Exploring EnKF stability under different type reservoir model

Darwis, Sutawanir; Gunawan, Agus Yodi; Mutaqin, Aceng Komarudin; Fitriyati, Nina
May 2012
AIP Conference Proceedings;5/22/2012, Vol. 1450 Issue 1, p132
Academic Journal
Reservoir properties estimation aims to maximize the recovery, and least squares history matching is one of strategies to achieve this target. A model is matched on historical data until it reproduced the production history. The model can be used to simulate future reservoir production. The Ensemble Kalman Filter (EnKF) is a method designed for non linear history matching. The success of EnKF affected by non linearity of reservoir model. If the reservoir model is highly non linear the EnKF might be unstable. This paper aims to explore the stability of EnKF under different type reservoir models. Three type reservoir models are considered: (1) line source (2) bounded constant rate no flow (3) three grid reservoir. The permeability is estimated based on the matching of observed pressure with predicted pressure (obtained from reservoir model). A simulation experiment is set up to show how EnKF iteration may be used to estimate the permeability. The simulation is set up for an experiment time T, time step Δt, and number of observations m. The prior for permeability is drawn from normal distribution. The stability of the results is presented as plots of pressure history match and permeability updating. The three grid reservoir model shows instability of EnKF method, and an iterative EnKF was proposed to improve the stability of EnKF. The proposed method is able to history match the reservoir models, and exploring the stability of a more general reservoir model is needed for the success of the method.


Related Articles

  • Kalman Filtering as an Alternative to Ordinary Least Squares--Some Theoretical Considerations and Empirical Results. Watson, P. K. // Empirical Economics;1983, Vol. 8 Issue 2, p71 

    The purpose of this paper is to highlight the superiority of the Kalman filter over Ordinary Least Squares for estimating the unknown coefficients of the classical linear regression model. Both methods are analyzed with respect to their optimality properties and their usefulness in dealing with...

  • Autoregressive conditional tail behavior and results on Government bond yield spreads. Wagner, Niklas // International Review of Financial Analysis;2005, Vol. 14 Issue 2, p247 

    Previous evidence in empirical finance indicates the potential usefulness of modeling time variation particularly in the tails of speculative return distributions. Based on results from extreme value theory, the present paper proposes a fixed changepoint Pareto-type autoregressive conditional...

  • Comparison of Ensemble Kalman Filter-Based Forecasts to Traditional Ensemble and Deterministic Forecasts for a Case Study of Banded Snow. Suarez, Astrid; Reeves, Heather Dawn; Wheatley, Dustan; Coniglio, Michael // Weather & Forecasting;Feb2012, Vol. 27 Issue 1, p85 

    The ensemble Kalman filter (EnKF) technique is compared to other modeling approaches for a case study of banded snow. The forecasts include a 12- and 3-km grid-spaced deterministic forecast (D12 and D3), a 12-km 30-member ensemble (E12), and a 12-km 30-member ensemble with EnKF-based...

  • Handling Nonlinearity in an Ensemble Kalman Filter: Experiments with the Three-Variable Lorenz Model. Yang, Shu-Chih; Kalnay, Eugenia; Hunt, Brian // Monthly Weather Review;Aug2012, Vol. 140 Issue 8, p2628 

    An ensemble Kalman filter (EnKF) is optimal only for linear models because it assumes Gaussian distributions. A new type of outer loop, different from the one used in 3D and 4D variational data assimilation (Var), is proposed for EnKF to improve its ability to handle nonlinear dynamics,...

  • General formulation of HDMR component functions with independent and correlated variables. Li, Genyuan; Rabitz, Herschel // Journal of Mathematical Chemistry;Jan2012, Vol. 50 Issue 1, p99 

    The High Dimensional Model Representation (HDMR) technique decomposes an n-variate function f ( x) into a finite hierarchical expansion of component functions in terms of the input variables x = ( x, x, . . . , x). The uniqueness of the HDMR component functions is crucial for performing global...

  • Time-Dependent Structural Equations Modeling: A Methodology for Analyzing the Dynamic Attitude-Behavior Relationship. Lyon, Patricia K. // Transportation Science;Nov84, Vol. 18 Issue 4, p395 

    The attitude-behavior relationship is expressed as a set of time- dependent structural equations whose endogenous variables are measures of attitudes (continuous) and behavior (binary). The dynamic aspect of the structure is incorporated in three ways: the equations have a lagged endogenous...

  • A robust stage-discharge rating curve model based on critical flow from a reservoir. Petersen-Øverleir, Asgeir // Nordic Hydrology;2006, Vol. 37 Issue 3, p217 

    Hydrometric gauging stations are often sited in reservoirs such as lakes and river pools since they possess favourable features for streamflow determination. By applying the equations which govern critical flow from a reservoir along with a spline representation of the known geometric...

  • A Instantaneous Availability Modelling Method for Repairable System. Dezhi Kong; Xiaobing Li // Applied Mechanics & Materials;2014, Vol. 724, p334 

    A Instantaneous availability modelling method for repair system was introduced. The empirical modeling method was used to the reliability function, and the life distribution selection method was proposed. Then, under the assumption that the maintainability was decreasing with the calendar time,...

  • Constrained State Estimation for Nonlinear Systems with Unknown Input. Luo, Zhen; Fang, Huajing; Luo, Yuanhao // Circuits, Systems & Signal Processing;Oct2013, Vol. 32 Issue 5, p2199 

    This paper extends the problem of state estimation for linear discrete-time systems with unknown input to the nonlinear systems. Based on physical consideration, the constraints of state are also considered. And the constraints which can improve the quality of estimation are imposed on...


Read the Article


Sorry, but this item is not currently available from your library.

Try another library?
Sign out of this library

Other Topics