论文部分内容阅读
为了更好地确定井间岩石属性,特别是进行储层研究的需要,已经导致不断利用地震道属性作为岩石属性的标记。如果在井位处能够确定地震和岩石属性的有效的关系,可以大大地降低井间的不确定性。采用这种关系,随时可以利用地震资料指导井间的预测。但是,从地震属性资料预测岩石属性时存在三个关键性的问题:(1)地震属性的选择;(2)评价的准确性;(3)外围数值的确定。我们提出通过保持来自校验或者感知步骤中的资料并且利用这种隐含的资料检测结果,可以决定利用多少地震属性。这与地质统计中通常采用的横向验证方法相似。假定我们的控制资料代表了整个资料组分,这项技术也提供了有关我们预测值的不确定性的数值(即该区中钻探的任何新井落入我们已经钻探的井的范围)。这项技术还能够确定统计上存在差异的井(代表新的信息),或者存在误差的井。如果校验数据库在统计意义上与验证数据库不同,或者从统计意义上能够代表地震和测井资料,那么我们可以合理地假定验证数据库的准确结果意味着可以准确应用于地震资料。
In order to better determine crosswell rock properties, and in particular the need for reservoir studies, it has led to continuous use of the attributes of the seismic trace as a marker of rock properties. If we can determine the effective relationship between seismic and rock properties at the well site, we can greatly reduce the uncertainty between wells. Using this relationship, seismological data can be used at any time to guide the prediction of wells. However, there are three key problems in predicting rock properties from seismic attributes: (1) the choice of seismic attributes; (2) the accuracy of the evaluation; and (3) the determination of the external values. We propose that by preserving the data from the calibration or sensing steps and using this implicit data to detect the results, we can decide how much seismic attributes to utilize. This is similar to the horizontal verification method commonly used in geostatistics. Assuming that our control data represents the entire data component, this technique also provides a numerical value about the uncertainty of our predictions (ie, any new wells drilled in the area fall within the range of wells we have drilled). This technique also allows for the identification of wells that are statistically different (representing new information), or wells that have errors. If the verification database is statistically different from the verification database or can statistically represent seismic and well logging data, then we can reasonably assume that validating the accuracy of the database implies that it can be accurately applied to seismic data.