东海西湖N区块致密砂岩气藏甜点预测研究

“Sweet Spot” prediction of a tight sandstone gas reservoir in the N Block in Xihu Sag,China

东海西湖N区块HG组H3b段为致密砂岩储层,具有砂体厚度大、低孔、低渗及横向差异变化大等特点,致密砂岩与泥岩、优质储层分布与含气性变化比较快,寻找物性较好、含气丰度较高的砂岩“甜点”是该区天然气高效开发的关键所在。在砂体、储层、含气性岩石物理特征综合研究基础上,探索了一套基于敏感弹性参数优选与合成的甜点预测技术,采用砂体优质储层含气优质储层逐级递进预测流程,达到准确预测甜点的目的。利用叠前资料反演纵波阻抗、纵横波速度比数据体,采用坐标旋转技术合成具有岩性区分识别能力的砂泥岩SynVpVs敏感属性参数,该参数可用来识别砂泥岩;利用优质储层的低纵波阻抗值特征划分识别优质储层;利用E/λ(杨氏模量与拉梅参数比值)含气性敏感属性参数预测优质储层含气性。实际应用结果表明,该方法在一定程度上降低了多解性,提高甜点预测的准确性和精度。

The H3b section of the HG group in the N block of the Xihu Sag,East China Sea,is a tight sandstone reservoir,in which the sand body exhibits a large thickness,low porosity,low permeability,and high variation of lateral differences.For tight sandstone and mudstone,the distribution of high-quality reservoirs and the gas-bearing properties are highly variable.The search for “sweet spots” in sandstone,having good physical properties and high abundance of gas,is critical for an efficient development of natural gas exploitation in the area.Based on a comprehensive analysis of the sandstone,reservoir,and gas-bearing petrophysical characteristics,a set of techniques,based on a sensitive elastic parameter optimization and a synthetic sweet spot prediction,was explored.This technology utilizes a progressive prediction process,searching for a sand body first,then for a high-quality reservoir,and finally for a high-quality gas-bearing reservoir,so as to accurately predict the location of the “sweet spot”.Based on P-wave impedance and v P/ v S volume obtained from pre-stack inversion,a sensitive attribute SynV pV s which has the ability to distinguish lithology was synthesized using the coordinate rotation technique to identify sand and mudstone.P-wave impedance was used to identify the high-quality reservoirs.The ratio of Young s modulus to the Lame s parameter (E/λ) is a gas-sensitive parameter that can be used to predict gas-bearing properties of high-quality reservoirs.Practical applications of the proposed methods can reduce the need for multiple solutions (model indeterminateness) and improve the accuracy of “sweet spot” prediction.