求解非线性反问题的大范围收敛梯度正则化算法

基于同伦映射的思想,改进了求解非线性反问题的梯度正则化算法。通过路径跟踪有效地拓宽了梯度正则化算法求解的收敛范围。对于正则化参数的修正,通过引入拟Sigmoid函数,提出了一种下降速率可调的连续化参数修正方法,在保证迭代稳定的条件下,得到较好的计算效率,同时保证该算法具有很好的抵抗观测噪声能力。实际算例表明,该方法收敛范围宽,计算效率高,在存在较强观测噪声的条件下也能得到很好的反演结果。     

基于同伦方法反演非饱和土中镉离子传输参数

依据实验室土柱实验结果，引入同伦方法对镉离子在不同非饱和土样中的传输参数进行了反演计算，得到了相应的反演结果．     

跨音速翼型反设计的一种大范围收敛方法

求解跨音速翼型的反设计问题时，传统的梯度型方法一般均为局部收敛. 为增大求解的收敛范围，依据同伦方法的思想，通过构造不动点同伦，将原问题的求解 转化为其同伦函数的求解，并依据拟Sigmoid函数调整同伦参数以提高计算效率，进而构造 出一种具有较高计算效率的大范围收敛反设计方法. 数值算例以RAE2822翼型的表面压力分 布为拟合目标，分别采用B样条方法, PARSEC方法及正交形函数方法等3种不同的 参数化方法，并分别以NACA0012, OAF139及VR15翼型为初始翼型进行迭代计 算. 计算结果证明，该方法适用于多种参数化方法，且具有较好的计算效率，从多 个不同的初始翼型出发，经较少次数迭代后， 均能与目标翼型很好地拟合，是一种高效的大范围收敛方法.     

地下水流并行有限层方法及同伦反演研究

根据有限层求解格式存在的解耦性，实现了地下水三维流问题的高效并行化计算。在此基础上，结合非线性同伦方法，提出了地下水参数反演分析的并行同伦算法，利用MATLAB编译了相应的正反演计算程序。与已有解析解和有限差分解的对比以及数值算例，验证了并行化正反演方法及程序的正确性，探讨了并行算法的计算效率。研究表明，并行方法可以有效提高计算速度，较串行方法具有明显优势，同时同伦反演方法具有大范围收敛的特点，不依赖于参数值的初始选取。     

多宗量瞬态热传导反演识别

引入Bregman距离加权函数,建立了多宗量瞬态热传导反演的一种求解模式.时域上采用精细算法,分别建立了便于敏度分析的有限元正/反演模型,应用同伦算法进行反问题求解,对导热系数和边界条件等宗量进行有效的组合识别.对信息误差和计算效率作了探讨,并给出了相应的数值验证.     

基于同伦技术的偶应力反问题求解

引入Bregman距离构造同伦函数,建立偶应力反问题的一种求解模式,可以对偶应力理论相关参数进行识别.利用有限元技术,建立了偶应力正/反问题数值求解模型,该模型不仅考虑了非均质的影响,而且也便于反问题的敏度分析.文中给出了相关的数值算例,并对信息误差和不同的函数形式计算结果做了初步探讨,得到满意的结果.数值算例表明该模...     

超越摄动:同伦分析方法基本思想及其应用

介绍一种新的、求解强非线性问题解析近似的一般方法------同伦分析方法.该方法从根本上克服了摄动理论对小参数的过分依赖, 其有效性与所研究的非线性问题是否含有小参数无关, 因此,适用范围广.此外, 不同于所有其他解析近似方法,同伦分析方法提供了一个简单的途径, 确保所得到的级数解收敛, 从而获得足够精确的解析近似.而且, 不同于所有其他解析近似方法, 同伦分析方法(HAM)提供了选取基函数之自由, 从而可以选择较好的基函数, 更有效地逼近问题的解.同伦分析方法为非线性问题的解析近似求解提供了一个全新的思路, 为非线性问题(特别是不含小参数的强非线性问题)的求解开辟了一个全新的途径.简要描述同伦分析方法的基本思想, 其在非线性力学、物理、化学、生物、金融、工程和计算数学等领域的应用举例, 以及与摄动方法、Lyapunov 人工小参数法、$\delta$展开法、Adomian 分解法、同伦摄动方法之区别和联系.      

基于精细积分技术的非线性动力学方程的同伦摄动法

将精细积分技术（PIM）和同伦摄动方法（HPM）相结合,给出了一种求解非线性动力学方程的新的渐近数值方法。采用精细积分法求解非线性问题时,需要将非线性项对时间参数按Taylor级数展开,在展开项少时,计算精度对时间步长敏感;随着展开项的增加,计算格式会变得越来越复杂。采用同伦摄动法,则具有相对筒单的计算格式,但计算精度较差,应用范围也限于低维非线性微分方程。将这两种方法相结合得到的新的渐近数值方法则同时具备了两者的优点,既使同伦摄动方法的应用范围推广到高维非线性动力学方程的求解,又使精细积分方法在求解非线性问题时具有较简单的计算格式。数值算例表明,该方法具有较高的数值精度和计算效率。     

同伦分析方法：一种不依赖于小参数的非线性分析方法

本文进一步一般化了作者所提出的一种新的非线性分析方法，称为“同伦分析方法”。“同伦分析方法”的最大优点，在于其不依赖小参数，从而可求解更多的非线性问题，甚至包括那些不含小参数的问题。本文给出了几个应用实例，以说明“同伦分析方法”的有效性及巨大的潜力。     

建立在同伦基础之上的一种非线性分析方法

本文描述了一种新的，不依赖于小参数的非线性分析方法，并分析，比较了该方法与摄动展开方法的优缺点，本文所述方法建立在拓扑的同伦理论之上。彻底抛弃了摄动方法的小参数假设，从而可以求解更强的非线性问题，特别是那些不含小参数的非线性问题。     

双相介质波动方程孔隙率反演的同伦方法

从材料响应的理论合成应与实际测量数据相拟合这一出发点，将双相介质波劝方程参数的反演问题转化为非线性算子方程的零点求解问题，从而应用一种大范围收敛的同伦方尘土注来解非线性算子方程，并把这种方法用于Simon(1984)给出的具有解析的一维双相介质模型的数值模拟，最后的数值结果表明，给出的算法是十分有效的。     

Numerical Calculation of Effective Diffusion in Unsaturated Porous Media by the TRT Lattice Boltzmann Method

Numerical models that solve transport of pollutants at the macroscopic scale in unsaturated porous media need the effective diffusion dependence on saturation as an input. We conducted numerical computations at the pore scale in order to obtain the effective diffusion curve as a function of saturation for an academic sphere packing porous medium and for a real porous medium where pore structure knowledge was obtained through X-ray tomography. The computations were performed using a combination of lattice Boltzmann models based on two relaxation time (TRT) scheme. The first stage of the calculations consisted in recovering the water spatial distribution into the pore structure for several fixed saturations using a phase separation TRT lattice Boltzmann model. Then, we performed diffusion computation of a non-reactive solute in the connected water structure using a diffusion TRT lattice Boltzmann model. Finally, the effective diffusion for each selected saturation value was estimated through inversion of a macroscopic classical analytical solution.     

On the solution of MHD flow over a nonlinear stretching sheet by an efficient semi‐analytical technique

The problem of magneto‐hydrodynamic fluid flow past a nonlinear stretching sheet in the presence of a transverse magnetic field is analyzed. The governing equations are transformed into a nonlinear ordinary differential equation that is solved using a novel spectral homotopy analysis method and the Matlab in‐built numerical solverttbvp4c. The new technique removes some known limitations of the homotopy analysis method and offers a more systematic way of selecting initial approximations and the optimal auxiliary parameter ?. A comparison with the numerical solution confirms the robustness, the computational efficiency, and the accuracy of the technique. Copyright © 2011 John Wiley & Sons, Ltd.     

Parameterizing the Leaching Surface by Combining Curve-Fitting for Solute Breakthrough and for Spatial Solute Distribution

Multi-compartment samplers (MCSs) measure unsaturated solute transport in space and time at a given depth. Sorting the breakthrough curves (BTCs) for individual compartments in descending order of total solute amount and plotting in 3D produces the leaching surface. The leaching surface is a useful tool to organize, present, and analyze MCS data. We present a novel method to quantitatively characterize leaching surfaces. We fitted a mean pore-water velocity and a dispersion coefficient to each BTC, and then approximated their values by functions of the rank order of the BTCs. By combining the parameters of these functions with those of the Beta distribution fitted to the spatial distribution of solutes, we described an entire leaching surface by four to eight parameters. This direct characterization method allows trends to be subtracted from the observations, and incorporates the effects of local heterogeneity. The parametric fit creates the possibility to quantify concisely the leaching behavior of a soil in a given climate under given land use, and eases the quantitative comparison of spatio-temporal leaching behavior in different soils and climates.     

Solution of boundary layer flow and heat transfer of an electrically conducting micropolar fluid in a non-Darcian porous medium

In this paper, we have studied the effects of radiation on the boundary layer flow and heat transfer of an electrically conducting micropolar fluid over a continuously moving stretching surface embedded in a non-Darcian porous medium with a uniform magnetic field has been analyzed analytically. The governing fundamental equations are approximated by a system of nonlinear locally similar ordinary differential equations which are solved analytically by applying homotopy analysis method (HAM). The effects of Darcy number, heat generation parameter and inertia coefficient parameter are determined on the flow. Convergence of the obtained series solution is discussed. The homotopy analysis method provides us with a new way to obtain series solutions of such problems. This method contains the auxiliary parameter which provides us with a simple way to adjust and control the convergence region of series solution. By suitable choice of the auxiliary parameter, we can obtain reasonable solutions for large modulus.     

The numerical analysis of water and solute movement in scale heterogeneous profiles

A computer based numerical method is presented for the analysis of water and solute movement in unsaturated heterogeneous porous materials. Such a method is necessary since, for those field studies where solute movement is of concern, the soil profiles under consideration are invariably heterogeneous. The numerical analysis is based on a general one-dimensional finite difference soil water flow model which includes a numerical technique combining the concepts of scale heterogeneity with an interpolative soil water hysteresis model. An explicit finite difference solute movement subroutine is incorporated into the unsaturated flow model to describe the transport of nonreactive solutes. A velocity dependent longitudinal dispersion coefficient is used in the solution of the hydrodynamic dispersion equation. The resulting hysteretic scale heterogeneous solute movement model permits the study of solute dynamics during infiltrating and redistribution in realistically complex spatially varying soil profiles. Results are presented for the leaching of both coarse grading to fine and fine grading to coarse sand profiles. Both vertical and horizontal profiles are studied using either a constant flux or a constant concentration input boundary condition. The four cases studied demonstrate the versatility of the numerical method and emphasise the substantial differences in transport behavior that can arise between heterogeneous and homogeneous profiles.Now with BHP Petroleum Pty. Ltd., GPO Box 1911R, Melbourne, Vic. 3001, Australia.