Understanding Reservoir Performance and Uncertainty Using a Multiple History Matching Process

Abstract

This paper describes the history matching and predictive case studies of twodeepwater Gulf of Mexico (GOM) fields using an advanced Bayes linear estimationtool.Advantages of the tool include significant acceleration of thehistory matching process, identification and quality measurements of multiplehistory matches, quantification of reservoir uncertainty, and an improvedunderstanding of reservoir performance.Additionally, a statisticalestimator of predictive simulation results is created to generate statisticallyvalid confidence intervals around performance predictions.This paperdescribes a practical workflow incorporating this tool to rapidly evaluatedeepwater producing gas fields, and illustrates its use to determine remainingfield potential and future development requirements of two fields, the Harrierand the Raptor Fields in the Pioneer Natural Resources-operated FalconCorridor. Introduction The deepwater GOM can contain fields with very prolific wells that can behighly profitable for an Operator.The loss of even one of these wells canadversely impact both short and long-term field production forecasts, thus cashflow and profitability.These impacts are especially significant whenthere are few, very high rate wells that contribute to the total fieldproduction.When such a well fails, it is crucial to understand the causesin order to determine how and if the situation can be remedied, the costnecessary to do so, and the risks involved.The goals are then tounderstand and reduce risks, to minimize cycle time and capital exposure, andto maximize profitability. If a well's failure to meet forecast expectations is attributed to reservoirperformance, a number of tools ranging from the very simple to the very complexcan be used to evaluate reservoir performance.The choice as to whichtools to use is dependent upon the amount and quality of data available, thecomplexity of the problem, the time available in which to make a decision, andthe magnitude of the capital required to execute the decision.Oftenhistory matching with 3-dimensional (3D) reservoir simulation is the tool ofchoice used to evaluate and explain production performance.However, thehistory-matching process can be very frustrating and time-consuming, even forfields that appear relatively simple in nature, because of the reservoirprocesses involved and the non-unique nature of the solution.[1]Consequently, much time and many resources can be spent in attempting toachieve even one history match.Frequently multiple solutions can be foundthat can satisfy history-match criteria but which yield divergent predictionoutcomes. Because of the high production rates in both the Harrier and Raptor Fields, rapid analysis and integration of production data were necessary to providequick answers to reservoir analysis and reservoir management questions, and toaddress well-intervention and deepwater rig availabilitydecisions.Pioneer selected 3D simulation and the implementation of anadvanced linear Bayesian tool to expedite the history matching and uncertaintyanalyses process.[2]3D static geologic models for both fields, built andupscaled for dynamic flow simulation prior to production start-up, had beenused for predictive simulations.Good quality pressure information frompermanent downhole gauges and daily gas production data were available for thecalibration of these models.Although there is a global workflow processthat encompasses the ‘seismic to simulation’ process, this paper focusesprimarily upon the dynamic history matching and predictive portion of theoverall process.

Bridging the Gap Between Deterministic and Probabilistic Uncertainty Quantification Using Advanced Proxy Based Methods

“…They show a remarkable efficiency compared to other approaches. (Bustamante, Keller, & Monson, 2005) (Jutila & Goodwin, 2006) (Elrafie, White, & Awami, 2008) (Junker, Dose, Plas, & Little, 2006) (Gruenwalder, Poellitzer, & Clemens, 2007) (Elrafie, White, & Awami, 2008) (Shenawi, White, Elrafie, & El-Kilany, 2007).…”

Work Flow for Reservoir Study and Challenges

“…Thus, history matching of geologically complex reservoirs are challenging aspects for efficient reservoir management which is normally the most time-consuming phase of a simulation study that every dynamic reservoir engineer will want to avoid if possible. The foremost reason is the high level of uncertainty that exists in the reservoir models because of the limited, sparse, and multiscaled reservoir data available (Bustamante, 2005). It helps to identify the weaknesses in the available data, keeps the model up to data and consistent, improves the reservoir description and forms the basis for future performance predictions.…”

Modern Approach to Estimation of Uncertainty of Predictions With Dynamic Reservoir Simulation - A Case Study of a German Rotliegend Gasfield

“…76% of reserves. 'White' wells (1,2,4,8,10) show contributions of 23%, whilst 'grey' wells (3,7,9) will contribute only around 1%.…”