The conventional deconvolution method is based on traditional convolution model, which is not well suitable for the actual data. The method based on dynamic convolution model assumes that reflectivity is white noise, but the actual reflectivity series is not white noise, which impacts the effectiveness of deconvolution. Therefore,we studied the relationship between seismic signal, reflectivity and wavelet based on dynamic convolution model, and the result show that the logarithmic time-frequency spectrum of the seismic signal is the sum of that of the other two. Then, the timefrequency spectrum of wavelet was obtained by time-frequency spectrum and logarithmic time-frequency spectrum of synthetic seismogram by model testing. The result shows that if the reflectivity sequence is not white noise, better time-frequency spectrum of wavelet will be obtained by flattening the logarithmic time-frequency spectrum of seismic signal. Actual data application proved the above achievement.%基于动态褶积模型的反褶积方法假设反射系数序列是白噪的,而实际地震资料的反射系数序列通常是非白噪的,因而会影响反褶积效果.基于动态褶积模型研究了反射系数白噪和非白噪情况下地震道信息、反射系数和子波三者之间的关系,指出地震信号的对数时频谱是子波的对数时频谱和反射系数的对数时频谱之和.通过白噪和非白噪反射系数序列模型试验,研究了利用合成地震道的时频谱和对数时频谱求取子波时频谱的效果.试验结果表明:用对数时频谱求取的子波谱在频率方向精度更高；在反射系数序列非白噪的情况下,通过平滑合成地震道的对数时频谱获取子波时频谱,动态反褶积的效果更好.某油田实际地震资料处理结果验证了上述结论.
展开▼
