Reducing Uncertainties in Production Forecasts by Constraining Geological Modeling to Dynamic Data

Hu, L.Y.; Ravalec, M.; Blanc, G.; Roggero, F.; Noetinger, B.; Haas, A.; Corre, B.

doi:10.2523/56703-ms

Cited by 3 publications

(4 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We can find many papers in the literaturedealing withthe effect of geological uncertainty in a reservoir reserve estimation and how to handle orreduce the geological uncertainty in production prediction such as Ballin et al (1993), Hu et al (1999), Massonnate (2000), Elfeki and Dekking (2007) and Chen et al (2012).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Robust Optimization of Unconventional Reservoirs under Uncertainties

Nguyen

Dang²,

Nghiem³

et al. 2016

SPE Europec Featured at 78th EAGE Conference and Exhibition

View full text Add to dashboard Cite

shows that the robust optimization increases all the NPV percentile values. This indicates the increased robustness of the optimal solution under uncertainties and thus an adoption of the robust optimization procedure can significantly reduce the project risk.This research is to present an innovative optimization approach which is fast and robust and can be used to efficiently capture the uncertainties of the key factors including geology, reservoir properties, and economic variation.

show abstract

Section: Introductionmentioning

confidence: 99%

“…The uncertainty in the geological model was generated by geostatistical stochastic simulation. Hu et al (1999) presented how to decrease uncertainties in production forecasts by constraining geological modeling to dynamic data. They used approaches of stochactic modelling and geostatistical simulation to reduce a model uncertainty.…”

Section: Introductionmentioning

confidence: 99%

Robust Optimization of Unconventional Reservoirs under Uncertainties

Nguyen

Dang²,

Nghiem³

et al. 2016

SPE Europec Featured at 78th EAGE Conference and Exhibition

View full text Add to dashboard Cite

show abstract

“…They did not use this approach in example problems, however, as in general it is not feasible to do a Schur decomposition of the Hessian for large-scale problems. Other re-parameterization methods that have been considered include the pilot point method RamaRao et al (1995);de Marsily et al (1984) which can introduce non-physical artifacts in the rock property fields (Oliver et al, 2008), and gradual deformation Hu et al (1999) which tends to yield reasonable geological models but may encounter difficulties in obtaining a good data match (Gao et al, 2007).…”

Section: Introductionmentioning

confidence: 99%

History Matching With Parametrization Based on the SVD of a Dimensionless Sensitivity Matrix

Tavakoli

Reynolds

2009

SPE Reservoir Simulation Symposium

View full text Add to dashboard Cite

In gradient based automatic history matching, calculation of the derivatives of all production data with respect to gridblock rock properties (sensitivities) and other model parameters is not feasible for large-scale problems. Thus, the Gauss-Newton method and Levenberg-Marquardt algorithm, which require calculation of all sensitivities to form the Hessian, are seldom viable. For such problems, the quasi-Newton and nonlinear conjugate gradient algorithms present reasonable alternatives as these two methods do not require explicit calculation of the complete sensitivity matrix or the Hessian. Another possibility, the one explored here, is to define a new parameterization to radically reduced the number of model parameters.We provide a theoretical argument which indicates that reparametrization based on the principal right singular vectors of the dimensionless sensitivity matrix provides an optimal basis for re-parametrization of the vector of model parameters. We present and apply to example problems two algorithms for using this parametrization. Like LBFGS, these algorithms avoid explicit computation of individual sensitivity coefficients. Explicit computation of the sensitivities is avoided by using a partial singular value decomposition based on a form of the Lanczos algorithm. At least for all synthetic problems that we have considered, the reliability, computational efficiency and robustness of the methods presented here are as good as those obtained with quasi-Newton methods.

show abstract

“…Analytical coefficients were then derived in the tracer case. [13][14][15][16][17][18] To apply the gradual deformation in our context, we will consider a set of different realizations of a permeability field. 10 This method has been further extended to compute sensitivity coefficients of pressures, saturations, and production data for two-phase flow with nonunit mobility ratios and varying boundary conditions using analytical solutions along streamlines.…”

Section: Introductionmentioning

confidence: 99%

History Matching Using a Streamline-Based Approach and Gradual Deformation

2004

Self Cite

View full text Add to dashboard Cite

Reservoir engineers often have to deal with history-matching problems. This is time-consuming because of the many numerical simulations that have to be run and also because of the size of the models. Optimization, coupled with gradient-based methods, enables engineers to find efficiently a reservoir representation that respects all static and dynamic data. Nevertheless, for multiphase flow or for compositional problems, only relatively small models can be handled with a finite-volume flow simulator. On the other hand, streamline-based simulation provides accurate and fast estimates to various flow problems. Simulations can be performed on larger models than with a regular fluid flow simulator. Recently, streamline-based simulation has been used for history-matching problems. First, analytical gradients of production data with respect to a geostatistical parameterization have been computed for the tracer-flow case. The saturations along streamlines were moved using an analytical solution of the transport equation. This approach then has been extended to waterflood problems with varying boundary conditions.In our method, we compute gradients of the production data, of the saturation, and of the pressure. We use a 1D numerical simulator to move the saturation along streamlines. This method is more general than when analytical calculations are used; it is more accurate for complex flow patterns. It also can be extended to take into account other parameters such as gravity. The major drawback is that it is more time-consuming; the cost for the computation of each gradient is slightly smaller than the cost of a simulation. Nevertheless, this cost can be dramatically decreased by computing multiple gradients at the same time on different processors using parallelism.These gradients will be used to perform history matching using a nonlinear optimization package in conjunction with a geostatistical method, the gradual deformation method, enables the combination of multiple realizations of a permeability field and the continuous variation from one to another.We will apply our new method to perform a history match on a large-scale 3D synthetic example of a waterflood with a complex injection history. The Streamline-Based Gradient ApproachIn this section, we will present our methodology to perform history matching. We will first introduce the Gradual Deformation Method. We will then describe the calculation of the streamlinebased gradients.

show abstract

Reducing Uncertainties in Production Forecasts by Constraining Geological Modeling to Dynamic Data

Cited by 3 publications

References 8 publications

Robust Optimization of Unconventional Reservoirs under Uncertainties

Robust Optimization of Unconventional Reservoirs under Uncertainties

History Matching With Parametrization Based on the SVD of a Dimensionless Sensitivity Matrix

History Matching Using a Streamline-Based Approach and Gradual Deformation

Contact Info

Product

Resources

About