Upscaling of Nonwetting Phase Residual Transport in Porous Media: A Network Approach

Based on the volume averaging method, a macroscopic model is developed for the upscaling of NAPL transport in a porous medium idealised by a network model. Under the assumption of local mass non-equilibrium, a macroscopic equation involving a dispersion tensor, additional convective terms and a linear form for the interfacial mass flux is obtained. The resolution of the two local closure problems obtained allow the determination of the local properties without adjustable parmeters. These problems are solved in a semi-analytical, semi-numerical manner on the network. The originality of this work is the association of the upscaling by volume averaging method with the network approach. The local properties, including the dispersion tensor and the mass exchange coefficient, can therefore be calculated over a large number of pore-bodies and pore-throats in a computationaly tractable manner, thus leading to more significant results. Results are presented for 3D, spatially periodic models of porous media.     

The Method of Averaging for Euler's Equations of Rigid Body Motion

We formulate the method of averaging for perturbations of Euler's equations of rotational motion. Euler's equations are three strongly nonlinear coupled differential equations that can be viewed as a three dimensional oscillator. The method of averaging is used to determine the long-term influence of perturbation terms on the motion by averaging about the nominal rigid body motion. The treatment is applicable to a large class of motions including precession with large nutation – it is not restricted to small motions about simple spins or nearly axi-symmetric bodies. Three examples are shown that demonstrate the accuracy of the method's predictions.     

范式理论与平均法的等价性分析

分析了范式理论与平均法的等价性。得到的结论是：对含有一对纯虚根的二维非线性系统，使用两种方法得到的结果是等价的，并提供了两个算例来证实其结论的正确性。虽然分析是针对一类二维非线性系统，但其结论同样适合于高维非线性系统。     

Averaging methods for numerical simulations of flows through heterogeneous porous media

Many natural rock systems contain small patches of different permeability which affect the flow of fluids through them. As these heterogeneities become smaller and more numerous, they become harder to model numerically. We consider how to reduce the computational effort required in simulations by incorporating their effects in the boundary conditions at the edges of each grid block. This is in contrast with current methods which involve often arbitrary changes in the fluid properties. The method is restricted to the case of widely-spaced patches, which simplifies interaction effects. The system then reduces to an array of dipoles, and two averaging methods are proposed for finite grid blocks. Several infinite systems, including vertical and horizontal bands, are also considered as further approximations. There is a great wealth of existing results from different fields which lead to identical mathematical problems and which can be used in these cases. Finally, we consider how to use these techniques when the precise configuration of the grid block is not known, but only its statistical properties. This can lead to results which are very different from the deterministic case.     

A study on the measurement of mean velocity and its convergence in molecular dynamics simulations

In many particle‐based simulations, measurement of local mean flow velocity and other continuum‐based properties are of utmost importance. Macroscopic quantities, such as mean flow velocity, temperature, and density, can be estimated by averaging the corresponding microscopic behavior of the particles. The two main subjects that should be considered in the averaging over the particles in a specific problem are spatial and temporal behaviors of them. In this paper, we study the latter. Because of the chaotic nature of the collisions among the molecules and consequently their random path, extracted macroscopic values fluctuate about their average values causing statistical errors. In this paper, an averaging method called SAM‐Modified‐CAM (SMC) will be proposed for the measurement of mean velocity that reduces statistical errors in its calculation. This proposal is based on the study conducted here on the implementations of two common averaging methods, sample‐averaged measurement (SAM) and cumulative average measurement (CAM) in molecular dynamics. In addition, convergence of mean flow velocity measurement is thoroughly discussed, and a convergence criterion is proposed for this purpose. Implementation of the proposed method in different test cases has approved its reliable performance. Copyright © 2010 John Wiley & Sons, Ltd.     

The reduced detailed configuration accounting (RDCA) model for NLTE plasma calculations

The motivation for this work is to provide a more accurate and cost effective in-line NLTE capability for calculating plasma properties in large scale radiation-transport hydrodynamic simulations. A method is developed to transform the large detailed atomic models to very small models that can be used for fast in-line calculations. An averaging technique is used to reduce detailed models involving tens of thousands of states into just a few, 10–40 states, per ionization state. The reduced model is more accurate than the average atom models conventionally used in such simulations. In the present work, the averaging scheme is presented and results of the reduced model are compared to the original detailed model and the average atom model. The average ionization state of iron and uranium plasmas under various conditions of material temperature, atom number density, and radiation temperature are compared.     

Structural averaging of flow processes in nonhomogeneous porous media

The problem of constructing macroscopic analogs for the equations describing processes in nonhomogeneous porous media is considered. The classical results of the theory relate to the case in which the averaging procedure leads to the smoothing of the coefficients describing the inhomogeneity without modifying the structure of the equations of the process. It is natural to call such averaging coefficient averaging. In this paper another approach — structural averaging, in which the type of the equations themselves or their qualitative structure is modified, is investigated. In the overwhelming majority of cases, in addition to a small scale of inhomogeneity, these systems also contain one or more small (large) parameters reflecting important differences in the properties of the individual components of the medium or the physical components of the transport process itself. A typical example of the structural averaging problems generated by processes in highly nonhomogeneous media and, moreover, processes with nonequivalent diffusion and convective transport is investigated. The methods of asymptotic averaging [1,2] are employed. Processes in highly nonhomogeneous media were investigated in [3–6]. Studies [4, 8, 9] are concerned with the averaging of convection-diffusion systems.Based on paper read at the Seventh Congress on Theoretical and Applied Mechanics, Moscow, August 1991. Presented by V. N. Nikolaevskii.Translated from Izvestiya Rossiiskoi Akademii Nauk, Mekhanika Zhidkosti i Gaza, No.6, pp. 103–116, November–December, 1992.     

Random perturbations of nonlinear oscillators: Homogenization and large deviations

This paper considers two prototypical strongly nonlinear oscillators with weak dissipation and noise: a pendulum which is acted on by a small noisy torque and opposed by a friction; and the noisy Duffing-van der Pol oscillator. But over a long time the weak dissipative and noise effects can be significant. This paper develops asymptotic techniques based on averaging and large deviations to study the effects of weak noise on the escape from the domain of attraction of stable equilibrium points and limit cycles in phase-space.     

A generalization of the concept of equivalent linearization

A method is presented whereby a non-linear second order dynamical system is replaced by a linear system in such a way that an average of the difference between the two systems is minimized. Provided the averaging operator possesses certain properties, it is shown that the replacement is unique and can be accomplished in a straightforward manner. The parameters of the replacement linear system are expressed in terms of averages of functions of the linearized solution.     

Dynamics of self-excited oscillators with neutral delay coupling

The work is devoted to analytic and numeric investigation of dynamical behavior in a system of two Van der Pol (VdP) oscillators coupled by a non-dispersive elastic rod. The model is rigorously reduced to a system of nonlinear neutral differential delay equations. For the case of relatively small coupling and moderate delay, an approximate analytic investigation can be accomplished by means of an averaging procedure. The region of synchronization in the space of parameters is established and characteristic bifurcations are revealed. A numeric study confirms the validity of the analytic approach in the synchronization region. Beyond this region, the averaging approach is no more valid. A multitude of quasiperiodic and chaotic-like orbits has been revealed. Especially interesting behavior occurs in the case of relatively large delays and corresponds to sequential quenching and excitation of the VdP oscillators. This regime is also explored analytically, by means of a large-delay approximation, which reduces the system to a perturbed discrete map.     

Responses of a non-linearly coupled pitch-roll ship model under harmonic excitation

Stationary responses of nonlinearly coupled pitch and roll ship modes are studied using a modified averaging method, along with two second order multiple time scale (MTS) methods for comparison. Stability of the solutions is also studied. In the case of harmonic excitation all the three methods give fairly accurate results to the original problem but the averaging method is the most efficient. Analytic solutions are obtained from the averaged equations, which can be used to predict stationary responses both for small and for large excitations. From the averaging method several qualitatively different phenomena which cannot be addressed by the first order theory have been obtained: (i) the saturation phenomenon is lost, (ii) the bifurcation points are altered and (iii) a drift term is present which, although small, appears to have a significant effect on the accuracy of the solutions.     

A spectral approach to inertial confined thin‐film flow

A spectral approach is proposed to determine the flow field of a thin film inside narrow channels of arbitrary shape. Although the method is easily extended to transient flow, only steady flow is considered here. The flow field is represented spectrally in the depthwise direction in terms of orthonormal shape functions, which together with the Galerkin projection lead to a system of ordinary differential equations that can be solved using standard methods. The method is particularly effective for nonlinear flow, including nonlinearities of geometrical or material origins. The validity of the proposed method is demonstrated for a flow with inertia, and, unlike the depth‐averaging method, is not limited to a flow at small Reynolds number. The problem is closely related to high‐speed lubrication flow. The validity of the spectral representation is assessed by examining the convergence of the method, and comparing it with the fully two‐dimensional finite‐element solution, and the widely used depth‐averaging method from shallow‐water theory. It is found that a low number of modes are usually sufficient to secure convergence and accuracy. The influence of inertia is examined on the velocity and pressure fields. The pressure distributions reflect excellent agreement between the low‐order spectral method and the finite‐element solution, even at moderately high Reynolds number. The depth‐averaging solution is unable to predict accurately (qualitatively and quantitatively) the high‐inertia flow. Comparison of the velocity field reflects the expected discrepancy in a boundary layer formulation. Copyright © 2012 John Wiley & Sons, Ltd.     

Laminate system schemes for effective property estimates of architectured composites with co-(dis)continuous phases

Effective properties of fiber and laminate systems of ellipsoidal overall symmetry result from a weighted spatial average of either the effective stiffness or the effective compliance of directional fibers or laminate structures. Such systems can be considered for various (elastic, dielectric, magnetic, etc.) types of properties, including coupling. Effective properties from such fiber and laminate systems have been shown relevant for a large variety of material architectures with various phase continuity/connectivity features (Franciosi et al., 2011). For laminate systems (LS) here of concern, it has been further argued (Franciosi and El Omri, 2011) why a stiffness (resp. compliance) averaging relevantly describes two-phase assemblages with both (resp. none) through-sample continuous phases. The provided microstructural interpretations from morphological considerations resulted from a comparison of these LS schemes with the self-consistent approximation in the two-phase case. It is here shown that (i) the so established relevancy of laminate systems to represent two-phase materials with co-continuous or co-discontinuous phases does not extend to n-phase composites, (ii) for large classes of structures with any number n of co-(dis)continuous phases, referring to an equivalent two-phase material is an alternative way to establish this relevancy.     

Parametric excitation in non-linear dynamics

Consider a one-mass system with two degrees of freedom, non-linearly coupled, with parametric excitation in one direction. Assuming the internal resonance 1:2 and parametric resonance 1:2 we derive conditions for stability of the trivial solution by using both the harmonic balance method and the normal form method of averaging. If the trivial solution becomes unstable, a stable periodic solution may emerge, there are also cases where the trivial solution is stable and co-exists with a stable periodic solution; if both the trivial solution and the periodic solution(s) are unstable, we find an attracting torus with large amplitudes by a Neimark-Sacker bifurcation. The results of the harmonic balance method and averaging are compared, as well as the results on the Neimark-Sacker bifurcation obtained by the numerical software package CONTENT and by averaging. In all cases we have good agreement.     

Bin size determination for the measurement of mean flow velocity in molecular dynamics simulations

Extracting macroscopic properties from molecular simulation of fluids is required in most of the molecular dynamics problems. However, methods used for this purpose, their accuracy, and their dependence on the sampling and averaging schemes are still ambiguous. Macroscopic properties at any point of a molecular domain can be extracted via sampling and averaging of molecular behavior within a control region around that point, called a bin. The size of this bin has a significant impact on the accuracy of the macroscopic properties' measurement. In this research, we will focus on the measurement of mean flow velocity using ‘binning method’. On the basis of the results of this study, the most appropriate range of the bin size in which mean velocity can be measured accurately is determined. Although this range is determined based on the measurement of mean flow velocity, we believe that it can be used to estimate the proper range for the measurement of other flow quantities as well. Copyright © 2012 John Wiley & Sons, Ltd.     

Damping of steady-state waves in systems described by a nonlinear Klein-Gordon equation

The damping of a nonsinusoidal wave in systems described by a Klein-Gordon equation is investigated by the method of averaging. An explicit solution is given for an initial-value problem. It is shown that in certain cases the prolonged existence of a steady-state wave is impossible. Dissipation can lead to the damping out of the wave. The characteristic features of the boundary-value problem are discussed. Formulas are obtained describing the damping of single pulses (solitons).Translated from Zhurnal Prikladnoi Mekhaniki i Tekhnicheskoi Fiziki, No. 5, pp. 55–59, September–October, 1974.     

Double-mode modeling of chaotic and bifurcation dynamics for a simply supported rectangular plate in large deflection

This paper presents a new approach to characterize the conditions that can possibly lead to chaotic motion for a simply supported large deflection rectangular plate by utilizing the criteria of the fractal dimension and the maximum Lyapunov exponent. The governing partial differential equation of the simply supported rectangular plate is first derived and simplified to a set of two ordinary differential equations by the Galerkin method. Several different features including Fourier spectra, state-space plot, Poinca?e map and bifurcation diagram are then numerically computed by using a double-mode approach. These features are used to characterize the dynamic behavior of the plate subjected to various excitation conditions. Numerical examples are presented to verify the validity of the conditions that lead to chaotic motion and the effectiveness of the proposed modeling approach. The numerical results indicate that large deflection motion of a rectangular plate possesses many bifurcation points, two different chaotic motions and some jump phenomena under various lateral loading. The results of numerical simulation indicate that the computed bifurcation points can lead to either a transcritical bifurcation or a pitchfork bifurcation for the motion of a large deflection rectangular plate. Meanwhile, the points of pitchfork bifurcation can gradually lead to chaotic motion in some specific loading conditions. The modeling result thus obtained by using the method proposed in this paper can be employed to predict the instability induced by the dynamics of a large deflection plate.     

Optimisation of temporal averaging processes in PIV

A hybrid of correlation and vector averaging is introduced to capitalise on the advantages of each process. An extensive series of Monte Carlo simulations have been conducted to investigate hybrid averaging and evaluate it against both vector and correlation averaging. The simulations show that hybrid averaging improves the measurement accuracy over both correlation and vector averaging over a wide range of imaging conditions. The simulations are validated by applying hybrid averaging to experimental micro- and macro-flows. In pulsatile conditions, correlation averaging yields an averaged correlation function that is multi-modal, which can result in unpredictable measurements. A Monte Carlo simulation shows the benefits of hybrid averaging over correlation averaging in such conditions. This has been experimentally validated on the unsteady wake behind a shedding circular cylinder at Re = 98.     

Convergence of local atomistic stress based on periodic lattice

Local stress in an atomic system, which provides an average stress measurement within a spatial volume containing a collection of atoms, is essential for determining the mechanical properties of a nanoscale structure as well as developing a proper multiscale modeling technique. Theoretically, the smaller averaging volume where a local stress can converge, the closer this atomistic stress definition can approach the ideal continuum stress. As a result, the more accurate stress concentration can be evaluated for the inhomogeneous case. With reference to the previous studies focusing on the spherical averaging volume, dependent on the type of crystals, the convergent radius of the virial stress or Hardy stress usually spans the size of several lattice constants. In this paper, we find that, once the averaging volume is periodic, the convergence of the virial stress and Hardy stress can be accomplished within one single lattice, which is much smaller than what is required by other non-periodic volumes such as a sphere. In the final section, a cracked sodium chloride crystal is considered to demonstrate that the crack opening stress described by the periodic lattices captures the stress concentration near the crack tip.     

Rational approaches for combining redundant,independent measurements to minimize combined experimental uncertainty

General methods for combining multiple, independent measurements for minimized uncertainty are presented. Although the methods are general, the presentation is motivated by the reduction of heat exchanger performance data. It is demonstrated that the prevalent practice of using the arithmetic mean of two measured heat transfer rates for the evaluation of heat exchanger performance, which is widely adopted in the literature and used in engineering standards, does not always lead to reduced experimental uncertainty and very rarely—if ever—leads to a minimized experimental uncertainty. For the case of averaging two redundant measurements, simple criteria are developed to determine whether using the arithmetic mean or a single measurement yields a smaller absolute or relative uncertainty. Then, more general methods are proposed in the form of weighted-linear averages, in which the weight factors are determined to minimize the combined uncertainty. The methods are generally applicable to averaging any number of redundant measurements with varying uncertainties.