高精度地震属性储层预测技术研究
|
摘要
地震属性技术是储层预测的重要手段,但地震储层预测存在多解性,可靠性程度低问题,预测出的最终图件难以进行合理的地质解释,文章基于地震属性与地质属性随时空变化的关系,根据地震属性与储层属性相关程度,以及地震属性对储层参数敏感程度来进行地震属性有效性分析;然后将搜索算法与神经网络相结合来实现地震属性优化,用优化出的地震属性再进行多元储层预测。实际工区的砂岩厚度预测结果表明,上述方法可明显提高储层预测的精度。
Seismic attribute technique is an important means of reservoir prediction. However, the seismic reservoir prediction is of multi-solvability and low reliability and the final maps predicted are difficult to make a reasonable geologic interpretation. It is pointed out in the paper that the effectiveness analysis of seismic attribute can be done according to the time-space change relation between seismic attribute and geologic attribute, the correlativity of seismic attribute with reservoir attribute and the sensibility of seismic attribute to reservoir parameter; then the seismic attribute optimization may be realized by combining the searching algorithm with the neural network; and then the multivariate reservoir prediction can be finished finally by use of the optimized seismic attribute. The prediction result of sandstone thickness in work area indicated that the accuracy of reservoir prediction might be obviously raised by the method mentioned above.
Seismic attribute technique is an important means of reservoir prediction. However, the seismic reservoir prediction is of multi-solvability and low reliability and the final maps predicted are difficult to make a reasonable geologic interpretation. It is pointed out in the paper that the effectiveness analysis of seismic attribute can be done according to the time-space change relation between seismic attribute and geologic attribute, the correlativity of seismic attribute with reservoir attribute and the sensibility of seismic attribute to reservoir parameter; then the seismic attribute optimization may be realized by combining the searching algorithm with the neural network; and then the multivariate reservoir prediction can be finished finally by use of the optimized seismic attribute. The prediction result of sandstone thickness in work area indicated that the accuracy of reservoir prediction might be obviously raised by the method mentioned above.
引文
1 Leiphart D J, Hart B S. Case history-comparison of linear regression and a probabilistic neural network to predict porosity from 3-D seismic attributes in lower brushy canyon channeled sandstones,southeast New Mexico. GEOPHYSICS,2001;66(5):1349-1358
2周守信,徐严波,李士伦等．致密泥质砂岩储层的物性预 测方法及应用．天然气工业,2004;24(1):40-43
3李学义．利用地震资料作储层预测需注意的问题．天然气 工业,2004;24(9):44-47
4陈遵德．地震数据油气预测中的属性优化方法．石油物 探,1998;37(4):32-47
5 Dorrington K P et at. Genetic-algorithm/neural-network approach to seismic attribute selection for well-log prediction. Geophysics,2004;69(1):212-221
6顾汉明．地震特征空间到储层特征空间的非线性映射．石 油地球物理勘探,1997;32(1):111-116
2周守信,徐严波,李士伦等．致密泥质砂岩储层的物性预 测方法及应用．天然气工业,2004;24(1):40-43
3李学义．利用地震资料作储层预测需注意的问题．天然气 工业,2004;24(9):44-47
4陈遵德．地震数据油气预测中的属性优化方法．石油物 探,1998;37(4):32-47
5 Dorrington K P et at. Genetic-algorithm/neural-network approach to seismic attribute selection for well-log prediction. Geophysics,2004;69(1):212-221
6顾汉明．地震特征空间到储层特征空间的非线性映射．石 油地球物理勘探,1997;32(1):111-116
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心