【作者】 孙山；

【导师】 宋维琪；

【作者基本信息】 中国石油大学 ， 地球探测与信息技术， 2007， 硕士

【摘要】 本文针对滩坝砂岩薄层的地质和地震特征,研究了多方向分形分维方法和复数子波匹配追踪方法识别薄层砂体空间展布。同时,也研究了瞬时平均频率峰值预测滩坝砂岩薄层厚度的方法。在研究分形多种算法的基础上,针对滩坝砂的特征,研究了简化的多方向分形分维方法,经过模型论证,得到了滩坝砂薄层的横向上展布。针对多方向分形分维方法纵向上识别薄层砂岩能力较弱,研究了经过改进的复数子波匹配追踪方法。采用复数子波匹配追踪方法,经过模型论证,证明该方法能够从纵向识别滩坝砂薄层的地震信息。简化的多方向分维与改进的复数子波匹配追踪方法加快了计算速度,节省了计算时间。二者对于薄层砂体的识别相互补充,较好的识别出滩坝砂岩的空间展布。在复数子波匹配识别滩坝砂薄层的基础上,运用瞬时平均频率峰值理论预测薄层砂体的厚度,得到了较好的应用效果。更多还原

【Abstract】 In this paper, for the geologic feature and seismic character of thinbeach-bar sand, I have studied the multi-direction fractal dimensionmethod and the matching pursuit complex wavelet method, aiming atrecognizing thespace distribution of the thin sandbody.At the same time,I have researched the instantaneous average frequency peak method topredict the thickness of thin beach-bar sandbody. For the feature ofbeach-bar sand, I have studied the simplified multi-direction fractaldimension method also, based on the research on various fractalalgorithms, and have obtained the horizontal distribution of thin beach-barsand by the proof of the model. Because of the poor ability of themulti-direction fractal dimension method in recognizing thin sand atlongitudinal direction, advanced matching pursuit complex waveletmethod is studied. After proof-texted by the model, this method can beproved that it does much better in recognizing seismic information of thinbeach-bar sand at longitudinal direction. Simplified multi-direction fractaldimension method and advanced matching pursuit complex waveletmethod compute much faster than before. Each of the methods can be a supplement to another, and do better in recognizing the space distributionof the beach-bar sand. Based on recognizing thin beach-bar sand withmatching pursuit complex wavelet method, the instantaneous averagefrequency peak method is applied to predict the thickness of thinbeach-bar sand, and preferable effect is obtained.更多还原