三种国外常用蒸汽驱解析模型剖析及其适用性研究

发布时间：2018-03-18 00:37

本文选题：蒸汽驱　切入点：产量预测　出处：《东北石油大学》2015年硕士论文　论文类型：学位论文

【摘要】：稠油在世界油气资源中占有较大的比例。约占我国石油总资源量的28%,稠油开采在我国整个石油工业可持续发展中扮演着十分重要的角色。稠油油藏只进行蒸汽吞吐开采,采收率仅为10~20%,剩余死油区太多,剩余储量太大,必须接着进行蒸汽驱开采,将采收率提高到50%以上。当前,有一些数值模拟软件可以用于蒸汽驱的开发和预测,虽然预测效果不错,但是这些软件往往本身价格不菲,而且运行需要应用大型计算机,此外还会花费大量的运算时间。因此用分析预测模型替代数值模拟软件也不失为一个好的方法。然而,蒸汽驱的驱替过程很难进行分析,相同的预测模型对不同油藏类型,不同地质条件的油田的预测效果也差异很大,对蒸汽驱的机理分析以及对预测模型进行研究,是非常有必要的,会对稠油的生产实践产生重要的指导意义。本文综合对比了众多蒸汽驱模型,选取了国外现今最常用的三种蒸汽驱模型,深入分析研究了这三个模型对于不同油藏类型的适用性。本文分别对Jones模型,Farouq Ali模型以及Miller-Leung模型进行了深入研究,编制了计算程序,通过历史拟合发现,Jones蒸汽驱模型在对和Kern River油田情况类似的油田的预测中,效果更好,适用性更强。在对其他类型的油田产能预测上则需要对输入的一些经验参数做一些调整以适应历史拟合。在对采油速度的预测上,Jones模型在开采初期预测偏低,在之后的时间里又预测的偏高,在对产量的预测上,开始阶段预测的准确性不是很高,但随着时间的推移,预测结果将越来越精准。修正后的Farouq Ali模型不需要任何经验参数就可以拟合的很好。但是模型需要的Sors,Sorw,,Sst和Swir等参数却很不容易得到,需要简要的估算。将Farouq Ali模型同CMG数值模拟软件中理想模型进行对比,可以知道修正后的Farouq Ali模型对Kern River-A油田的拟合效果非常理想。Miller-Leung模型引入了面积波及系数这一概念,增大了模型的适用范围。在对采油速度的预测上,Miller-Leung模型的曲线更加的平缓,通过对比发现,在对产量的预测上,尤其是大型油田区块的应用上,Miller-Leung模型比Jones模型更加的精准和优秀,但对于个别油田,由于一些因素的影响Miller-Leung模型的预测时间会过短。通过对注汽温度,净总厚度比,初始含油饱和度等一系列敏感参数的研究,探明了这些参数对油田生产影响的大小,可以有针对性的对具体的油藏在生产时对注汽参数进行调节。
[Abstract]:Heavy oil occupies a large proportion of the world's oil and gas resources. It accounts for about 28 percent of the total oil resources in China. Heavy oil production plays a very important role in the sustainable development of the whole petroleum industry in China. Heavy oil reservoirs only perform steam huff and puff production. The oil recovery is only 10 ~ 20%, there are too many remaining dead oil areas, and the remaining reserves are too large. It is necessary to proceed with steam drive production to increase the recovery factor to more than 50%. At present, there are some numerical simulation software that can be used for steam drive development and prediction. Although the prediction works well, the software itself is often expensive and runs on mainframe computers. Therefore, it is a good method to replace the numerical simulation software with the analytical prediction model. However, it is difficult to analyze the displacement process of steam flooding, and the same prediction model can be used for different reservoir types. The prediction results of oil fields with different geological conditions are also very different, so it is very necessary to analyze the mechanism of steam flooding and to study the prediction model. This paper synthetically compares many steam drive models and selects the three most commonly used steam drive models in foreign countries. The applicability of these three models to different reservoir types is analyzed and studied in depth. In this paper, the Jones model, the Farouq Ali model and the Miller-Leung model are studied in depth, and the calculation program is compiled. By historical fitting, it is found that the model is more effective in predicting oil fields similar to those in Kern River oilfield. In order to predict the productivity of other types of oilfields, we need to adjust some of the input empirical parameters to fit the historical fitting. In the prediction of oil recovery rate, the Jones model is low in the early stage of production. In the later period of time, the prediction was on the high side. In the prediction of production, the accuracy of the initial stage was not very high, but with the passage of time, The prediction results will be more and more accurate. The modified Farouq Ali model can fit well without any empirical parameters. But the parameters such as Sorsn Sorwfen Sst and Swir are not easy to get. By comparing the Farouq Ali model with the ideal model in CMG numerical simulation software, we can know that the modified Farouq Ali model has a very ideal fitting effect on Kern River-A oil field. Miller-Leung model introduces the concept of area sweep coefficient. The curve of Miller-Leung model is more smooth in predicting the oil production rate. Through comparison, it is found that, in the prediction of production, Especially in the application of large oilfield blocks, the Miller-Leung model is more accurate and excellent than the Jones model, but for some oil fields, the prediction time of Miller-Leung model is too short because of some factors. The study of a series of sensitive parameters, such as initial oil saturation, has proved the influence of these parameters on oilfield production, which can be adjusted to the steam injection parameters of specific reservoirs in production.
【学位授予单位】：东北石油大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TE357.44

【相似文献】