Comparison of Empirical and Analytical Methods for Production Forecasting in Unconventional Reservoirs: Lessons from North America

Abstract

The prediction of Estimated Ultimate Recovery (EUR) for a well or group of wells in a development project is critical to accurate reserves estimation. A number of techniques, many of which can be used deterministically or probabilistically, are employed in EUR prediction in mature and maturing unconventional gas and oil plays in North America. These include the use of geological and production data from analogous reservoirs, the use of volumetric methods and recovery factors, analytical models, numerical reservoir simulation and production decline curve analysis (DCA). Decline curve analysis is arguably the most commonly used method for forecasting reserves in unconventional reservoirs. This paper discusses its basic theory and application, together with the potential pitfalls of using simple empirical production forecasting methods in complex reservoirs. We analyse production data from several US unconventional oil and gas plays and carry out production forecasting using the traditional Arps' methods as a basis for comparison, and newer empirical solutions including the Power Law, Stretched Exponential Decline Model, Duong (and variations thereof). The range of production forecasts provided by these methods is examined, together with methodologies for developing statistically valid type wells in unconventional plays, and how best to determine valid input parameters for the various empirical solutions. The effect of the variable length of production history available in the various plays, and how it impacts the accuracy of the forecasts is also examined. The results of the analyses are compared with analytical models developed for each play to determine the suitability of each decline curve analysis method: in which plays and under which circumstances they can be applied, and suggest reasonable input parameters and data requirements for each method. Finally, the potential future use of the methods in emerging plays outside of North America is presented.

Integrated Hierarchy–Correlation Model for Evaluating Water-Driven Oil Reservoirs

“…The different timescale data were selected to train the machine learning models. The numerical results confirmed the superiority of the SGB model in predicting the oil recovery factor. ,− ,− …”

Impact of gas desorption on production behavior of shale gas

“…Also, he found that most shale models have "a" values ranging from 0 to 3 and m values ranging from 0.9 to 1.3. His model gives the reasonable estimation of cumulative production compared to the power law and Arp's model ( Dutta et al, 2014). Nelson et al (2012) investigated the application of various decline curves to production data form hydraulically fractured horizontal wells completed in a shale foramtion.…”