共查询到20条相似文献，搜索用时 46 毫秒
1.
Combining sensitivities and prior information for covariance localization in the ensemble Kalman filter for petroleum reservoir applications 总被引：1，自引：0，他引：1
Sampling errors can severely degrade the reliability of estimates of conditional means and uncertainty quantification obtained
by the application of the ensemble Kalman filter (EnKF) for data assimilation. A standard recommendation for reducing the
spurious correlations and loss of variance due to sampling errors is to use covariance localization. In distance-based localization,
the prior (forecast) covariance matrix at each data assimilation step is replaced with the Schur product of a correlation
matrix with compact support and the forecast covariance matrix. The most important decision to be made in this localization
procedure is the choice of the critical length(s) used to generate this correlation matrix. Here, we give a simple argument
that the appropriate choice of critical length(s) should be based both on the underlying principal correlation length(s) of
the geological model and the range of the sensitivity matrices. Based on this result, we implement a procedure for covariance
localization and demonstrate with a set of distinctive reservoir history-matching examples that this procedure yields improved
results over the standard EnKF implementation and over covariance localization with other choices of critical length. 相似文献
2.
Wiktoria Lawniczak Remus Hanea Arnold Heemink Dennis McLaughlin 《Computational Geosciences》2009,13(2):245-254
Reservoir management requires periodic updates of the simulation models using the production data available over time. Traditionally,
validation of reservoir models with production data is done using a history matching process. Uncertainties in the data, as
well as in the model, lead to a nonunique history matching inverse problem. It has been shown that the ensemble Kalman filter
(EnKF) is an adequate method for predicting the dynamics of the reservoir. The EnKF is a sequential Monte-Carlo approach that
uses an ensemble of reservoir models. For realistic, large-scale applications, the ensemble size needs to be kept small due
to computational inefficiency. Consequently, the error space is not well covered (poor cross-correlation matrix approximations)
and the updated parameter field becomes scattered and loses important geological features (for example, the contact between
high- and low-permeability values). The prior geological knowledge present in the initial time is not found anymore in the
final updated parameter. We propose a new approach to overcome some of the EnKF limitations. This paper shows the specifications
and results of the ensemble multiscale filter (EnMSF) for automatic history matching. EnMSF replaces, at each update time,
the prior sample covariance with a multiscale tree. The global dependence is preserved via the parent–child relation in the
tree (nodes at the adjacent scales). After constructing the tree, the Kalman update is performed. The properties of the EnMSF
are presented here with a 2D, two-phase (oil and water) small twin experiment, and the results are compared to the EnKF. The
advantages of using EnMSF are localization in space and scale, adaptability to prior information, and efficiency in case many
measurements are available. These advantages make the EnMSF a practical tool for many data assimilation problems. 相似文献
3.
In an attempt to derive more information on the parameters driving compaction, this paper explores the feasibility of a method
utilizing data on compaction-induced subsidence. We commence by using a Bayesian inversion scheme to infer the reservoir compaction
from subsidence observations. The method’s strength is that it incorporates all the spatial and temporal correlations imposed
by the geology and reservoir data. Subsequently, the contributions of the driving parameters are unravelled. We apply the
approach to a synthetic model of an upscaled gas field in the northern Netherlands. We find that the inversion procedure leads
to coupling between the driving parameters, as it does not discriminate between the individual contributions to the compaction.
The provisional assessment of the parameter values shows that, in order to identify adequate estimate ranges for the driving
parameters, a proper parameter estimation procedure (Markov Chain Monte Carlo, data assimilation) is necessary. 相似文献
4.
针对目前常用的储层地质模型不确定性评价方法存在的主要问题,提出了先用距离函数计算模型之间的差异,再以差异的大小来判断不确定性大小的方法.以WZ油田西区为例,采用相控物性参数建模技术,利用顺序高斯模拟方法建立渗透率的三维模型.对各种度量差异的距离函数进行对比研究和分析,结果显示曼哈顿距离函数和欧氏距离函数能较好地刻画模型之间的差异.选用欧氏距离函数计算模型之间的差异,其原理是先计算每两个模型之间相对应的每一网格节点的渗透率值差的平方和,然后取平方根,得到一个表征各模型之间差异的矩阵.根据该矩阵可得到各个模型之间的差异程度,差异越大,不确定性就越大.最后通过对比模型过井剖面图分析结果与距离矩阵分析结果,说明了本方法的正确性,结果显示该方法能有效评价随机模拟生成的储层地质模型的不确定性. 相似文献
5.
6.
Large-scale history matching with quadratic interpolation models 总被引：2，自引：0，他引：2
7.
In this paper, we describe a method of history matching in which changes to the reservoir model are constructed from a limited set of basis vectors. The purpose of this reparameterization is to reduce the cost of a Newton iteration, without altering the final estimate of model parameters and without substantially slowing the rate of convergence. The utility of a subspace method depends on several factors, including the choice and number of the subspace vectors to be used. Computational gains in efficiency result partly from a reduction in the size of the matrix system that must be solved in a Newton iteration. More important contributions, however, result from a reduction in the number of sensitivity coefficients that must be computed, reduction in the dimensions of the matrices that must be multiplied, and elimination of matrix products involving the inverse of the prior model covariance matrix. These factors affect the efficiency of each Newton iteration. Although computation of the optimal set of subspace vectors may be expensive, we show that the rate of convergence and the final results are somewhat insensitive to the choice of subspace vectors. We also show that it is desirable to start with a small number of subspace vectors and gradually increase the number at each Newton iteration until an acceptable level of data mismatch is obtained. 相似文献
8.
Analysis of remote images, elevation surveys, stratigraphic cross-sections, and hydrocarbon production data demonstrates that extensive areas of wetland loss in the northern Gulf Coast region of the United States were associated with large-volume fluid production from mature petroleum fields. Interior wetland losses at many sites in coastal Louisiana and Texas are attributed largely to accelerated land subsidence and fault reactivation induced by decreased reservoir pressures as a result of rapid or prolonged extraction of gas, oil, and associated brines. Evidence that moderately-deep hydrocarbon production has induced land-surface subsidence and reactivated faults that intersect the surface include: (1) close temporal and spatial correlation of fluid production with surficial changes including rapid subsidence of wetland sediments near producing fields, (2) measurable offsets of shallow strata across the zones of wetland loss, (3) large reductions in subsurface pressures where subsidence rates are high, (4) coincidence of orientation and direction of displacement between surface fault traces and faults that bound the reservoirs, and (5) accelerated subsidence rates near producing fields compared to subsidence rates in surrounding areas or compared to geological rates of subsidence. Based on historical trends, subsidence rates in the Gulf Coast region near producing fields most likely will decrease in the future because most petroleum fields are nearly depleted. Alternatively, continued extraction of conventional energy resources as well as potential production of alternative energy resources (geopressured-geothermal fluids) in the Gulf Coast region could increase subsidence and land losses and also contribute to inundation of areas of higher elevation. 相似文献
9.
A geomechanical model can reveal the mechanical behavior of rocks and be used to manage the reservoir programs in a better mode. Fluid pressure will be reduced during hydrocarbon production from a reservoir. This reduction of pressure will increase the effective stress due to overburden sediments and will cause porous media compaction and surface subsidence. In some oil fields, the compacting reservoir can support oil and gas production. However, the phenomena can also cause the loss of wells and reduced production and also cause irreparable damage to the surface structures and affect the surrounding environment. For a detailed study of the geomechanical behavior of a hydrocarbon field, a 3D numerical model to describe the reservoir geomechanical characteristics is essential. During this study, using available data and information, a coupled fluid flow-geomechanic model of Fahlian reservoir formation in X-field in SW of Iran was constructed to estimate the amount of land subsidence. According to the prepared model, in this field, the maximum amount of the vertical stress is 110 MPa and the maximum amount of the horizontal stress is 94 MPa. At last, this model is used for the prediction of reservoir compaction and subsidence of the surface. The maximum value of estimated ground subsidence in the study equals to 29 mm. It is considered that according to the obtained values of horizontal and vertical movement in the wall of different wells, those movements are not problematic for casing and well production and also the surrounding environment. 相似文献
10.
I. C. Kroon B.-L. Nguyen P. A. Fokker A. G. Muntendam-Bos G. de Lange 《Mathematical Geosciences》2009,41(5):571-584
Understanding and predicting surface movement is important both technically and for social reasons. The shallow processes
contributing to subsidence include construction works, peat oxidation, clay compaction, and groundwater withdrawal; deep causes
are hydrocarbon and salt production. We describe an inversion procedure we have devised to disentangle the deep and shallow
causes of surface movement. It employs a Bayesian inversion scheme, using forward models and other ‘a priori’ information
about shallow and deep compaction. Parameter estimation thus takes place at two different depths, thereby disentangling the
deep and shallow compaction processes responsible for surface movement. The uncertainty in the surface measurements and ‘a
priori’ estimates is naturally incorporated. Furthermore, spatial and temporal correlations can be taken into account through
inclusion of the covariance matrix. The inversion scheme is demonstrated for two synthetic cases. The first combines a compacting
gas field and a compacting shallow peat layer. We demonstrate that assumptions on the shape of the subsidence bowl are not
necessary. We also show how neglecting either deep or shallow causes of subsidence can produce spurious results. The advantage
of using the ‘a priori’ estimates of the compaction and the covariance matrix obtained by Monte Carlo simulations is demonstrated
with a second synthetic example involving two polders and different depths of their water table. A robust solution is obtained
for each polder unit, while a simpler (and faster) ‘a priori’ estimate based on the expected average clay thickness fails
to reproduce the actual compaction. Monte Carlo simulations can also be applied to compaction in depleting gas reservoirs.
Information on spatial correlations is often available, even when the absolute values of the ‘a priori’ compaction data are
quite uncertain. Explicitly incorporating such ‘a priori’ known spatial correlations improves the result significantly. 相似文献
11.
12.
In a recent paper, we developed a physics-based data-driven model referred to as INSIM-FT and showed that it can be used for history matching and future reservoir performance predictions even when no prior geological model is available. The model requires no prior knowledge of petrophysical properties. In this work, we explore the possibility of using INSIM-FT in place of a reservoir simulation model when estimating the well controls that optimize water flooding performance where we use the net present value (NPV) of life-cycle production as our cost (objective) function. The well controls are either the flowing bottom-hole pressure (BHP) or total liquid rates at injectors and producers on the time intervals which represent the prescribed control steps. The optimal well controls that maximize NPV are estimated with an ensemble-based optimization algorithm using the history-matched INSIM-FT model as the forward model. We compare the optimal NPV obtained using INSIM-FT as the forward model with the estimate of the optimal NPV obtained using the correct full-scale reservoir simulation model when performing waterflooding optimization. 相似文献
13.
Ensemble size is critical to the efficiency and performance of the ensemble Kalman filter, but when the ensemble size is small,
the Kalman gain generally cannot be well estimated. To reduce the negative effect of spurious correlations, a regularization
process applied on either the covariance or the Kalman gain seems to be necessary. In this paper, we evaluate and compare
the estimation errors when two regularization methods including the distance-dependent localization and the bootstrap-based
screening are applied on the covariance and on the Kalman gain. The investigations were carried out through two examples:
1D linear problem without dynamics but for which the true Kalman gain can be computed and a 2D highly nonlinear reservoir
fluid flow problem. The investigation resulted in three primary conclusions. First, if localizations of two covariance matrices
are not consistent, the estimate of the Kalman gain will generally be poor at the observation location. The consistency condition
can be difficult to apply for nonlocal observations. Second, the estimate of the Kalman gain that results from covariance
regularization is generally subject to greater errors than the estimate of the Kalman gain that results from Kalman gain regularization.
Third, in terms of removing spurious correlations in the estimation of spatially correlated variables, the performance of
screening Kalman gain is comparable as the performance of localization methods (applied on either covariance or Kalman gain),
but screening Kalman gain outperforms the localization methods in terms of generality for application, as the screening method
can be used for estimating both spatially correlated and uncorrelated variables, and moreover, no assumption about the prior
covariance is required for the screening method. 相似文献
14.
水库库区地形地质和水位地质条件复杂,蓄水后受降雨和库水位变动影响容易产生滑坡、崩塌等次生地质灾害,严重威胁水库安全运行和附近居民安全. 本文依托层次分析法,以某蓄水水库为研究对象,在充分收集其地形地质和水文条件资料的基础上,选取地形地貌、地层岩性、坡度、坡向、地灾点密度、地灾点面积、降雨、库水变动幅度和地震强度等9个致滑因子,构建评价矩阵和滑坡危险性计算评价方法. 依据评价成果划分4个滑坡危险性等级,借助MapGIS软件生成库区潜在滑坡危险性分区图. 该分区图与遥感解译的库区滑坡体分布点高度吻合,验证了评价模型的合理性. 相似文献
15.
在储层建模中利用多点地质统计学整合地质概念模型及其解释(英文) 总被引：1，自引：0，他引：1
基于三维空间中稀疏的观测数据,地质学家和储层建模人员尝试预测井间的地质沉积相的空间非均质性时,地质概念模型和先验认识在其中扮演着重要的角色。这种整合先验模型或解释的过程有时是隐蔽或不易察觉的,正如在手工绘等值线图中的情形;它也能够被显式地运用到某种算法当中,比如数字绘图中的算法。新近兴起的多点地质统计学为地质学家和储层建模人员提供了一种有力工具,它强调使用训练图像把先验模型明确而定量地引入到储层建模当中。先验地质模型包含了被研究的真实储层中确信存在的样式,而训练图像则是该模型的定量化表达。通过再现高阶统计量,多点算法能够从训练图像中捕捉复杂的(非线性)特征样式并把它们锚定到观测的井位数据。文中描述了多点地质统计学原理,以突出训练图像概念重要性为主线,描述了多点地质统计学在建立三维储层模型中的应用。 相似文献
16.
Alexey Viktorov 《Mathematical Geology》2007,39(8):735-748
The estimation of the risk caused by hazardous geological processes is a common problem. The aim of this paper is to present
a method for solving problems involving a wide spectrum of diffuse exogenous geological processes, based on the mathematical
morphology of landscapes. Diffuse processes develop as random rounded sites within certain areas and include karstification,
subsidence, thermo-karstification and aeolian processes. Model assumptions take into consideration spatial and temporal randomness
of occurrence, independent growth under occasional conditions, and cessation of growth given a ‘stop factor’, with transformation
into degenerate sites. Based on these model assumptions, the following probabilities of impact by diffuse exogenous geological
processes are solved for analytically. Firstly, the probability of impact on engineering constructions of round shape given
the size of the diffuse processes and the known average impact risk; secondly, the probability of impact on small engineering
constructions (points); and lastly, the probability of impact on linear engineering constructions of specified length. The
computed impact probabilities are verified in various areas within Western Siberia. 相似文献
17.
Giulia Codegone Vera Rocca Francesca Verga Christian Coti 《Geotechnical and Geological Engineering》2016,34(6):1749-1763
The multi-disciplinary work described in the paper was aimed at analyzing and predicting the cyclical ground surface movements induced by underground gas storage (UGS) activities in a depleted gas field located in the Po Plain (Italy). The field has been operated as a storage facility for nearly three decades. Currently, the possibility of delta-pressuring the reservoir (i.e. to increase the maximum operating pressure above the initial reservoir pressure) to enhance the storage performance is being considered. Significant information was collected over time: 2/3D seismic surveys, geological and sedimentological studies, 60+ logged wells, geotechnical lab tests and 50+ years of production history and monitoring were available for the development of a fully integrated static–dynamic–geomechanical analysis. The mechanical aspects of the study are the focus of this paper. The data coming from different sources at different scales were analysed and integrated to set up and characterize a 3D finite element method mechanical model to calculate the surface movements induced by UGS activity by adopting an elasto-plastic constitutive law. The model was then calibrated via a back analysis approach, i.e. the model parameters were fine-tuned so that the simulated subsidence/uplift would compare satisfactorily with the ground movements collected over nearly 10 years of monitoring via interferometric synthetic aperture radar analysis in the region under investigation. Eventually, the calibrated model was used as a forecasting tool for subsidence evaluation under different future storage strategies, including delta-pressuring conditions. Results proved that no significant subsidence is expected even if the maximum operating pressure reached 120 % of the initial formation pressure. 相似文献
18.
Roussos Dimitrakopoulos 《Mathematical Geology》1990,22(3):361-380
Conditional simulation of intrinsic random functions of orderk is a stochastic method that generates realizations which mimic the spatial fluctuation of nonstationary phenomena, reproduce their generalized covariance and honor the available data at sampled locations. The technique proposed here requires the following steps: (i) on-line simulation of Wiener-Levy processes and of their integrations; (ii) use of the turning-bands method to generate realizations in Rn; (iii) conditioning to available data; and (iv) verification of the reproduced generalized covariance using generalized variograms. The applicational aspects of the technique are demonstrated in two and three dimensions. Examples include the conditional simulation of geological variates of the Crystal Viking petroleum reservoir, Alberta, Canada. 相似文献
19.
Roussos Dimitrakopoulos 《Mathematical Geosciences》1990,22(3):361-380
Conditional simulation of intrinsic random functions of orderk is a stochastic method that generates realizations which mimic the spatial fluctuation of nonstationary phenomena, reproduce their generalized covariance and honor the available data at sampled locations. The technique proposed here requires the following steps: (i) on-line simulation of Wiener-Levy processes and of their integrations; (ii) use of the turning-bands method to generate realizations in Rn; (iii) conditioning to available data; and (iv) verification of the reproduced generalized covariance using generalized variograms. The applicational aspects of the technique are demonstrated in two and three dimensions. Examples include the conditional simulation of geological variates of the Crystal Viking petroleum reservoir, Alberta, Canada. 相似文献
20.
Assessment of ground subsidence hazard near an abandoned underground coal mine using GIS 总被引：15，自引：0，他引：15
Ki-Dong Kim Saro Lee Hyun-Joo Oh Jong-Kuk Choi Joong-Sun Won 《Environmental Geology》2006,50(8):1183-1191
This study constructs a hazard map for ground subsidence around abandoned underground coal mines (AUCMs) at Samcheok City in Korea using a probability (frequency ratio) model, a statistical (logistic regression) model, and a Geographic Information System (GIS). To evaluate the factors related to ground subsidence, an image database was constructed from a topographical map, geological map, mining tunnel map, Global Positioning System (GPS) data, land use map, lineaments, digital elevation model (DEM) data, and borehole data. An attribute database was also constructed from field investigations and reports on the existing ground subsidence areas at the study site. Nine major factors causing ground subsidence were extracted from the probability analysis of the existing ground subsidence area: (1) depth of drift; (2) DEM and slope gradient; (3) groundwater level, permeability, and rock mass rating (RMR); (4) lineaments and geology; and (5) land use. The frequency ratio and logistic regression models were applied to determine each factor’s rating, and the ratings were overlain for ground subsidence hazard mapping. The ground subsidence hazard map was then verified and compared with existing subsidence areas. The verification results showed that the logistic regression model (accuracy of 95.01%) is better in prediction than the frequency ratio model (accuracy of 93.29%). The verification results showed sufficient agreement between the hazard map and the existing data on ground subsidence area. Analysis of ground subsidence with the frequency ratio and logistic regression models suggests that quantitative analysis of ground subsidence near AUCMs is possible. 相似文献