Approximate importance sampling Monte Carlo for data assimilation

Importance sampling Monte Carlo offers powerful approaches to approximating Bayesian updating in sequential problems. Specific classes of such approaches are known as particle filters. These procedures rely on the simulation of samples or ensembles of the unknown quantities and the calculation of associated weights for the ensemble members. As time evolves and/or when applied in high-dimensional settings, such as those of interest in many data assimilation problems, these weights typically display undesirable features. The key difficulty involves a collapse toward approximate distributions concentrating virtually all of their probability on an implausibly few ensemble members.

After reviewing ensembling, Monte Carlo, importance sampling and particle filters, we present some approximations intended to moderate the problem of collapsing weights. The motivations for these suggestions are combinations of (i) the idea that key dynamical behavior in many systems actually takes place on a low dimensional manifold, and (ii) notions of statistical dimension reduction. We illustrate our suggestions in a problem of inference for ocean surface winds and atmospheric pressure. Real observational data are used.     

On rainbows of inhomogeneous spherical droplets

The paper is devoted to the study of the intensity distributions and the angular spectra of the second and fifth rainbows of homogeneous and inhomogeneous spherical particles predicted by Lorenz-Mie theory. The results show that the distribution around the second rainbow angle for a homogeneous sphere of refractive index between 1.32 and 1.33 is due to the interference of the light after two or five internal reflections. The structure of the scattering diagram and the angular spectrum for homogeneous and radially inhomogeneous spheres are studied. For a homogeneous sphere we show that the second and fifth rainbows can be independently reconstructed by filtering the calculated spectrum. Since each order of rainbow penetrates the particle to different depths, such methods could be used to provide information about the refractive index profile or the temperature gradient of an inhomogeneous sphere. The Airy-like peaks of the second and fifth rainbows are closely connected with the refractive index profile, which is beneficial to the measurement of its refractive index profile or temperature gradient.     

Use of elemental size distributions in identifying particle formation modes

The chemical composition of particles generated during pulverized coal combustion is the consequence of their formation processes. This work aims to use the size resolved elemental composition of coal-derived particles to identify their formation modes. A size-classified bituminous coal is burnt in a laboratory drop tube furnace at 1150, 1250, and 1350 °C, respectively. The elemental composition of the size-segregated particles from coal combustion is analyzed and the total mass fraction size distributions of Si and Al are obtained. Three particle formation modes are observed in these distribution profiles. The coarse mode has the highest value of the total mass fraction of Si and Al while the ultrafine mode has the lowest one. The total mass fraction of Si and Al in these two modes is nearly independent of particle size. It is believed that the coarse mode is formed by the mineral coalescence mechanism and the ultrafine mode by the vaporization–condensation mechanism. The difference in the total mass fraction of Si and Al between the central mode and the other two indicates that the central mode is formed by different mechanisms. Based on the observation that the total mass fraction of Si and Al in this mode increases with increasing particle size, heterogeneous condensation of vaporized species on existing fine residual ash particles is proposed to account for the formation of these particles. The study of the elemental composition of the three modes represented in five categories verifies the proposed formation mechanisms for them to some extent.     

PIV水下流场测试系统试验研究

基于粒子图像测速技术,设计研制了一套适用于水洞洞体内高速、封闭试验条件的流场测量与显示设备.该系统由光源、成像、图像处理与分析、控制和粒子投放等分系统组成，具有复杂环境下全流场的无接触、无扰动、高准确度测量和显示能力，流场测试误差小于2%，可为水下航行器外形设计及其改进提供相关的实测数据.     

聚变实验增殖堆FEB-E粒子抽除和抽气系统(英文)

ＦＥＢ－Ｅ的粒子抽除是通过偏滤器进行的。由４８ 个楔形气室模件环形组装而成的ＦＥＢ－Ｅ偏滤器，位于真空室的下部，与抽气系统和冷却系统相连。ＦＥＢ－Ｅ抽气系统有二个子系统：环粗抽系统和环高真空系统。环高真空系统是由一组处于真空室内１６ 个下部舱口内的低温泵和一组处于生物屏蔽层外的附加涡轮分子泵组成的。这些低温泵能提供的名义总抽速为５７６ｍ ３·ｓ－ １。在偏滤器高中性压力（＞ １Ｐａ）情况下，低温泵入口阀节流控制抽气粒子流。由于偏滤器抽气槽路以及偏滤器下侧通至真空室下部舱口的有限的通导能力，这些低温泵在偏滤器幽僻区域有效抽速为１６０ｍ ３·ｓ－ １。这意味着偏滤器幽僻区域的中性压力应在０．５－ １．０Ｐａ 范围内，以得到８０－ １６０Ｐａ·ｍ ３·ｓ－ １（在预期的偏滤器抽气槽路温度为４７３Ｋ时）范围内的抽气流量。低温泵每次在聚变实验增殖堆燃烧１０００ｓ 后受激运行     

Identification of ground water flow patterns using particle models

Particle models are often used to simulate transport processes in ground water. The ground water flow pattern is one of the driving parameters of the transport model. In this paper a parameter identification algorithm is developed for particle type models to identify the underlying flow pattern from concentration measurements. The estimation problem is solved with a gradient based algorithm. For each generated particle track, the adjoint track is determined to efficiently compute gradient of the criterion.     

落体法测g实验误差的研究

本对落体法测g实验误差进行了深入研究。使该实验精度大为提高。     

两性聚氨酯纳米水分散液的制备和性质

制备出了分子链上同时带有酸、碱基团的两性聚氨酯（APU），该两性聚氨酯在酸碱介质中均能形成稳定的纳米水分散液。同时还测定了APU的等电点和水分散液中APU分子链上可离子化基团的含量；实验结果还表明：中和剂浓度对两性聚氨酯纳米水分散液的粒径及稳定性有很大的影响。     

Particle creation in a time-dependent electric field revisited

We adopt the general formalism for analyzing evolution of gaussian states of quantized fields in time-dependent backgrounds in the Schrodinger picture (presented in detail in Mahajan and Padmanabhan [G. Mahajan, T. Padmanabhan, Gen. Rel. Grav. 40 (2008) 661]) to study the example of a spatially uniform electric field background (in a time-dependent gauge) which is kept turned on for a finite duration of time. In particular, we study the time-dependent particle content, defined in terms of the concept of instantaneous eigenstates, and describe how it captures the time evolution of the quantized field modes. The actual particle creation process occurs over a relatively short interval in time, and the particle content saturates rather quickly. We also compare the power spectrum of the field modes, computed in the asymptotic limit, with the corresponding situation in a cosmological de Sitter background. Particle creation under the influence of a spiked electric field localized in time, as a particular limiting case of the above general model, is also considered.