Search

International audience; Personal Data Management Systems (PDMSs) provide individuals with appropriate tools to collect, manage and share their personal data under control. A founding principle of PDMSs is to move the computation code to the user's data, not the other way around. This opens up new uses for personal data, wherein the entire personal database of the individuals is operated within their local environment and never exposed outside, but only aggregated computed results are externalized. Yet, whenever arbitrary aggregation function code, provided by a third-party service or application, is evaluated on large datasets, as envisioned for typical PDMS use-cases, can the potential leakage of the user's personal information, through the legitimate results of that function, be bounded and kept small? This paper aims at providing a positive answer to this question, which is essential to demonstrate the rationale of the PDMS paradigm. We resort to an architecture for PDMSs based on Trusted Execution Environments to evaluate any classical user-defined aggregate PDMS function. We show that an upper bound on leakage exists and we sketch remaining research issues.

Sa sve većim porastom računarstva u oblaku raste i potreba za sigurnim skladištenjem podataka. Postoje mnoga rješenja koja pokušavaju osigurati povjerljivost podataka u oblaku, a jedno od njih je CryptDB. CryptDB je programsko rješenje koje pruža dokazivu i praktičnu privatnost u slučaju kompromitiranog servera baze podataka ili znatiželjnih administratora. Za to koristi šifriranu bazu, posredničkog poslužitelja i korisnički definirane funkcije. U ovom radu obrađena su rješenja koja pokušavaju rješiti isti problem. Zatim je detaljno objašnjen CryptDB i njegove funkcionalnosti, kao i kriptografski algoritmi korišteni za njegovu implementaciju. Za praktični dio rada implementirano je rješenje po uzoru na CryptDB, te su objašnjene glavne značajke rješenja, kao i razlike u odnosu na CryptDB. The constant growth of cloud computing services is also the reason for rising need for safety of the data that is resting there. There are many solutions that try to provide confidentiality of the data in the cloud, one of them being CryptDB. CryptDB is a software solution that gives provides provable and practical privacy in the face of a compromised database server or curious database administrators. The encrypted database, proxy server and user defined functions are characteristics of the solution. This thesis covers related work, then goes in depth on features and cryptographic algorithms in CryptDB. In the practical part solution is implemented inspired by CryptDB. Its main features are explained, as well as main differences with CryptDB.

A computational 2-D Eulerian-Eulerian approach was developed to simulate the hydrodynamics and heat transfer of a biomass gasification process in a pilot-scale bubbling fluidized bed reactor. The mathematical model was validated under experimental results collected from fluidization curves gathered at different temperatures in a pilot-scale reactor (75 kWth). Own user defined functions (UDFs) were developed in C programming and included to improve drag and heat transfer phenomena, as well to minimize deviations between experimental and numerical data found in previous works. Mesh selection was achieved by comparing solid fraction and pressure drop contours with grids comprised of different number of cells. A comparative study for particle diameter and inlet gas velocity was conducted for three different biomass feedstocks’ and their impact in the mixing and segregation index was studied. Mixing and segregation index were measured by implementing the standard deviation concept. Results indicated that UDFs significantly improved the mathematical model predictions on the reactor’s fluidization curves. Biomass and sand particles size and density showed direct influence on the solids distribution along the bed height. Smaller biomass particles revealed faster heat conduction and improved mixing properties. IF/01772/ 2014 info:eu-repo/semantics/publishedVersion

In this study, for a near-room-temperature magnetic cooling system, a decoupled multi-physics numerical approach (Magnetism, Fluid Flow, and Heat Transfer) is developed using a commercial CFD solver, ANSYS-FLUENT, as a design tool. User defined functions are incorporated into the software in order to take into account the magnetocaloric effect. Magnetic flux density is assumed to be linear during the magnetization and demagnetization processes. Furthermore, the minimum and maximum magnetic flux densities (Bmin and Bmax) are defined as 0.27 and 0.98, respectively. Two different sets of analyses are conducted by assuming an insulated cold heat exchanger (CHEX) and by defining an artificial cooling load in the CHEX. As a validation case, experimental work from the literature is reproduced numerically, and the results show that the current methodology is fairly accurate. Moreover, parametric analyses are conducted to investigate the effect of the velocity of heat transfer fluid (HTF) and types of HTF on the performance of the magnetic cooling system. Also, the performance metrics of the magnetic cooling system are investigated with regards to the temperature span of the magnetic cooling unit, and the cooling load. It is concluded that reducing the cycle duration ensures reaching lower temperature values. Similarly, reducing the velocity of the HTF allows reducing the outlet temperature of the HTF. In the current system, the highest temperature spans are obtained numerically as around 6 K, 5.2 K and 4.1 K for the cycle durations of 4.2 s, 6.2 s and 8.2 s, respectively., Scientific and Technological Research Council of Turkey (TUBITAK 114M829)

The paper reports about a pilot study that gives a numerical solution to the solid waste accumulation problem (SWAP). The purpose is to show both a simple and effective way to implement the theory using the technology of the Spatial DataBase Management Systems (SDBMSs), and the versatility of the proposed solution from the point of view of those responsible for the MSW management who, in fact, are offered a dual-mode display of the results: one tabular (typical of relational databases) and the other based on geographical maps, the latter particularly useful to highlight the spatial component of the data of the SWAP.

The thesis deals with connecting SAP ERP system via local database system MS SQL Server using the tools SAP BI, data synchronization between systems and advanced usage of T-SQL language for preparing data for web applications and reports written in PHP. The thesis contains a brief overview of the SAP system and the possibility of connecting to the SAP system. The general principles of described solution can be used in conjunction with other systems and programming languages.

[EN] This paper describes the methodology followed to perform a co-simulation between 1D (OpenWAM) and 3D (FLUENT) CFD codes. The Method of Characteristics (MoC) has been chosen to transfer the information between the two domains by properly updating the boundary condition at the shared interface. A short explanation of the MoC is provided, including the modifications needed by the Riemann invariants when dealing with non-homentropic flow. The implementation of the coupling is explained, focusing on the particular approach required by FLUENT in order to obtain the Riemann invariants. Two validation tests have been performed. The Sod's problem has been used to test the numerical accuracy of the coupling methodology. On the other hand, an impulse test rig configuration has been simulated to show the potential capability of a co-simulation in terms of reducing the computational cost. In both cases a good agreement in the solution is found. © 2010 Elsevier Ltd., Roberto Navarro is indebted to the Universidad Politecnica de Valencia for its support through Grant Beca de Excelencia 2009. PAID-09-09.

A three dimensional, steady state, non-isothermal, single phase model was developed and simulations were carried out in order to find the effect of cathode channel dimensions (width, depth and height) on the performance of an air-breathing fuel cell. The model was solved using commercial CFD package Fluent (version 6.3). Separate user defined functions were written to solve the electrochemical equations and the water transport through the membrane along with the other governing equations. Analyses were carried out for three different channel widths (2, 4 and 6�mm), for three different channel depths (2, 6 and 10�mm) and for three different cell heights (15, 30 and 45�mm). Cell characteristics like current distribution, species distribution, oxygen mass transfer coefficient, cell temperature, cathode channel velocities and net water transport coefficients are reported. The results show that the cell performance improves with increase in cathode channel width, channel depth and with decrease in cell height. Maximum power density obtained was 240�mW/cm2 for a channel width of 4�mm and channel depth of 6�mm. When the channel depth was 2�mm the performance was limited mainly due to the resistance offered by the channel for the buoyancy induced flow. For channel depths higher than 2�mm, the diffusion resistance of the porous GDL also contributed significantly to limit the performance to low current densities. At low current densities the fuel cell is prone to flooding whereas at high current densities ohmic overpotential due to dehydration of the membrane significantly contributes to the overall voltage loss. � 2009 Elsevier Masson SAS. All rights reserved.

In a tangentially coal-fired boiler, for locations inside and near the combustor, heat-transfer by radiation is significant, and hence, ash particles arrive in molten state. The aim of the present study is to adopt a mechanistic modeling approach which incorporates energy-conservation principles to address slag-layer growth. In order to determine the outcome of molten ash impaction, a mechanistic bouncing potential model, incorporating the phenomenon of recoiling of molten ash droplets after impaction, is employed. The bouncing potential is a representation of the excess energy possessed by the recoiling splat, and is used to determine the outcome of molten ash impaction - to stick or to bounce. Computational fluid dynamics techniques, incorporating the effect of thermophoresis, are adopted to estimate the arrival rate of ash particles, and the bouncing potential model, as a user-defined function, is incorporated in the simulation package to determine the status of the droplets after impaction. Two coals of Indian origin are simulated for slag-layer growth for a period of 100min. The simulation results, when compared with field data provided by BHEL-Trichy, indicate that the model qualitatively predicts the growth of slag-layers. It has been further inferred that smaller particles dominate deposit formation and its growth. � 2014 Elsevier Ltd.

Concentration polarization (CP) is a critical issue during desalination by cross-flow reverse osmosis (CFRO) filtration system and it is known that shear rate on the membrane surface can help its mitigation. Roto-dynamic filtration systems provide an easy way to generate high shear rate on membrane surface, and hence reducing CP. However the studies related to the growth of CP layer in roto-dynamic CFRO filtration is very limited. In this paper we developed a computational frame-work where flow and solute transport were simulated using ANSYS-Fluent V14.5 but physics of reverse osmosis through the membrane were implemented using "user defined functions" (UDFs). The development of the CP layer in a roto-dynamic CFRO filtration system has been discussed for different operating conditions such as inlet location, rotation speed, gap between the disk and the membrane and feed pressure. The effects of these parameters on permeate flux are also discussed. It is noted that angular speed of rotor and feed pressure are very effective parameters than others to enhance the permeate discharge. The understanding about the CP layer growth and its influence on permeate flux from the present study will help in better and efficient design of roto-dynamic CFRO filtration systems. � 2015 Elsevier B.V.

Part 4: Pattern Recognition and Image Processing; International audience; In this paper the possibilities for construction of an ad hoc search system to examine large-sized raster image data sets, e.g. rock images or medical images, for analysis of its characteristic parameters are presented. A new solution for image exploration based on any attributes extracted with computer image analysis by using extensions for server-side operations is proposed.

A new combustor concept referred as the trapped vortex combustor (TVC) employs a vortex that is trapped inside a cavity to stabilise the flame. The cavity is formed between two axisymmetric disks mounted in tandem. TVC offers many advantages when compared to conventional swirl stabilisers. In the present work, numerical investigation of cold flow (non-reacting) through trapped vortex combustor is performed. The numerical simulation involves passive flow through TVC to obtain an optimum cavity size to trap stable vortices inside the cavity and to observe the important characteristics of TVC. One of the main objectives is to evaluate various two equation turbulence models for the aerodynamic predictions of TVC. Commercial CFD software Fluent is used for the present study. In addition to many models available, Non-linear k-ω and modified k-ω models are incorporated through user defined functions. Results obtained include streamlines, residence time and entrainments for all models. The reattachment length obtained by non-linear k-ω model closely matches with that obtained by DNS in the case of forebody-spindle alone. Non-linear k-ω model alone captures the corner vortices while all the other models failed to capture. From the entrainment characteristics study, it is inferred that the primary air needs to be injected for accommodating the decrease in oxidizer inside the cavity to obtain better performance from the TVC.