameter setting and fast running speed.However,many processing technicians,for the lack of detailed research on the impact of the two methods on parameters,computational efficiency,amplitude preservation and other issues,often present unsatisfactory processing effect.This paper mainly compares the methods of temporal spatial prediction deconvolution and τ-p domain prediction deconvolution,expounds the basic principles,analyzes the effect parameters,and compares the spectrum change and amplitude preserving effect of the two methods.It is proved that both methods are relatively amplitude-preserving,and the main influencing parameters of temporal spatial prediction deconvolution are prediction step size and operator length.However,τ-p domain prediction deconvolution will bring false frequency in the process of forward and inverse transformation due to approximate formula and discrete values.Meanwhile,this method is deeply influenced by such parameters as the maximum number of channels in the τ-p domain,sampling interval and maximum frequency,which needs more experiments to further demonstrate the processing validity of parameters.The actual data show that the computational efficiency of temporal spatial prediction deconvolution is 1.5 times that of τ-p domain prediction deconvolution,and the relative amplitude preservation is better.But τ-p domain prediction deconvolution is better in removing multiple waves,broadening frequency band,improving low frequency and high frequency information.This research method is of great reference value for practitioners' method selection and parameter setting,and has practical application value for improving the effect of multiple wave removal,broadening frequency band and amplitude preserving processing in actual data processing.