基于Q值自适应优化的稳定反Q滤波方法

发布时间：2018-03-11 11:51

本文选题：反Q滤波　切入点：稳定性控制　出处：《东北石油大学》2017年硕士论文　论文类型：学位论文

【摘要】：由于实际地球介质存在粘性吸收,地球介质的小尺度非均匀也产生类似于粘性吸收的幅值衰减效应,这些客观存在导致了地震波在传播过程中发生幅值的吸收衰减,而在油田勘探、开发的中后期,提高对深层薄砂体、薄互层、微幅构造等隐蔽油气藏的勘探能力,增强对小断层、裂缝发育带等油气疏导体系的识别能力,进而识别有利的油气储层,是目前地震勘探的主要目标。这就需要针对粘性介质吸收与频散进行有效补偿和校正。其中反Q滤波就是一种目前常用的补偿方法。常规的反Q滤波方法一般基于波场延拓理论,具有不稳定性或振幅补偿不足的缺点,补偿高频能量的同时也不可避免地增强了高频噪声,出现不稳定现象,且通常的时不变增益限约束的反Q滤波振幅补偿虽可控制地震资料信噪比,但其经常会压制深层地震波的高频成分,降低深层地震资料的分辨率。因此,本文针对上述两个问题,在前人研究的基础上,提出了二次函数法、三角函数法2种新的稳定性控制方法,设计了两种新的反Q滤波补偿算子。并利用理论合成数据模型验证期有效性,并与稳定因子法对比分析,分析结果表明:2种新的稳定性控制方法均可以有效增强反Q滤波过程中的稳定性,并且相比于稳定因子法具有更强的振幅恢复效果,更加明显的改善深层的分辨率。以这两种稳定性方法为基础,将自适应增益限的思想与稳定性控制方法结合,使优化的振幅补偿函数的增益限自适应于地震数据有效频带的截止频率,通过理论模型试验,分析其时域及频域补偿效果,深层补偿效果明显提升,验证了其有效性。由于薄层调谐等现象存在,基于反射地震资料确定Q值则变的更加困难,因此,在以上研究的基础上,本文还对克服薄层调谐现象的Q值建模方法进行研究,提出一套基于频率导数的算术平均值来剔除薄层调谐效应的影响,利用Q值扫描来判断频率恢复状况,从而实现用反射资料估计等效Q值的方法;通过改变以往从地震衰减数据出发求取叠加Q值的思想,提出了先对地震数据进行粘性补偿,将地震波损失的能量进行补偿,使补偿后的数据展现出没有粘性衰减特征。消除地震数据经过一定的传播距离后,幅值过低而受到背景噪声和多次波等因素的影响。算法实质就是选取正确的叠加Q值使深、浅地层子波的频谱谱宽保持一致。并将这种Q值建模方法应用与理论模型和实际地震资料处理中,充分验证了其可实用性。
[Abstract]:Due to the existence of viscous absorption in the actual earth medium, the small scale nonuniformity of the earth medium also produces amplitude attenuation effects similar to the viscous absorption, which leads to the absorption and attenuation of the seismic wave in the process of propagation. In the middle and late stages of oil field exploration and development, the ability to explore subtle reservoirs, such as deep thin sand bodies, thin interbeds, micro-structures, and so on, is improved, and the ability to identify oil and gas dredging systems such as small faults and fracture development zones is enhanced. In order to identify favorable oil and gas reservoirs, It is necessary to compensate and correct the absorption and dispersion of viscous media effectively. The inverse Q filter is one of the commonly used compensation methods. Yu Bo's field extension theory, It has the disadvantage of instability or insufficient amplitude compensation. The compensation of high frequency energy inevitably increases the high frequency noise and instabilities. The amplitude compensation of inverse Q filter with time-invariant gain limit can control the signal-to-noise ratio of seismic data, but it often suppresses the high frequency component of deep seismic wave and reduces the resolution of deep seismic data. In this paper, two new stability control methods, quadratic function method and trigonometric function method, are proposed on the basis of previous studies. Two new inverse Q filter compensation operators are designed, and the validity of the period is verified by the theoretical synthesis data model, and compared with the stability factor method. The analysis results show that the two new stability control methods can effectively enhance the stability in the inverse Q filter process, and have a stronger amplitude recovery effect than the stability factor method. Based on these two stability methods, the idea of adaptive gain limit is combined with the stability control method. The gain limit of the optimized amplitude compensation function is adapted to the cutoff frequency of the effective frequency band of seismic data. Through theoretical model test, the compensation effect in time domain and frequency domain is analyzed, and the deep compensation effect is improved obviously. Because of the phenomenon of thin layer tuning, it is more difficult to determine Q value based on reflection seismic data. Therefore, based on the above research, the Q value modeling method to overcome thin layer tuning phenomenon is also studied in this paper. A set of arithmetic average method based on frequency derivative is proposed to eliminate the influence of thin layer tuning effect, and the frequency recovery condition is judged by Q value scan, so that the equivalent Q value can be estimated by reflection data. By changing the idea of calculating the stack Q value from seismic attenuation data in the past, the viscous compensation of seismic data is put forward, and the energy loss of seismic wave is compensated. The compensated data show that there is no viscous attenuation, and the amplitude of seismic data is too low after a certain propagation distance, which is influenced by background noise and multiple waves. The algorithm is essentially to select the correct superposition Q value to make the seismic data deep. The spectral width of wavelet in shallow strata is consistent, and its practicability is fully verified by applying this Q value modeling method to theoretical model and practical seismic data processing.
【学位授予单位】：东北石油大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P631.4

【相似文献】