基于粒子滤波的碳酸盐岩缝洞型储层AVO反演研究

The commonly used pre-stack AVO inversion method is not effective in carbonates’ reservoir prediction and fluid identification. There are two reasons (1) The deeply buried carbonate reservoirs in China have strong heterogeneity and complex media, which result in difficulty in reservoir prediction and fluid identification. (2) The conventional AVO inversion’s accuracy is severely restrict by the low signal to noise ratio. Trying to address this problem, the main objective of this study is to modify pre-stack inversion method in order to make it suitable for carbonates’ reservoir prediction. Therefore, this project focuses on the following three aspects: (1) The random theoretical model characterizing the fractured-vuggy carbonates is established via investigating the deriving method of the characteristic function of the fractured-vuggy carbonates. (2) Investigate the influence of the parameters of fractured-vuggy carbonate reservoir on AVO, On that basis, the angle is adjusted according to the curvature. Due to the scattering of the heterogeneous body there is deviation exists between the incident angles of horizontal layering homogeneous medium from conventional hypothesis. (3) In order to improve the inversion accuracy, we mainly introduce MCMC particle filter to reduce the dependence on low-frequency model, and embedding the fractured-vuggy characteristic function to constrain the particles’ move. Once the project is completed, we will obtain a new carbonate reservoirs’ fluid identification method, and it will provide some ideas for carbonates’ reservoir prediction in China.

地震叠前AVO反演是目前传统的储层预测和流体识别方法，但在碳酸盐岩储层中的运用效果一般，针对我国碳酸盐岩储层埋藏深，储层介质复杂，非均质性强，储层预测和流体识别难度大以及AVO反演精度低的问题，本课题着重研究以下三方面内容：（1）优选碳酸盐岩缝洞体描述的特征函数，建立通过随机模拟来表征碳酸岩缝洞体的理论模型；（2）研究碳酸盐岩缝洞型储层的缝洞参数对AVO反演的影响，探讨非均质体的散射，以及射线入射角与传统假设水平层状均匀介质入射角的偏差问题，并根据曲率对角度进行校正；（3）在提高反演精度方面，由于地震信号信噪比低，我们通过引入MCMC粒子滤波来降低对低频模型的依赖，并嵌入缝洞特征函数进行约束，形成不同于传统的AVO反演方法。通过对以上三方面内容的详细研究和深入探讨，推导并总结碳酸盐岩流体识别的新方法，从而有望为我国碳酸盐岩储层预测提供新思路.

本研究针对碳酸盐岩储层中AVO应用效果不佳，储层预测精度低的难题，主要设置了储层模拟、碳酸盐岩反演角度校正以及基于粒子滤波的叠前AVO反演方法研究三方面的研究内容。通过三年来项目的实施，主要获得了以下四方面的进展： 一、通过对碳酸盐岩储层进行实际建模，通过数值模拟研究了其波场特征；二、在对波场特征有了一定了解的基础上，提出了自适应走时校正的 Marchenko 成像方法，该方法在初始速度模型不正确的情况下，仍然能够获得较高成像质量的地震剖面；三、考虑非匀质体散射，提出了基于梯度结构张量的角度校正AVO反演方法；四、AVO反演由于近似方程存在诸多假设，与实际地下地质情况存在误差，为了进一步提高反演的精度，我们在正演时采用了精确Zoeppritz方程，在贝叶斯框架下，利用鲸群算法来改进粒子滤波贫化问题，新方法在理论和实际应用中均比传统线性AVO方法具有更高的反演精度。在全球油气探明程度不断提高和勘探难度日益增大,碳酸盐岩油气藏逐渐成为重要勘探领域的背景下，本项目的顺利实施对提高碳酸盐岩储层预测精度具有重要意义。