Destruction of carbon tetrachloride in the presence of hydrogen-supplying compounds with ionisation and catalytic oxidation: Part 2. Methane as hydrogen font

The performance of methane (CH₄) as hydrogen donor molecule for the deep and clean oxidation of tetracloromethane (CCl₄) has been experimented. The oxidation of CCl₄–CH₄ admixtures were performed by catalytic combustion with and without pre-catalytic actions (i.e. ionisation/ozonisation) over manganese oxide based catalyst. The extent of CCl₄ and CH₄ combustion was controlled by quantitative determinations of both CO₂ and HCl formed and of CCl₄ and CH₄ converted. Experiments were performed in the 100–500 °C reaction interval in a continuous reaction line at laboratory scale. Fixed CCl₄ concentration (2000 ppmv) with/without CH₄ addition to air constituted the feed. Methane was added to the feedstream in variable concentration (1000–3000 ppmv), the obtained organic mixtures had H/Cl atomic ratios (R_H/Cl) from 0.5 to 1.5. Comparing the tests of ionisation/ozonisation (I/O) of pure CCl₄ streams and CCl₄–CH₄ admixtures, the efficiency of the hydrogen donor role of CH₄ was evidenced. The residue CH₄ concentration was not affected by the residence time (14–42 s) in the ionisation reactor while the residue concentration of CCl₄ decreased by increasing it. A first-order rate equation in CCl₄ concentration was found to represent the ionisation process both in the absence and in the presence of hydrocarbon. On the manganese oxide catalyst, CCl₄ catalytic combustion (CC) was not remarkably influenced by CH₄ presence. CCl₄ destruction by catalytic combustion assisted with ionisation (IOCC) gave rise to higher HCl yield in comparison with conventional combustion. The HCl formation increased with increasing the R_H/Cl value both in the I/O and IOCC experiments, while the CO₂ formation lightly decreased for high R_H/Cl in the CC and IOCC experiments due to incomplete oxidation of CH₄.     

Energy savings in chemical plants: a vinyl chloride case history

To optimise the energy utilisation in the processes (and in particular in the chemical industry) it is useful to identify (when possible) process configurations in which the driving forces are uniform and the entropy generation approaches zero when the size of the plant increases. Such ideal configurations can be used as references and, to obtain practical plant flow diagrams, they have to be modified, according to the peculiar features of the process and the plant, the economical factors and the safety and environmental requirements. This paper presents, after a short methodological introduction, a vinyl chloride plant optimisation, based on such concepts and realised in the Porto Torres plant by the European Vinyl Corporation. Vinyl chloride monomer (VCM) is obtained by ethylene dichloride (EDC) cracking; furnace effluents containing VCM, EDC, hydrogen chloride (HCl) and chlorinated byproducts are processed through quench towers and a distillation train in order to separate the various components. The key of the new technology is the distillation sequence, which is adjusted to make it similar to the thermodynamically optimal configuration. Due to the new temperature levels of reboilers and condensers, a better heat cascading and a reduction of the steam consumption are obtained. The modification of the existing plant with a nominal capacity of 150,000 t/a has been realised using mainly the existing columns and exchangers; a HCl distillation column and three heat exchangers have been added; a new screw compressor allows the use of the existing cooling system. In a 10-day test run carried out at 50% capacity after the modification of the plant a reduction of the steam consumption of 50% was achieved.     

Weakly supervised learning of interactions between humans and objects

We introduce a weakly supervised approach for learning human actions modeled as interactions between humans and objects. Our approach is human-centric: We first localize a human in the image and then determine the object relevant for the action and its spatial relation with the human. The model is learned automatically from a set of still images annotated only with the action label. Our approach relies on a human detector to initialize the model learning. For robustness to various degrees of visibility, we build a detector that learns to combine a set of existing part detectors. Starting from humans detected in a set of images depicting the action, our approach determines the action object and its spatial relation to the human. Its final output is a probabilistic model of the human-object interaction, i.e., the spatial relation between the human and the object. We present an extensive experimental evaluation on the sports action data set from [1], the PASCAL Action 2010 data set [2], and a new human-object interaction data set.     

Weakly Supervised Localization and Learning with Generic Knowledge

Learning a new object class from cluttered training images is very challenging when the location of object instances is unknown, i.e. in a weakly supervised setting. Many previous works require objects covering a large portion of the images. We present a novel approach that can cope with extensive clutter as well as large scale and appearance variations between object instances. To make this possible we exploit generic knowledge learned beforehand from images of other classes for which location annotation is available. Generic knowledge facilitates learning any new class from weakly supervised images, because it reduces the uncertainty in the location of its object instances. We propose a conditional random field that starts from generic knowledge and then progressively adapts to the new class. Our approach simultaneously localizes object instances while learning an appearance model specific for the class. We demonstrate this on several datasets, including the very challenging Pascal VOC 2007. Furthermore, our method allows training any state-of-the-art object detector in a weakly supervised fashion, although it would normally require object location annotations.     

Remote sensing of water quality in an Australian tropical freshwater impoundment using matrix inversion and MERIS images

The purpose of this study was to investigate how semi-analytical inversion techniques developed for the remote sensing of water quality parameters (chlorophyll a, tripton and coloured dissolved organic matter (CDOM)) in inland waters could be adapted or improved for application to Australian tropical and sub-tropical water bodies. The Matrix Inversion Method (MIM) with a semi-analytic model of the anisotropy of the in-water light field was applied to MERIS images of Burdekin Falls Dam, Australia, a tropical freshwater impoundment. Specific attention was required to improve the atmospheric correction of the MERIS data. The performance of the conventional three band exact solution of the MIM was compared to that of over-determined solutions that used constant and differential weighting for each sensor band.The results of the application of the MIM algorithm showed that the best weighting scheme had a mean chlorophyll a retrieval difference of 1.0 μgl^− 1, the three band direct matrix inversion scheme had a mean difference of 4.2 μgl^− 1 and the constant weight scheme had a mean difference of 5.5 μgl^− 1. For tripton, the best performed weighting scheme had a mean difference of 1.2 mgl^− 1, the three band scheme had a mean difference of 3.4 mgl^− 1 and the constant weight scheme had a mean difference of 1.8 mgl^− 1. For the CDOM retrieval, the mean difference was found to be 0.12 m^− 1 for the best performed weighting scheme, 0.25 m^− 1 for the three band scheme and 0.52 m^− 1 for the constant weight scheme. It was found that significant improvements in the accuracy and precision of retrieved water quality parameter values can be obtained by using differentially weighted, over-determined systems of equations, rather than exact solutions. These more reliable estimates of water quality parameters will allow water resource managers to improve their monitoring regimes.     

Graphical Congestion Games

We consider congestion games with linear latency functions in which each player is aware only of a subset of all the other players. This is modeled by means of a social knowledge graph G in which nodes represent players and there is an edge from i to j if i knows j. Under the assumption that the payoff of each player is affected only by the strategies of the adjacent ones, we first give a complete characterization of the games possessing pure Nash equilibria. Namely, if the social graph G is undirected, the game is an exact potential game and thus isomorphic to a classical congestion game. As a consequence, it always converges and possesses Nash equilibria. On the other hand, if G is directed an equilibrium is not guaranteed to exist, but the game is always convergent and an equilibrium can be found in polynomial time if G is acyclic, even if finding the best equilibrium remains an intractable problem.     

ImageNet Auto-Annotation with Segmentation Propagation

ImageNet is a large-scale hierarchical database of object classes with millions of images.We propose to automatically populate it with pixelwise object-background segmentations, by leveraging existing manual annotations in the form of class labels and bounding-boxes. The key idea is to recursively exploit images segmented so far to guide the segmentation of new images. At each stage this propagation process expands into the images which are easiest to segment at that point in time, e.g. by moving to the semantically most related classes to those segmented so far. The propagation of segmentation occurs both (a) at the image level, by transferring existing segmentations to estimate the probability of a pixel to be foreground, and (b) at the class level, by jointly segmenting images of the same class and by importing the appearance models of classes that are already segmented. Through experiments on 577 classes and 500k images we show that our technique (i) annotates a wide range of classes with accurate segmentations; (ii) effectively exploits the hierarchical structure of ImageNet; (iii) scales efficiently, especially when implemented on superpixels; (iv) outperforms a baseline GrabCut (Rother et al. 2004) initialized on the image center, as well as segmentation transfer from a fixed source pool and run independently on each target image (Kuettel and Ferrari 2012). Moreover, our method also delivers state-of-the-art results on the recent iCoseg dataset for co-segmentation.     

PET synthesis in the presence of lanthanide catalysts

Syntheses of poly(ethylene terephthalate), PET, by transesterification of dimethyl terephthalate with ethylene glycol in the presence of well-known catalysts and various lanthanide compounds were performed. Lanthanide catalysts appeared to be much more efficient in the first stage of the process (transesterification in the presence of an excess of ethylene glycol), and less active in polycondensation. PET produced with lanthanides was found to possess enhanced thermal and hydrolytic stability as compared to PET synthesized with well-known catalysts and commercial PET. © 1995 John Wiley & Sons, Inc.