Modeling scale formation in flat‐sheet membrane modules during water desalination

Modeling the operation of spiral‐wound membrane modules is essential for their successful design and optimization. Such models must include the main types of membrane fouling, degrading desalination plant performance, including scaling due to sparingly soluble salts. Unfortunately, the complexity of underlying physicochemical processes and the coexistence of several spatial and temporal scales render intractable modeling of membrane scaling based on first principles. Therefore, a suitable (albeit simplified) framework is developed for incorporating scaling dynamics into a fluid flow model formulated at an intermediate (i.e., mesoscopic) length scale of membrane operation. The general mesoscopic approach involves integration of spatially distributed submodels, thereby allowing predictions at the large (entire membrane sheet) scale; these submodels comprise constitutive laws and kinetic rate expressions derived at fine scales. A submodel for the effect of pre‐existing bulk particles on scale formation is developed herein. Several numerical results are presented to exemplify the potential of the proposed framework. © 2013 American Institute of Chemical Engineers AIChE J, 59: 2917–2927, 2013     

Experimental Investigations on Condensation in the Framework of ENhanced COndensers in Microgravity (ENCOM-2) Project

We report preliminary results on experimental investigations on condensation in the framework of the European Space Agency funded programme Enhanced Condensers in Microgravity (ENCOM-2) which aims at better understanding underlying phenomena during condensation. The first experiment is a study on condensation of HFE on external curvilinear surface of 15 mm height during reduced gravity experiments. It is found that the local minimum of the film thickness exists at the conjugation area of condensed film and the meniscus at the bottom of the fin; this leads to the local maximum of the heat transfer coefficient, which we also found moves towards the fin tip. The second experiment is a study of falling films hydrodynamics inside a vertical long pipe. In particular, characteristics of wavy falling films produced employing intermittent liquid feed are examined in order to assess wave effects on film condensation. Preliminary results suggest that intermittent feed simply divides the film in two autonomous regions with the wave feature of each one depending only on its flow rate. The processing of registered film thickness data can lead to the estimation of the transverse velocity profile in the film, which is mainly responsible for heat transfer during condensation. The third experiment looks at in-tube convective condensation at low mass fluxes (typical of Loop Heat Pipes and Capillary Pumped Loops) of n-pentane inside a 0.56 mm diameter channel. The results show that the mean heat transfer in the annular zone when it is elongated may be less than the mean heat transfer when it is shorter, due to the interface deformation involved by surface tension effect. When the length of this annular zone reaches a critical value, the interface becomes unstable, and a liquid bridge forms, involving the release of a bubble. The heat transfer due to the phase-change in this isolated bubble zone appears to be very small compared to the sensible heat transfer: the bubbles evolve and collapse in a highly subcooled liquid. The last experiment concerns in-tube condensation of R134a inside a square channel of 1.23 mm hydraulic diameter at mass fluxes of 135 kg m^?2 s^?1 and 390 kg m^?2 s^?1 for three different configurations: horizontal, vertical downflow and vertical upflow. For the calculated heat transfer coefficient it is found that gravity has no effect on condensation in downflow configurations at 390 kg m^?2 s^?1 and in upflow conditions at both values of mass velocity. The effect of gravity on the condensation heat transfer coefficient becomes noteworthy in downflow at mass velocity G = 135 kg m^?2 s^?1 and vapour quality lower than 0.6.     

Long-term expression of human clotting factor IX from retrovirally transduced primary human keratinocytes in vivo

A persistent obstacle that has hampered gene transfer experiments is the short-term nature of transgene expression in vivo. In this article we present evidence for sustained expression from primary human keratinocytes, using the retroviral vector MFG. Primary keratinocytes were transduced in culture with the MFG retroviral vector containing the coding region from factor IX cDNA. Transduced keratinocytes, which secreted on average 830 ng of factor IX/10(6) cells/24 hr in tissue culture, were used to form a bilayered skin equivalent and grafted onto nude mice under a silicone transplantation chamber. Between 0.1 and 2.75 ng of human factor IX per milliliter was found in mouse plasma for more than 1 year, suggesting that keratinocyte stem cells were both transduced and grafted. The results show, for the first time, that long-term expression is obtainable in retrovirally transduced keratinocytes after transplantation.     

Multi-channel simulation of regeneration in honeycomb monolithic diesel particulate filters

In recent years advanced computational tools of diesel particulate filter (DPF) regeneration have been developed to assist in the systematic and cost-effective optimization of next generation particulate trap systems.In the present study, we employ a previously validated, state-of-the-art multichannel DPF simulator to study the regeneration process over the entire spatial domain of the filter. Particular attention is placed on identifying the effect of inlet cones and boundary conditions, filter can insulation and the dynamics of “hot spots” induced by localized external energy deposition. Lateral heat losses through the insulation and the periphery of the filter can, as captured by the magnitude of the Nusselt number, Nu, are detrimental to the effectiveness of the regeneration process. A filter can Nu number less than 10 and preferably less than 5 is a good design target for high regeneration efficiency. For the case studied, insulation of the inlet cones can lead to a gain of 30% in regeneration efficiency by eliminating radial temperature gradients at the inlet filter face. The multichannel simulator provides an instructive illustration of the well-appreciated effects of localized hot spot on filter regeneration: hot spots play a more significant role (spread over) when located near the entrance of the filter.     

Nephele: Scalable Access Control for Federated File Services

The integration of storage resources across different administrative domains can serve as building block for the development of efficient collaboration environments. In order to improve application portability across such environments, we target data sharing facilities that securely span multiple domains at the filesystem rather than the application level. We introduce the hypergroup as an heterogeneous two-layer construct, where the upper layer consists of administrative domains and the lower layer of users from each participating domain. We use public keys to uniquely identify users and domains, but rely on credentials to securely bind users and domains with hypergroups. Each domain is responsible for authenticating its local users across the federation, and employs access control lists to specify the rights of individual users and hypergroups over local storage resources. In comparison to existing systems, we show both analytically and experimentally reduced transfer cost of remote authorizations and improved scalability properties.     

Large wave characteristics and their downstream evolution at high Reynolds number falling films

There are many open questions regarding the evolution of waves, especially for the case of turbulent films. To resolve the complexity in modeling wavy turbulent films, more information needs to be derived from experimental data. On this account, a new way is proposed herein to analyze experimental film thickness traces, replacing the usual statistical analysis. Large waves are identified in experimental traces, and their shape is described by approximation with a few parameters curve. The probability density functions of these parameters are identified and the whole procedure can be considered as a compression method of the information content of experimental data series. By comparing results at several downstream locations, information on the evolution of waves along the flow is derived. This information indicates a 3D character of the flow, customary neglected in modeling efforts. In addition, the current results can be used for the numerical reconstruction of experimental film thickness traces. © 2009 American Institute of Chemical Engineers AIChE J, 2009     

Aspects of the Two-Layer Model for Direct Contact Condensation of Steam on Wavy Falling Films

Detailed physical modeling of the direct contact heat transfer process of steam condensing on a falling liquid film is a very difficult task due to the complex hydrodynamics of the film. The present state of the art is restricted to film Reynolds number of the order of 100. On the other hand, empirical relations cannot offer any insight into the mechanism and features of heat transfer in liquid film. Phenomenological models are needed to bridge the gap between empirical relations and direct physical simulations. One of these models is the so-called two-layer model, which divides the falling film into a laminar conduction-dominated (substrate) layer flowing over the solid wall and a completely mixed layer representing the waves. This model is further developed here by focusing specifically on two aspects. First, the influence of temperature-dependent physical properties of the liquid on the structure and heat transfer characteristics of the substrate layer is studied, and typical results are presented in the absence of waves. Second, the relation of the main parameter of the model (i.e., thickness of the substrate layer) to local film state is discussed in detail. Generalized constitutive laws and an approach based on utilization of the experimental film thickness time series are proposed and discussed. The proposed procedures can be integrated to a generalized two-layer model for direct contact condensation.     

Is naturalistic driving research possible with highly instrumented cars? Lessons learnt in three research centres

This paper provides an overview of the experiences using Highly Instrumented Cars (HICs) in three research Centres across Europe; Spain, the UK and Greece. The data collection capability of each car is described and an overview presented relating to the relationship between the level of instrumentation and the research possible. A discussion then follows which considers the advantages and disadvantages of using HICs for ND research. This includes the obtrusive nature of the data collection equipment, the cost of equipping the vehicles with sophisticated Data Acquisition Systems (DAS) and the challenges for data storage and analysis particularly with respect to video data. It is concluded that the use of HICs substantially increases the depth of knowledge relating to the driver's behaviour and their interaction with the vehicle and surroundings. With careful study design and integration into larger studies with Low(ly) instrumented Cars (LICs), HICs can contribute significantly and in a relatively naturalistic manner to the driver behaviour research.     

Efficient Markov Network Structure Discovery Using Independence Tests

We present two algorithms for learning the structure of a Markov network from data: GSMN* and GSIMN. Both algorithms use statistical independence tests to infer the structure by successively constraining the set of structures consistent with the results of these tests. Until very recently, algorithms for structure learning were based on maximum likelihood estimation, which has been proved to be NP-hard for Markov networks due to the difficulty of estimating the parameters of the network, needed for the computation of the data likelihood. The independence-based approach does not require the computation of the likelihood, and thus both GSMN* and GSIMN can compute the structure efficiently (as shown in our experiments). GSMN* is an adaptation of the Grow-Shrink algorithm of Margaritis and Thrun for learning the structure of Bayesian networks. GSIMN extends GSMN* by additionally exploiting Pearl's well-known properties of the conditional independence relation to infer novel independences from known ones, thus avoiding the performance of statistical tests to estimate them. To accomplish this efficiently GSIMN uses the Triangle theorem, also introduced in this work, which is a simplified version of the set of Markov axioms. Experimental comparisons on artificial and real-world data sets show GSIMN can yield significant savings with respect to GSMN*, while generating a Markov network with comparable or in some cases improved quality. We also compare GSIMN to a forward-chaining implementation, called GSIMN-FCH, that produces all possible conditional independences resulting from repeatedly applying Pearl's theorems on the known conditional independence tests. The results of this comparison show that GSIMN, by the sole use of the Triangle theorem, is nearly optimal in terms of the set of independences tests that it infers.     

DIRECT CONTACT CONDENSATION OF DILUTE STEAM/AIR MIXTURES ON WAVY FALLING FILMS

Condensation from air-steam mixtures on falling water layers is investigated experimentally and theoretically. The thin film flows on the inner surface of a 5cmi.d. vertical pipe. This film is wavy turbulent while the gas phase is kept saturated with steam. Experiments are conducted with the gas mixture effectively stagnant, compared with the fast moving liquid film. Measurements are also made under a mild vacuum applied on the gas phase. Heat transfer coefficients averaged over the entire length of the condensing surface, tend to increase by decreasing the liquid flowrate, by increasing the steam fraction, and by applying a mild vacuum on the gas phase. However, for the cases examined, there is a liquid flowrate above which the heat transfer coefficient becomes almost constant.

Numerical predictions are made for a fully developed turbulent film using an eddy diffusivity model. The results indicate that for a system with a large amount of noncondensable gases-as in this study-the temperature profile in the liquid film is nearly uniform and that the major resistance to condensation resides in the gas phase. The analysis also shows that the relative contribution of sensible heat transferred through the gas phase is small relative to the latent heat released upon condensation. Comparison of predictions with experimental data suggests that a significant parameter in these analyses is the gas diffusion boundary layer thickness which seems to be comparable in size with the liquid film thickness. Finally, the possibility is discussed of correlating condensation heat transfer coefficients with already available statistical characteristics of the falling wavy layer. Theoretical predictions based on this idea are in good agreement with data.