Forecasting Production in Undeveloped, Unconventional Plays Using Rate Transient Analysis and Experimental Design
MetadataShow full item record
A complete inventorying of resources, under the Petroleum Resources Management System (PRMS), requires credible low, best, and high case forecasts at all resources classification levels (Reserves, Contingent resources, and Prospective resources). Repeatable and accepted methodology for forecasting production and calculating EURs for each of these classification levels are not available: current methods to forecast production are inadequate for undeveloped resources, as they require production or pressure history, are overly simplified, or are time consuming and financially burdensome. Additionally, these methods do not quantify the level of uncertainty associated with a given forecast, which is needed to comply with the low, best, and high forecasts (often associated with a probability, P10, P50, P90), needed for inventorying under PRMS framework. RTA has been hailed as a happy medium between empirical and numerical simulation techniques to forecast production in unconventional, undeveloped plays in that it considers the completion and reservoir mechanics of the well and of the formation from which it produces (like numerical simulation techniques, unlike empirical techniques), and is straight-forward and user-friendly (like empirical techniques, unlike numerical simulation techniques). RTA also, does not require production history to generate a production forecast. However, there are currently few practical methods in industry which allow for the probabilistic forecasting of production using RTA. While we can consider a “best match” (or P50) forecast generated with RTA as a 2P (i.e., best or most likely) forecast, regulators and investors are far more interested in 1P (lower volume, high confidence) forecasts. The purpose of this work is to develop a workflow to generate a range of production forecasts using RTA techniques, from which probabilistic forecasts can be extracted. The methods involve first history-matching available production data, by varying critical reservoir and completion parameters to find the reservoir and completion parameter combinations which yield a best-fit (via least deviation calculated rate trends from observed rate trends). From this condensed number of best-fit history matches, appropriate probabilistic production forecasts for a certain well can be extrapolated. In this work, we show that incorporating Experimental Design (ED/DOE) techniques makes RTA a more practical production forecasting technique, reducing the number of history matches that need to be assessed, from which production forecasts can be generated. From this reduced set of best-fit history matches, appropriate probabilistic forecasts and EURs in accordance with PRMS and SEC standards, can be extracted.
Quist, Morgan Anne (2018). Forecasting Production in Undeveloped, Unconventional Plays Using Rate Transient Analysis and Experimental Design. Master's thesis, Texas A & M University. Available electronically from