Approximate Derivative Computations for the Gradient-Based Optimization Methods in the Local Gradual Deformation for History Matching

Reservoir characterization needs the integration of various data through history matching, especially dynamic information such as production or four-dimensional seismic data. To update geostatistical realizations, the local gradual deformation method can be used. However, history matching is a complex inverse problem, and the computational effort in terms of the number of reservoir simulations required in the optimization procedure increases with the number of matching parameters. History matching large fields with a large number of parameters has been an ongoing challenge in reservoir simulation. This paper presents a new technique to improve history matching with the local gradual deformation method using the gradient-based optimizations. The new approach is based on the approximate derivative calculations using the partial separability of the objective function. The objective function is first split into local components, and only the most influential parameters in each component are used for the derivative computation. A perturbation design is then proposed to simultaneously compute all the derivatives with only a few simulations. This new technique makes history matching using the local gradual deformation method with large numbers of parameters tractable.     

一井定向在立井联系测量中的应用

详细介绍了在立井平面联系测量中连接三角形法一井定向的测量和解算方法,特别提出了不同连接三角形解算适用不同的计算公式。并结合实践经验,阐述了一井定向连接三角形法和瞄直法的使用范围和测量注意事项。     

利用有限单元法模拟二维MT倾子响应

在二维地电模型中,考虑到大地电磁测深(MT)倾子响应依赖于横电(TE)模式的数值模拟问题,笔者采用矩形网格单元和双二次插值对MT的倾子响应进行求解。首先,给出了二维地电条件下的边值问题和变分问题,并通过有限单元法对模型进行单元剖分、插值、积分和整体合成;其次,通过求解复系数方程组得到了每个节点的电场值;最后,采用差分方法求得电场沿横向和垂向的偏导数值,从而计算出二维MT的倾子响应。通过对2个二维地电模型的倾子响应进行数值模拟,获得了倾子响应的实部、虚部和振幅,结果很好地反映了不均匀体的横向分布情况。     

电磁波场数值计算中时间域有限差分法与时间域伪谱法的对比研究

为提高电磁波场数值计算精度,对时间域有限差分法与时间域伪谱法进行了对比研究。时间域有限差分法是一种目前流行的电磁场时域数值计算方法,已被广泛应用于求解与时间有关的偏导数方程。对于大规模数值计算,时间域有限差分法需要较多的内存空间。时间域伪谱法基于时间域有限差分法,该方法使用快速傅里叶度换来计算麦克斯韦方程中的空间导数。由于傅里叶变换的准确性,时间域伪谱法使计算精度提高,数值计算时所需的格子数大大减少,这极大地节省了计算机内存空间,适合于大规模正反演问题的数值计算。     

基于三角元法计算三度体球冠模型的重力异常

提出利用高斯公式将体积分变成包围该积分域的全表面积,应用一系列三角形拟合计算表面,然后通过格林公式把对每个三角形的面积分变成对三角形每边的线积分之和的三角元法来计算三度体球冠模型重力异常,同时给出了合理的计算方案。计算了均匀密度球冠模型和非均匀密度球冠模型在水平地表面的重力异常,得到了2种模型的平面异常图。最后通过选择均匀密度球体作为误差对比的对象进行误差分析,验证了该方法的可行性以及计算结果的可靠性。     

水环境质量综合评价的多目标决策-理想区间法

对水环境质量综合评价中应用的多目标决策-理想点法进行了分析,认为实际上水环境质量评价标准是区间概念而非点的概念,因而传统的多目标决策-理想点法把评价标准处理成理想点存在一定缺陷。为了克服这种缺陷,提出了一种多目标决策-理想区间法(MODMIIM)。介绍了MODMIIM的原理,给出了它的构成方法,并把它应用到水环境质量综合评价中。MODMIIM与PP模型和神经网络法相比,简单、有效;与多目标决策理想点法相比,具有更大的适用性。MODMIIM可广泛应用于各种环境质量综合评价中。     

基于异常复电位2.5维CR有限元数值模拟

论述了基于异常复电位2.5维复电阻率法的正演计算技术。首先给出了二维地电断面、三维交流双点电源复电阻法的变分问题;然后利用有限单元法求解变分问题,在矩形网格中增加2条对角线形成三角网格对研究区域进行剖分,在单元中对复电位和岩、矿石复电导率进行线性插值,将变分方程化为线性代数方程组;最后,求解线性方程组,得到各节点的异常复电位值,并根据选定的测量装置计算出地表的视复电阻率,进而计算出复电阻率的振幅、相位或复电阻率的实、虚分量。通过理论模型的计算检验算法的有效性后构建模型算例,给出了其不同测量装置下的计算结果。     

Numerical modeling of the flow and transport of radionuclides in heterogeneous porous media

This paper is concerned with numerical methods for the modeling of flow and transport of contaminant in porous media. The numerical methods feature the mixed finite element method over triangles as a solver to the Darcy flow equation and a conservative finite volume scheme for the concentration equation. The convective term is approximated with a Godunov scheme over the dual finite volume mesh, whereas the diffusion–dispersion term is discretized by piecewise linear conforming triangular finite elements. It is shown that the scheme satisfies a discrete maximum principle. Numerical examples demonstrate the effectiveness of the methodology for a coupled system that includes an elliptic equation and a diffusion–convection–reaction equation arising when modeling flow and transport in heterogeneous porous media. The proposed scheme is robust, conservative, efficient, and stable, as confirmed by numerical simulations.      

二维扩散输移问题的一种新的有限体积算法

从空间离散格式上比较了有限元格式与有限体积格式的异同处,证明有限体积格式是有限元格式的一种特例,有限元格式也是一种守恒型格式。并根据有限元格式与有限体积格式的比较,提出了一种新的高精度、稳定性好的有限体积格式。通过模拟非定常的纯对流方程和二维对流扩散方程,证明了有限体积格式的优点。     

粗粒料多边形表征及二维FEM/DEM分析

考虑到颗粒形状对粗粒料的力学特性有重大影响,提出了一种新的表征颗粒形状的方法,即在椭圆上随机选取一系列点连接成多边形颗粒,表征狭长扁平的颗粒。新方法较圆上取点的方法能代表更多类型的颗粒形状,适用范围更广。提出了一种新的粗粒料投放算法,即先缩小颗粒,然采用随机算法将缩小的颗粒投放至给定区域,对颗粒划分好网格后,将颗粒放大到原来的大小,然后采用有限元-离散元(FEM/DEM)方法计算稳定后即生成了相应的试样。通过将上述颗粒生成及投放算法与FEM/DEM结合,应用于粗粒料的数值模拟。分析表明,FEM/DEM是研究粗粒料力学性质的较好方法,对复杂的颗粒形状也可简单建模,且因在颗粒内部划分了有限元网格,复杂的接触判断及接触力计算转化为标准统一的三角形和三角形之间的接触判断及接触力计算,所有的计算均可标准化、统一化。同时因为颗粒是可以变形的,依然保留了连续介质力学中应力和应变的概念,无须像PFC那样需通过测量圆来间接表示某点的应力、应变。最后,通过粗粒料的侧限压缩试验的数值模拟,展现了文中提出的一整套解决方案在模拟粗粒料方面的巨大潜力。     

位场垂向高阶导数的Tikhonov正则化迭代法

在位场数据处理中,垂向导数具有重要的物理意义。其在一定程度上可以划分不同深度和大小异常源产生的叠加异常,且导数的阶次越高,这种分辨能力就越强,但通常认为高阶导数的换算是不稳定的。本文在Tikhonov正则化求位场垂向高阶导数的基础上,结合迭代法进行逐次逼近,提出了位场高阶导数的Tikhonov正则化迭代法,并且得到Tikhonov正则化迭代法的递推公式。通过对该方法的滤波特性分析可以看出,该方法计算的位场垂向高阶导数具有一定的稳定性及保幅性。模型试验和实际数据的处理表明,该方法计算结果较常规FFT求导法有更高的稳定性和实用价值。     

Finite element modelling for evaluation of apparent resistivity over complex structures

The paper presents a computational algorithm designed for efficient modelling of apparent resistivity over complex geological structures, using finite element method. The algorithm can be used to study variations of apparent resistivities using any electrode configuration at any point on the earth’s surface, not necessarily regular. A Schlumberger apparent resistivity sounding curve over a buried anticline, is presented here as an example and compared with the corresponding analytical curve, to demonstrate the correctness of the FEM algorithm. The various potential derivatives required for the computation of apparent resistivities evaluated through different electrode configurations have been obtained by calculating the ‘influence coefficients’ using reciprocal theorems, an approach successfully applied in structural engineering. In essence, a set of self balancing nodal currents, obtained from the appropriate derivative(s) of the shape functions of the elements contributing to the point of observation, is applied as the load vector. The resulting quantities corresponding to the potential distribution in traditional finite element method, then, turn out to be the potential derivatives at the point of observation for different positions of the current electrodes. These are known as influence coefficients. The continuum nature of the domain beyond the region of interest has been modelled by using ‘infinite elements’ across which the potential is assumed to decay exponentially.     

Kalman滤波在气象数据同化中的发展与应用

气象学领域各种观测(特别是遥感遥测等非常规观测)数据的大量增多和数值天气预报模式的不断进步,推动气象数据同化技术不断发展。回顾了Kalman滤波在气象数据同化中的引入和几个发展阶段;介绍了Kalman滤波(尤其是简化Kalman滤波和总体Kalman滤波)在气象数据同化中的重要地位和应用进展。     

位场向下延拓的导数迭代法

针对向下延拓的不适定问题,提出了一种新的位场向下延拓方法-导数迭代法。从位场垂向一阶导数的定义出发,将观测面和向上延拓、向下延拓同等高度平面上的位场值近似联系起来,采用迭代法进行逐次逼近,推导出空间域位场向下延拓的导数迭代法的递推公式。考虑到空间域向上延拓积分方程实现的复杂性,对递推公式进行快速傅里叶变换,整理得到波数域中的导数迭代公式,同时证明了该方法的收敛性。模型检验和实例分析均表明迭代法相对直接FFT法具有稳定性强和下延深度大的优点。     

An unconditionally stable explicit and precise multiple timescale finite element modeling scheme for the fully coupled hydro-mechanical analysis of saturated poroelastic media

An unconditionally stable, fully explicit and highly precise multiple timescale finite element modeling scheme is described for a fully coupled hydro-mechanical (FCHM) analysis of saturated poroelastic media. The finite element method (FEM) is used for the discretization of the FCHM differential equation in the space domain. Direct integration is performed based on the precise time step integration method (PTSIM) for the time derivatives. Two configurations for the proposed scheme are constructed (abbreviated as PTSIM-f1 and -f2, respectively). The stability and convergence of the PTSIM-f1 and -f2 are proved using a matrix-based spectral analysis in the time domain. It is demonstrated that the explicit scheme proposed in this paper is unconditionally stable and independent of the time-step size. The algorithmic error estimation results indicate that the numerical modeling performed using PTSIM-f1 and -f2 in the time domain match the computer precision. Theoretically, the algorithmic error is caused by only the mesh discretization. Therefore, the proposed modeling scheme is a semi-analytical scheme. The applicability and accuracy of the proposed scheme are examined using sample calculations. By comparing with the analytical solutions, it is indicated that the modeling results have significant advantages over the standard FEM in terms of precision and computational efficiency for large timescales.     

Validation of theoretical footprint models using experimental measurements of turbulent fluxes over maize fields in Po Valley

Representative source area of turbulent fluxes measured by eddy covariance stations is an important issue which has not yet been fully investigated. In particular, the validation of the analytical footprint models is generally based on the comparison with Lagrangian model predictions, while experimental results are not largely diffused in literature. In this work, spatial distribution of carbon dioxide, latent and sensible heat fluxes across two different maize fields in Po Valley, is used to validate two theoretical footprint models. Experiments are performed in two totally different scenarios at bare and vegetated soils using two eddy covariance systems: one fixed station which is located about in the middle of the field and a mobile station which is placed at various distances from the field edge to investigate the horizontal variation of the vertical scalar fluxes. The first objective of this work is to provide detailed information about the spatial distribution of turbulent fluxes across Po Valley characteristic fields at bare and vegetated soils, highlighting peculiarities and uniqueness. The second objective consists in the comparison between mobile measurements of carbon dioxide, latent and sensible heat fluxes and the predictions of two analytical footprint models widely used in literature. Contemporaneously, the latter objective will permit to understand what is the best footprint model which, under typical Po Valley atmospheric turbulent conditions, describes a representative source area compatible with the field dimensions and the turbulent flux distributions. The results show that both models are in good agreement with experimental measurements. The results also show that the spatial distribution of turbulent fluxes is strongly influenced by the presence of vegetation in the field. Moreover, the representative source area is different for different scalar fluxes. Another result is about 10:1 fetch-to-height obtained for both field situations.     

三维地层的三棱柱剖分与土方计算

三维地层剖分算法是实现地层三维可视化与土方计算的重要基础。采用三棱柱体建模方法构造三维地层,分析任意平面切割三棱柱的不同情形,得出了针对三棱柱地层的剖分算法;提出了以切割面为分界面,将单个三棱柱剖分成多个小三棱柱的三角投影分割法--将三棱柱顶面、底面及切割截面按一定规则进行合理三角化划分,使划分后三角形在水平面上的投影相重合;将该3层三角形组合成以切割面为分界面的2层小三棱柱,从而完成剖分。土方计算中首先给出开挖体的边界控制方程,根据该控制方程所限定的内边界得到开挖区域,结合上述剖分算法,对该区域内所有小三棱柱体积求和,得到开挖土方量。将以上算法应用到广东省某重大水利工程中,实现了基坑模拟开挖及其土方计算。     

二维视电阻率断面的快速最小二乘反演

快速最小二乘反演是以平滑限定的最小二乘方法为基础,是对二维视电阻率断面进行反演的一种方法。反演过程不需要提供初始模型,在首次迭代时使用一均匀介质地下模型作为初始模型,该模型的视电阻率偏导数值可以用解析法得到。在后面的迭代中,使用了拟牛顿法去修改每一次迭代的偏导数矩阵,避免了偏导数矩阵的直接计算,从而减少了计算时间和存储空间。同时运用牛顿矩阵校正技术解最小二乘方程组也减少了大量的计算时间。总之,该方     

利用水平导数与垂向导数标准偏差的相关系数法识别磁源边界

基于导数换算的边界检测技术在位场数据处理与解释中具有重要应用,但是高阶导数算法极易受到干扰噪声的影响,稳定性较差。鉴于此,本文提出利用水平导数与垂向导数标准偏差的相关系数进行边界检测,该方法是基于水平导数与垂向导数的标准偏差在地质体边界位置处呈现负相关性实现的。理论模型试验表明:该方法不仅可以识别出不同深度的异常体的边界位置,而且受噪声影响较小,计算稳定性更强。将本文方法应用于海拉尔盆地实测航磁数据,检测结果与实际资料有着较好的对应关系。     

Short range prediction with a multi-level primitive equation model

A five-level primitive equation model in a (x, y, p, t) coordinate system has been developed. A fairly sophisticated scheme of physical processes has been incorporated in the model. The model physics include air-sea interaction, cumulus parametrization, large scale condensation, dry convective adjustment, horizontal and vertical diffusion and simulated radiation. The initial balance between mass and motion fields has been obtained through a dynamic initialization scheme. The model has been integrated upto 48 hr using input data of a case of monsoon depression. The results of initialization and forecast have been presented and discussed. Wind, temperature and vertical velocity fields have been found to retain the observed map features; after the initialization, however, the surface pressure has been considerably modified. The model produced a reasonably good forecast up to 24 hr as far as the flow fields, rainfall region, structure of the depression and the movement of cyclonic circulation were concerned and beyond that damped rapidly. The rainfall rates were underestimated. Some of the shortcomings of the model are also discussed.