基于最大似然属性的断层识别方法及应用

断层识别是断块型油气田勘探开发的重要研究内容,尤其是在复杂断块油气田的勘探开发中,准确合理的断层识别是落实油气田构造和确定注采井网的关键因素.方差体、相干体、曲率属性等常规方法在断层识别中发挥着重要作用,但在复杂断裂发育区地震资料品质较差,常规方法分辨率较低从而无法准确识别断层组合关系.基于相似系数改进的似然属性在已知断层倾向和倾角时可以精确表征断层,但由于断层的倾向和倾角是未知的,因此可以采用断层倾向和倾角扫描的方法计算最大似然属性来表征断层.本文对比分析了相似系数和最大似然属性的原理;并将最大似然属性应用于模型正演数据和实际地震数据进行断层识别分析,结果表明,最大似然属性在剖面上更符合断层展布特征,在平面上断层组合关系更加清晰,在断层识别上具有较好的应用效果.

Fault identification method based on the maximum likelihood attribute and its application

Fault identification plays an important role in the evaluation and development of the fault-block oil and gas fields.Especially in complex fault-block oil and gas fields,accurate fault identification is the key factor to implement structure features and determine the injection-production relationship.General methods such as variance,coherence and curvature attribute have mature application in fault identification.However,the imaging of seismic is poor in complex fault development area.Meanwhile,the geological bodies such as multi-stage channel sand body also lead to the break of seismic event axis which is similar to the main characteristic of fault identification.In these cases,these methods show some weakness such as unclear fault combination and disturbed by geological bodies.The likelihood attribute improved based on the semblance can accurate describe the fault occurrence with the maximum value when the fault tendency and dip angle are accurate,but the difficulty is that we don't know the fault tendency and dip angle in advance.An alternative method is calculate the maximum likelihood attribute based on the tendency and dip angle scanning to search the maximum value from the likelihood attribute.In this paper,we first analyze the algorithm of the semblance and the maximum likelihood attribute.Then we apply the maximum likelihood attribute on the forward modeling data and actual seismic data for fault identification.The results show that the maximum likelihood attribute describe fault plan more accurate in profile and clearer fault combination in plane.It shows that the maximum likelihood attribute has good application for fault identification.