论文部分内容阅读
计算机能力的提升和历史拟合方面的最新进展促进了对先前建立的储层模型的重新检验。为了节省工程师和CPU的时间,我们开发了4种独特的算法,来允许无需重新进行储层研究而重建现有模型。这些算法涉及的技术包括:优化、松弛、Wiener滤波或序贯重构。基本上,它们被用来确定一个随机函数和一系列随机数。给定一个随机函数,一族随机数将产生一个实现,这个实现和现有的储层模型十分接近。一旦随机数已知,现有的储层模型将被提交到一个历史拟合过程中,以此来改进数据拟合度或者考虑新收集到的数据。我们关注的是先前建立的相储层模型。虽然我们对模型模拟的方式一无所知,但是我们可以确定一系列随机数,再用多点统计模拟方法来建造一个和现有储层模型十分接近的实现。然后运行一种新的历史拟合程序来更新现有的储层模型,使其拟合两口新生产井的流量数据。
Recent advances in computer literacy and historical fittings have led to a re-examination of previously established reservoir models. To save engineers and CPU time, we developed four unique algorithms that allow the reconstruction of existing models without having to re-perform reservoir studies. Techniques covered by these algorithms include optimization, relaxation, Wiener filtering, or sequential reconstruction. Basically, they are used to determine a random function and a series of random numbers. Given a random function, a family of random numbers will produce an implementation that is very close to the existing reservoir model. Once the random number is known, the existing reservoir model will be submitted to a history fitting process to improve the data fit or consider the newly collected data. We are concerned with the previously established facies reservoir model. While we know nothing about the way the model is simulated, we can determine a series of random numbers and then use a multi-point statistical modeling approach to build an implementation that closely approximates the existing reservoir model. Then a new history fitting program was run to update the existing reservoir model to fit the flow data of two new production wells.