基于时频谱模拟的时变混合相位子波提取方法研究

发布时间：2019-01-12 11:56

【摘要】：由于地层的吸收衰减作用,实际地震记录是一个动态衰减信号,从中提取的子波也是随时间变化的。传统的子波提取方法是基于静态褶积模型,提取一个单一时不变的子波,这与实际情况不符。为解决地层衰减作用造成传统子波提取方法提取结果不准确的问题,本文在分析研究常规分段时变子波提取方法存在不足的基础上,提出了一种基于时频谱模拟的时变混合相位子波提取方法。与分段子波提取方法相比,该方法充分展现了子波的时变特性,提取的子波更符合实际。针对多项式谱模拟方法在拟合子波振幅谱方面存在的局限性,本文提出了基于时频二次谱模拟的时变子波提取方法,进一步提高了时变子波的提取精度。首先,针对常规分段时变子波提取方法不能充分还原子波时变性,且提取子波不够准确等问题,本论文提出了基于时频谱模拟的时变混合相位子波提取方法。首先对非平稳地震记录进行改进的广义S变换,并引入时频滤波处理,然后采用多项式谱模拟方法从处理后的地震记录时频谱中拟合出每一时刻的子波振幅谱,在子波相位时不变的假设条件下采用高阶累积量的双谱法重构出子波的混合相位谱,最后将振幅谱和相位谱结合实现时变混合相位子波的提取。该方法突破了常规分段时变子波提取方法的不足,充分保留了子波的时变性。仿真实验和实际地震资料处理结果表明,该方法能够提取出更符合实际的地震子波,在地震资料处理中表现出明显的应用效果。其次,针对多项式谱模拟方法所用的拟合多项式限制了子波振幅谱的形态,且求解稳定性较差等问题,本论文进一步提出了基于时频二次谱模拟的时变子波提取方法。该方法在基于时频谱模拟的时变混合相位子波提取方法的基础上,采用二次谱模拟代替谱模拟方法,提高了时变子波振幅谱的提取精度。仿真实验和实际地震资料处理结果表明,该方法在充分保留子波时变性的基础上能够更精确地提取时变子波。最后,将基于时频二次谱模拟的时变混合相位子波提取方法与分段子波提取方法进行对比,理论上分析总结了本论文所提方法在提取子波时变性和准确性方面明显优于分段子波提取方法。仿真实验和实际地震资料处理结果表明,该方法突破了常规分段时变子波方法的缺陷,充分展现了子波的时变性,为精确提取地震子波提供了一个新途径。
[Abstract]:Because of the absorption and attenuation of strata, the actual seismic record is a dynamic attenuation signal, and the wavelet extracted from it also varies with time. The traditional wavelet extraction method is based on the static convolution model to extract a single time-invariant wavelet which is inconsistent with the actual situation. In order to solve the problem that the traditional wavelet extraction method is inaccurate due to the formation attenuation, this paper analyzes and studies the shortcomings of the conventional piecewise time-varying wavelet extraction method. A time varying mixed phase wavelet extraction method based on time spectrum simulation is proposed. Compared with the piecewise wavelet extraction method, this method fully shows the time-varying characteristics of wavelet, and the extracted wavelet is more practical. In view of the limitation of polynomial spectral simulation in fitting wavelet amplitude spectrum, a time-varying wavelet extraction method based on time-frequency quadratic spectral simulation is proposed in this paper, which improves the accuracy of time-varying wavelet extraction. Firstly, aiming at the problem that the conventional piecewise time-varying wavelet extraction method can not fully reduce the wavelet time-variant, and the wavelet extraction is not accurate enough, a time-varying mixed phase wavelet extraction method based on time-frequency simulation is proposed in this paper. Firstly, an improved generalized S-transform is applied to the non-stationary seismic records, and a time-frequency filter is introduced, and then the wavelet amplitude spectrum at every moment is fitted from the time-frequency spectrum of the processed seismic records by the polynomial spectral simulation method. Under the assumption that the wavelet phase time is invariant, the mixed phase spectrum of the wavelet is reconstructed by using the bispectral method of higher order cumulant. Finally, the amplitude spectrum and the phase spectrum are combined to extract the time-varying mixed phase wavelet. This method breaks through the shortcomings of the conventional piecewise time-varying wavelet extraction method and fully preserves the time-varying wavelet. Simulation experiments and actual seismic data processing results show that the method can extract more practical seismic wavelet, and shows obvious application effect in seismic data processing. Secondly, the fitting polynomial used in polynomial spectral simulation method limits the shape of wavelet amplitude spectrum, and solves the problem of poor stability. In this paper, a time-varying wavelet extraction method based on time-frequency quadratic spectral simulation is proposed. Based on the time-varying mixed phase wavelet extraction method based on time-frequency simulation, the quadratic spectral simulation is used instead of the spectral simulation method to improve the accuracy of amplitude spectrum extraction of time-varying wavelet. The simulation results and the actual seismic data processing results show that the proposed method can extract time-varying wavelet more accurately on the basis of fully preserving wavelet time-variant. Finally, the time-varying mixed phase wavelet extraction method based on time-frequency quadratic spectrum simulation is compared with the piecewise wavelet extraction method. In theory, the method proposed in this paper is superior to the piecewise wavelet extraction method in terms of time variation and accuracy of wavelet extraction. The simulation results and the actual seismic data processing results show that this method breaks through the defects of the conventional piecewise time-varying wavelet method and fully shows the time-varying wavelet, which provides a new way for accurate seismic wavelet extraction.
【学位授予单位】：中国石油大学(华东)
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P631.4

【参考文献】