Methods Of Pre-stack AVO Simultaneous Inversion In Depth Domain

The development and widespread application of pre-stack depth migration have produce huge volumes of pre-stack seismic data in depth domain.Conventional seismic inversion on depth-domain seismic data firstly converts seismic data volume from time domain to depth domain,then performs inversion of elastic parameters characterizing and evaluating reservoirs in time domain,lastly converts elastic parameters from time domain to depth domain to establish models for geological explanation.Pre-stack seismic data sets contain abundant information of amplitude and travel time,especially the information of S-wave.The development of AVO technology lays the foundation for pre-stack inversion.Therefore,it is of great significance to avoid extra conversions of depth-to-time and time-to-depth and extend the methods of pre-stack AVO inversion to apply to seismic data volume produced by pre-stack depth migration in depth domain.The waveform of depth-domain seismic wavelet is the result of comprehensive contribution of velocities within propagating depth range of seismic wavelet,and the wavelength is positively correlated with the velocity.Velocity-dependent property of seismic wavelet results in that the convolution between seismic wavelet and reflection coefficients does not meet the condition of linear time invariant system in depth domain.The method of convolution in constant-velocity depth domain meets that condition by constant-velocity substitution.The approach of weighted superposition of variant seismic wavelets fits the character of linear time variant in convolutional system by means of space-variant seismic wavelets.Seismograms synthesized by both methods all reveal effectively the influence of the velocity on seismic waveform.The three-term AVO equation is ill-conditioned to be solved for.Pre-stack AVO simultaneous inversion in depth domain makes problem stable by two-term equation of P-and S-wave impedances.The objective function for inversion is formulated based on norm optimization,and its first order gradient are calculated by means of methods of interpolation deviation and Richardson extrapolation.We parameterize model to reduce inversion problem scale,solve for the objective function by using non-linear bound constrained conjugate gradient algorithm.In the case of pre-stack AVO simultaneous inversion of impedances while P-wave velocity is not solved for in depth domain,the inversion result is optimal when seismic wavelet is dependent on low frequency trend model of P-wave velocity.Adopting the optimal strategy,we carry out the inversion of P-and S-wave impedances on synthetic gathers and actual pre-stack data in depth domain,yield satisfactorily applied results.The inversion on pre-stack depth migration data in depth domain is feasible and provides a new way for reservoir prediction and evaluation.The study on pre-stack AVO simultaneous inversion of impedances establishes a method foundation for the inversion of other elastic parameters and creates conditions for further research on improving the stability,accuracy and efficiency of the inversion in depth domain.