多点地质统计学建模参数敏感性分析

发布时间：2018-04-21 01:01

本文选题：地质建模 + 多点模拟算法　；参考：《西安石油大学学报(自然科学版)》2017年01期

【摘要】：与基于像元的两点模拟和基于目标的模拟相比,多点地质统计学能较好地忠实于硬数据(井数据)和再现复杂的地质体形态。在介绍多点模拟算法(Snesim)实现流程的基础上,对Snesim算法中重要的输入参数进行了敏感性分析,结果表明:目标比率越接近训练图像的边缘相概率,模拟效果越好;目标比率一定时,提高伺服参数可使模拟相的比率更接近目标体,但以损失相结构信息为代价;搜索邻域的设计、网格级数的选择取决于训练图像的大小以及需重现的结构信息;此外,在一个较小的数据事件重复数下,随着最大条件数据的增加,其结构信息的再现效果越好,所需机时则呈线性增加。可见,参数设置对多点地质建模中模拟效果的好坏至关重要。
[Abstract]:Compared with the two point simulation based on pixel and target based simulation, multi point geostatistics can be faithful to hard data (well data) and complex geological form. Based on the introduction of the implementation process of the multi point simulation algorithm (Snesim), the sensitivity analysis of the important input parameters in the Snesim algorithm is analyzed. The results show that: The better the probability of the edge phase of the trained image, the better the simulation effect. When the target ratio is fixed, the ratio of the servo parameters can make the ratio of the simulated phase closer to the target body, but at the expense of the structure information of the loss phase; the selection of the grid series depends on the size of the training image and the structural information that needs to be reproduced; moreover, In a small number of data events, with the increase of the maximum condition data, the better the reconstruction effect of the structure information and the linear increase in the time required. It is very important to set the parameters to the simulation effect in the multi point geological modeling.

【作者单位】：中国石油大学(华东)地球科学与技术学院;
【基金】：国家科技重大专项“复杂油气藏精细表征及剩余油分布预测”(编号:2011ZX05009-003)
【分类号】：P628.2

【相似文献】