期刊
JOURNAL OF PETROLEUM SCIENCE AND ENGINEERING
卷 162, 期 -, 页码 617-632出版社
ELSEVIER
DOI: 10.1016/j.petrol.2017.10.079
关键词
Production forecasting; EUR estimates; Modified Hyperbolic Decline Model; Duong model; Logistic Growth Analysis Model; Power Law Exponential
资金
- Texas A&M Engineering Experiment Station
Various decline curve analysis (DCA) methods can be applied to forecast the production performance of hydrocarbon wells, including horizontal wells stimulated with hydraulic fracturing. Yet, which method is more preferable remains in doubt. The objective of this study is to evaluate various DCA methods by history matching and hindcasting both synthetic and field production data in order to assess for each method the reliability of production forecasts and the estimated ultimate recovery (EUR). Five DCA methods have been evaluated because of their computational simplicity and broad application. These methods are the Modified Hyperbolic Decline model (MHD), Duong model, Logistic Growth Analysis Model (LGM), and the Power Law Exponential (PLE, with D-infinity = 0 and D-infinity = 8) Decline models. Each method was evaluated using three data sets: synthetic shale oil production rates based on a CMG reservoir model using Eagle Ford reservoir characteristics, data from producing Eagle Ford wells and from Austin Chalk wells. The hindcast method was applied to the Eagle Ford synthetic wells to establish the EUR deviation between the synthetic reservoir model production forecast and the various DCA methods. The weighted residuals of history matched production rates for Eagle Ford indicate that the MHD and Duong models give the lowest matching errors (as low as 1.32%) and the highest EUR estimations (as high as 3,447,609 stb). The PLE model (with D-infinity not equal 0) gives the most conservative EUR estimation, and also the least reliable history matching method (on our Eagle Ford data sets), which therefore is the least preferable DCA method. For the Austin Chalk wells in our study, all methods generate fairly similar EUR predictions with similar matching errors (from 13% to 15%) so that the DCA method preference becomes indifferent. This study also touches upon the relationship between DCA results when well interference occurs due to various hydraulic fracture hits. Shale wells that primarily contain hydraulic fractures and few natural fractures (Eagle Ford) give the most reliable forecasts using the MHD and Duong DCA methods. However, wells in the Austin Chalk, a naturally fractured reservoir, show little difference between the forecast accuracy from the various DCA methods.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据