基于多射线联合反演的速度无关叠前地震数据Q值估计

描述地震波衰减特征的品质因子Q对地震数据处理和油藏描述非常重要,在地震勘探领域,Q值一般通过垂直地震剖面(VSP)数据或地面地震数据得到.由于叠前地面地震数据具有复杂的射线路径且存在噪声、调谐干涉效应等影响,从叠前地震数据中准确估计Q值相对困难.本文以地震波射线传播为基础,根据同相轴局部斜率和射线参数的映射关系,将多射线波形频谱同时带入谱比法联合反演估计Q值,提出了基于多射线联合反演的速度无关叠前Q值估计方法.该方法通过局部斜率属性避开了速度对Q值估计的影响,局部斜率携带地震波传播的速度信息,具有相同局部斜率的地震反射波具有相同的传播射线参数.同相轴局部斜率是地震数据域的属性,而速度是模型域的参数,在估计Q值中采用数据域的属性参数可以直接应用于数据的联合反演,而不需要通过速度对其做进一步的转化,从而提高了Q值估计的精度.同时,本方法采用预测映射(predictive mapping)技术将非零炮检距反射信息映射到零炮检距处,从而获得零偏移距走时对应的Q值.模拟和实际算例验证了本文方法的有效性.

Velocity-independent prestack seismic Q estimation based on multi-ray joint inversion

The quality factor Q, which describes the attenuation characteristics of seismic waves, is very important for seismic data processing and reservoir characterization. In the field of seismic exploration, Q value is usually obtained from vertical seismic profiling (VSP) data or surface seismic data. Because of the complex ray path, the influence of noise and tuning interference, it is relatively difficult to estimate Q value accurately from prestack seismic data. Based on the ray propagation of seismic wave and the mapping relationship between local slope of event and ray parameters, the spectrum of multi-ray waveform is simultaneously brought into the joint inversion of spectral ratio method to estimate Q value. A velocity-independent prestack Q value estimation method based on joint inversion of multi-rays is proposed. The method avoids the influence of velocity on Q value estimation by using local slope attribute. The local slope carries the velocity information of seismic wave propagation, and the seismic reflection wave with the same local slope has the same propagation ray parameter. Local slope of event is the attribute of seismic data domain, and velocity is the parameter of model domain. In estimating Q value, the attribute parameters of data domain can be directly applied to joint inversion of data without further transformation of velocity, thus improving the accuracy of Q value estimation. In addition, this method maps the reflection information of non-zero offset to zero offset using predictive mapping technology, so as to obtain the corresponding Q value of zero offset traveltime. The effectiveness of the proposed method is verified by synthetic and field examples.