基于Lucy-Richardson算法和广义S变换的Q值提取

品质因子Q能够定量衡量地震波的衰减。当地层厚度较薄时,常规的时频分析方法很难满足Q值估计的精度要求,其稳定性差。为此,将Lucy-Richardson算法与广义S变换相结合,推导出了新的谱比法品质因子估算公式。该方法引入高斯窗函数的尺度调节因子λ和p调节时窗函数的宽度和幅值,将广义S时频谱与高斯窗的Wigner-Ville分布代入Lucy-Richardson算法中,再利用对数谱比值与一个新定义的参数γ进行拟合,得到Q值。利用复杂模拟信号验证了Lucy-Richardson算法广义S变换的分辨率及聚集性,然后利用粘弹性层状模型进一步验证了该方法提取Q值的准确性。实际地震资料的应用结果表明,基于Lucy-Richardson算法广义S变换提取的Q值能准确区分含油气区域和围岩,与井资料吻合良好。

The Q factor can quantitatively measure the attenuation of seismic waves.However,challenges arise when the layer is thin and conventional time-frequency analysis is not able to meet the requirements of accurate Q estimation.In this study,the Lucy-Richardson algorithm and generalized S-transform are combined to derive a new formula for estimating the quality of the spectral ratio method.First,this method introduces the width and amplitude of the time window function of the parameters λ and p.Then,it substitutes the generalized S-time spectrum and the Wigner-Ville distribution of the Gaussian window into the Lucy-Richardson algorithm and uses the logarithmic spectral ratio to fit the parameters and obtain the Q value.Finally,the resolution and aggregation of the combined Lucy-Richardson algorithm and generalized S transform were verified by constructing complex analog signals.The accuracy of the Q value extraction using the proposed method was verified with a viscoelastic layered model.The results of application to actual seismic data show that the Q value obtained using the proposed method can accurately distinguish the oil-bearing area from surrounding rock,which is in good agreement with well data.

国家自然科学基金“基于频变信息的流体识别及流体可动性预测(41774142)”和国家自然科学基金联合基金“深层碳酸盐储层流体地震预测理论与方法(U1562111)”共同资助。