Information Preservation (IP) Method in Simulation

This paper reviews firstly methods for treating low speed rarefied gas flows: the linearised Boltzmann equation, the Lattice Boltzmann method (LBM), the Navier-Stokes equation plus slip boundary conditions and the DSMC method, and discusses the difficulties in simulating low speed transitional MEMS flows, especially the internal flows. In particular, the present version of the LBM is shown unfeasible for simulation of MEMS flow in transitional regime. The information preservation (IP) method overcomes the difficulty of the statistical simulation caused by the small information to noise ratio for low speed flows by preserving the average information of the enormous number of molecules a simulated molecule represents. A kind of validation of the method is given in this paper. The specificities of the internal flows in MEMS, i.e. the low speed and the large length to width ratio, result in the problem of elliptic nature of the necessity to regulate the inlet and outlet boundary conditions that influence each other. Through the example of the IP calculation of the microchannel (thousands long) flow it is shown that the adoption of the conservative scheme of the mass conservation equation and the super relaxation method resolves this problem successfully. With employment of the same measures the IP method solves the thin film air bearing problem in transitional regime for authentic hard disc write/read head length ( ) and provides pressure distribution in full agreement with the generalized Reynolds equation, while before this the DSMC check of the validity of the Reynolds equation was done only for short ( ) drive head. The author suggests degenerate the Reynolds equation to solve the microchannel flow problem in transitional regime, thus provides a means with merit of strict kinetic theory for testing various methods intending to treat the internal MEMS flows.

Information preservation (IP) method in simulation of internal rarefied gas flows in MEMS

This paper reviews firstly methods for treating low speed rarefied gas flows: the linearised Boltzmann equation, the Lattice Boltzmann method (LBM), the Navier-Stokes equation plus slip boundary conditions and the DSMC method, and discusses the difficulties in simulating low speed transitional MEMS flows, especially the internal flows. In particular, the present version of the LBM is shown unfeasible for simulation of MEMS flow in transitional regime. The information preservation (IP) method overcomes the difficulty of the statistical simulation caused by the small information to noise ratio for low speed flows by preserving the average information of the enormous number of molecules a simulated molecule represents. A kind of validation of the method is given in this paper. The specificities of the internal flows in MEMS, i.e. the low speed and the large length to width ratio, result in the problem of elliptic nature of the necessity to regulate the inlet and outlet boundary conditions that influence each other. Through the example of the IP calculation of the microchannel (thousands m ? long) flow it is shown that the adoption of the conservative scheme of the mass conservation equation and the super relaxation method resolves this problem successfully. With employment of the same measures the IP method solves the thin film air bearing problem in transitional regime for authentic hard disc write/read head length ( 1000 L m ? = ) and provides pressure distribution in full agreement with the generalized Reynolds equation, while before this the DSMC check of the validity of the Reynolds equation was done only for short ( 5 L m ? = ) drive head. The author suggests degenerate the Reynolds equation to solve the microchannel flow problem in transitional regime, thus provides a means with merit of strict kinetic theory for testing various methods intending to treat the internal MEMS flows.

IP Simulation of Micro Gas Flows under 3-D Head Sliders

Gas film lubrication of a three-dimensional flat read-write head slider is calculated using the information preservation (IP) method and the direct simulation Monte Carlo (DSMC) method, respectively. The pressure distributions on the head slider surface at different velocities and flying heights obtained by the two methods are in excellent agreement. IP method is also employed to deal with head slider with three-dimensional complex configuration. The pressure distribution on the head slider surface and the net lifting force obtained by the IP method also agree well with those of DSMC method. Much less (of the order about 10(2) less) computational time (the sum of the time used to reach a steady stage and the time used in sampling process) is needed by the IP method than the DSMC method and such an advantage is more remarkable as the gas velocity decreases.

IP方法模拟硬盘读写头气膜润滑问题

本论文中采用IP方法模拟硬盘读写头的气膜润滑问题，并选择耗时但精准的DSMC方法作为该问题中IP方法的检验标准，IP方法计算得到的读写头表面压力分布及读写头所受净升力为读写头的设计提供可靠的参考数据。 首先，将读写头复杂表面简化为平面并假定尺寸为微米（远小于真实的毫米），微米尺寸时可以得到DSMC的模拟结果用于比较。IP法对该问题的模拟结果与DSMC的模拟结果完全吻合。 其次，选定一个真实的读写头表面几何，但仍然假定尺寸为微米。IP法在该问题的应用过程中遇到四点困难，我们提出或采取了相应的解决办法。在复杂表面几何的读写头气膜润滑问题中，IP法的模拟结果与DSMC的模拟结果也相符一致。 最后完成真实几何、真实尺寸（毫米）的读写头气膜润滑问题的模拟，这也是本论文的最终研究目标。IP法模拟该问题时需要采用大网格，为此我们构造二维检验模型验证了大网格的合理性。 真实几何读写头气膜润滑问题的模拟结果表明，在流场的局部区域压力沿着高度方向变化显著，这意味着读写头问题的传统计算方法（即概括性雷诺方程）在该局部区域不成立，故用它计算得到的结果值得认真检验，但文献中尚未给出过这类检验。这正是该论文的研究背景和采用新方法（IP方法）的原因。 本论文研究的问题属于过渡领域、低速流动问题，故本论文还介绍了该领域另外一种粒子模拟方法（LVDSMC方法）以及我们对该方法提出的一些改进。 关键词：硬盘读写头，气膜润滑，概括性雷诺方程，IP方法，DSMC方法，过渡领域，LVDSMC方法

Efficient Methods for Robust Classification Under Uncertainty in Kernel Matrices

In this paper we study the problem of designing SVM classifiers when the kernel matrix, K, is affected by uncertainty. Specifically K is modeled as a positive affine combination of given positive semi definite kernels, with the coefficients ranging in a norm-bounded uncertainty set. We treat the problem using the Robust Optimization methodology. This reduces the uncertain SVM problem into a deterministic conic quadratic problem which can be solved in principle by a polynomial time Interior Point (IP) algorithm. However, for large-scale classification problems, IP methods become intractable and one has to resort to first-order gradient type methods. The strategy we use here is to reformulate the robust counterpart of the uncertain SVM problem as a saddle point problem and employ a special gradient scheme which works directly on the convex-concave saddle function. The algorithm is a simplified version of a general scheme due to Juditski and Nemirovski (2011). It achieves an O(1/T-2) reduction of the initial error after T iterations. A comprehensive empirical study on both synthetic data and real-world protein structure data sets show that the proposed formulations achieve the desired robustness, and the saddle point based algorithm outperforms the IP method significantly.

Statistical Simulation of Rarefied Gas Flows in Micro-Channels

Rarefied gas flows through micro-channels are simulated using particle approaches, named as the information preservation (IP) method and the direct simulation Monte Carlo (DSMC) method. In simulating the low speed flows in long micro-channels the DSMC method encounters the problem of large sample size demand and the difficulty of regulating boundary conditions at the inlet and outlet. Some important computational issues in the calculation of long micro-channel flows by using the IP method, such as the use the conservative form of the mass conservation equation to guarantee the adjustment of the inlet and outlet boundary conditions and the super-relaxation scheme to accelerate the convergence process, are addressed. Stream-wise pressure distributions and mass fluxes through micro-channels given by the IP method agree well with experimental data measured in long micro-channels by Pong et al. (with a height to length ratio of 1.2:3000), Shih et al. (l.2:4800), Arkilic et al. and Arkilic (l.3:7500), respectively. The famous Knudsen minimum of normalized mass flux is observed in IP and DSMC calculations of a short micro-channel over the entire flow regime from continuum to free molecular, whereas the slip Navier-Stokes solution fails to predict it.

Statistical Simulation of Thin-Film Bearings

The information preservation (IP) method and the direct simulation Monte Carlo (DSMC) method are used to simulate the gas flows between the write/read head and the platter of the disk drive (the slider bearing problem). The results of both methods are in good agreement with numerical solution of the Reynolds equation in the cases studied. However, the DSMC method owing to the problem of large sample size demand and the difficulty in regulating boundary conditions at the inlet and outlet was able to simulate only short bearings, while IP simulates the bearing of authentic length ~1000 m ? and can provide more detailed flow information.

Monte Carlo Modeling of Micro Scale Gas Flows

An information preservation (IP) method has been used to simulate many micro scale gas flows. It may efficiently reduce the statistical scatter inherent in conventional particle approaches such as the direct simulation Monte Carlo (DSMC) method. This paper reviews applications of IP to some benchmark problems. Comparison of the IP results with those given by experiment, DSMC, and the linearized Boltzmann equation, as well as the Navier-Stokes equations with a slip boundary condition, and the lattice Boltzmann equation, shows that the IP method is applicable to micro scale gas flows over the entire flow regime from continuum to free molecular.

Information Preservation Modeling of Rayleigh-Benard Transition from Thermal Conduction to Convection

Onset and evolution of the Rayleigh-Benard (R-B) convection are investigated using the Information Preservation (IP) method. The information velocity and temperature are updated using the Octant Flux Splitting (OFS) model developed by Masters & Ye based on the Maxwell transport equation suggested by Sun & Boyd. Statistical noise inherent in particle approaches such as the direct simulation Monte Carlo (DSMC) method is effectively reduced by the IP method, and therefore the evolutions from an initial quiescent fluid to a final steady state are shown clearly. An interesting phenomenon is observed: when the Rayleigh number (Ra) exceeds its critical value, there exists an obvious incubation stage. During the incubation stage, the vortex structure clearly appears and evolves, whereas the Nusselt number (Nu) of the lower plate is close to unity. After the incubation stage, the vortex velocity and Nu rapidly increase, and the flow field quickly reaches a steady, convective state. A relation of Nu to Ra given by IP agrees with those given by DSMC, the classical theory and experimental data.

Development of an efficient particle approach for micro-scale gas flow simulations

The micro-scale gas flows are usually low-speed flows and exhibit rarefied gas effects. It is challenging to simulate these flows because traditional CFD method is unable to capture the rarefied gas effects and the direct simulation Monte Carlo (DSMC) method is very inefficient for low-speed flows. In this study we combine two techniques to improve the efficiency of the DSMC method. The information preservation technique is used to reduce the statistical noise and the cell-size relaxed technique is employed to increase the effective cell size. The new cell-size relaxed IP method is found capable of simulating micro-scale gas flows as shown by the 2D lid-driven cavity flows.

Statistical analysis of microscale gas flows and stock price changes

First, recent studies on the information preservation (IP) method, a particle approach for low-speed micro-scale gas flows, are reviewed. The IP method was validated for benchmark issues such as Couette, Poiseuille and Rayleigh flows, compared well with measured data for typical internal flows through micro-channels and external flows past micro flat plates, and combined with the Navier-Stokes equations to be a hybrid scheme for subsonic, rarefied gas flows. Second, the focus is moved to the microscopic characteristic of China stock market, particularly the price correlation between stock deals. A very interesting phenomenon was found that showed a reverse transition behaviour between two neighbouring price changes. This behaviour significantly differs from the transition rules for atomic and molecular energy levels, and it is very helpful to understand the essential difference between stock markets and nature.

Kinetic Studies of Gas Flows

Our recent studies on kinetic behaviors of gas flows are reviewed in this paper. These flows have a wide range of background, but share a common feature that the flow Knudsen number is larger than 0.01. Thus kinetic approaches such as the direct simulation Monte Carlo method are required for their description. In the past few years, we studied several micro/nano-scale flows by developing novel particle simulation approach, and investigated the flows in low-pressure chambers and at high altitude. In addition, the microscopic behaviors of a couple of classical flow problems were analyzed, which shows the potential for kinetic approaches to reveal the microscopic mechanism of gas flows.

Étude des interactions protéiques impliquant NPM-MLF1 dans la leucémie myéloïde aiguë.

Différentes translocations génomiques sont fréquemment associées à l'apparition de leucémies myéloïdes aiguës (LMA). Ces translocations génomiques résultent de l’assemblage de deux gènes conduisant à la production d'une protéine de fusion. C'est le cas de la translocation t (3; 5) (q25.1; q34) impliquant le suppresseur tumoral NPM et l'oncogène MLF1 donnant naissance à la protéine de fusion NPM-MLF1. Généralement, les gènes impliqués dans ces translocations contrôlent la croissance cellulaire, la différenciation ou la survie cellulaire. Cependant, pour NPM-MLF1 les causes du gain ou de la perte de fonction associée à la translocation demeurent inconnues car nous ne savons pas comment cette translocation peut favoriser ou participer à l'avènement de la LMA. Le but de ce travail est d’analyser le rôle de NPM-MLF1 dans le cancer et d’examiner comment son activité contribue à la leucémie en faisant des études d’interactions protéine/protéine. En effet, l’étude de la fonction d’une protéine implique souvent de connaître ses partenaires d’interactions. Pour ce faire, la technique de double hybride dans la souche de levure AH109 a été utilisée. Tout d’abord, les ADN complémentaires (ADNc) de MLF1, NPM1 et de NPM-MLF1, MLF1-Like (une partie de MLF1 de l’acide aminé 94 à 157) normaux et mutés du domaine MTG8-Like constitué des acides aminés (a.a.) 151 à 164 de MLF1 (excepté NPM) ont été clonés dans un vecteur d'expression de levure pGBKT7. Les ADNc de GFI-1, mSin3A, PLZF, HDAC1 et HDAC3 ont été clonés dans le plasmide pGADT7 de façon à créer des protéines de fusion synthétiques avec le domaine de liaison à l'ADN et de trans-activation de la protéine GAL4. Le plasmide pGBKT7 possède un gène TRP1 et pGADT7 un gène LEU2 qui permettent la sélection des clones insérés dans la levure. Aussi, le pGBKT7 a un épitope c-myc et pGADT7 un épitope HA qui permet de voir l’expression des protéines par buvardage de type Western. Après la transformation des levures les interactions protéine/protéine ont été observées en vérifiant l’expression des gènes rapporteurs HIS3, LacZ, MEL1, ADE2 de la levure en utilisant des milieux de sélection YPD/-Leu/-Trp, YPD/-Leu/-Trp/-His, YPD/-Leu/-Trp/-His/-Ade, YPD/-Leu/-Trp/+ X-Gal, YPD/-Leu/-Trp/ + X-α-Gal. Ensuite, les interactions trouvées par double-hybride ont été vérifiées dans les cellules érythroleucémiques K562 par immuno-précipitation (IP) de protéines suivies de buvardages Westerns avec les anticorps appropriés. NPM-MLF1, MLF1, MTG8, MLF1-Like surexprimés dans les cellules K562 ont été clonés dans le plasmide pOZ-FH-N. pOZ-FH-N possède un récepteur IL-2 qui permet de sélectionner les cellules qui l’expriment ainsi qu’un tag Flag-HA qui permet de voir l’expression des protéines par buvardage-Western. Les résultats du double-hybride suggèrent une interaction faible de NPM-MLF1 avec HDAC1, HDAC3 et mSin3A ainsi qu’une interaction qui semble plus évidente entre NPM-MLF1 et PLZF, GFI-1. NPM interagit avec GFI-1 et mSin3A. Aussi, MLF1 et MLF1-Like interagissent avec HDAC1, HDAC3, GFI-1, PLZF mais pas avec mSin3A. Les IP suggèrent que NPM-MLF1 interagit avec HDAC1, HDAC3, mSin3A et PLZF. MLF1 et MLF1-Like interagissent avec HDAC1, HDAC3 et mSin3A. L’interaction de NPM-MLF1 avec GFI-1, MLF1 et MLF1-Like avec PLZF et GFI-1 n’a pas encore été vérifiée par IP. Ainsi, nos observations permettent de suggérer que NPM-MF1, MLF1 et NPM pourraient jouer un rôle dans la transcription et la régulation de l’expression de certains gènes importants dans l’hématopoïèse et une variété de processus cellulaires parce qu’ils interagissent avec différents corépresseurs. En déterminant les partenaires protéiques de MLF1, NPM et NPM-MLF1, leurs fonctions et comment NPM-MLF1 influence et modifie le fonctionnement cellulaire normal; il sera possible de renverser le processus de LMA favorisé par la t (3; 5) NPM-MLF1 par la technologie d’interférence à l’ARN.

Métodos geoelétricos empregados na caracterização e monitoramento de anomalias inerentes à infiltração controlada de vinhaça

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

On-chip immunoprecipitation for protein purification

Immunoprecipitation (IP) is one of the most widely used and selective techniques for protein purification. Here, a miniaturised, polymer-supported immunoprecipitation (µIP) method for the on-chip purification of proteins from complex mixtures is described. A 4 µl PDMS column functionalised with covalently bound antibodies was created and all critical aspects of the µIP protocol (antibody immobilisation, blocking of potential non-specific adsorption sites, sample incubation and washing conditions) were assessed and optimised. The optimised µIP method was used to obtain purified fractions of affinity-tagged protein from a bacterial lysate.