Preservation Studies In Deconvolution

Modern digital processing of seismic data requires more sophisticated, high resolution processing requirements gradually increased, deconvolution is the primary means of enhancing the seismic data resolution. But conventional deconvolution is often not preserving amplititude, too much emphasis on the high-resolution processing will result in preserving amplitude decreases. For deconvolution amplitude preservation quantitative evaluation, on the basis of the results of previous studies on the proposed impedance, waveform, wavelet and apparent wavelet which four kinds of evaluation methods of amplitude preservation; model (or logs) is given, based on the selected impulse deconvolution, predictive deconvolution, the maximum entropy deconvolution, harmonizer deconvolution, surface consistent deconvolution and robust surface consistent deconvolution which six methods to process the data; by changing the phase of the wavelet, the size of noise, formation thickness three sets of parameters to evaluate the above amplitude preservation deconvolution methods. Studies have shown that:If the phase of seismic wave is not minimum, strong energy spurious pulses will be come into being which will affect the amplitude preservation after predictive deconvolution and impulse deconvolution; maximum entropy deconvolution and harmonizer deconvolution for zero-phase wavelet, other sub-phase wave will reduce the amplitude preservation; when the thickness of target layer is less than 1/2 wavelength, the amplitude preservation of predictive deconvolution and impulse deconvolution come with thickness of stratigraphic will oscillate periodically; maximum entropy deconvolution and harmonizer deconvolution is not affected by the thickness of the stratum; noise for amplitude preservation effects of deconvolution is fatal, four deconvolution in the study will amplify random noise which will greatly reduce the amplitude preservation. The effect of amplititude preservation of Robust surface consistent deconvolution is better than surface consistent deconvolution. According to the problems of amplitude preservation of the prior existence deconvolution method, the paper proposed deconvolution method based on constraint layers through model and application of real data, and achieved good results.