50175 research outputs found
Sort by
Aggregation-induced enhanced photoluminescence in magnetic graphene oxide quantum dots as a fluorescence probe for As(iii) sensing
No full textClean drinking water is a basic need of human beings around the world. In the past few decades, many research groups have been trying different methods to exploit quantum dots (QDs) to decontaminate water but, due to either cost or cumbersome steps, they have been unsuccessful. We have tried exploiting graphene quantum dots (GQDs) for heavy metal ion detection, particularly for arsenic ions in contaminated water. Herein, we prepared highly sensitive and selective magnetic GQD (Fe-GQDs) based sensors for the turn on sensing of As3+ ions. Systematic characterization of the prepared magnetic GQDs was done using different spectroscopic techniques. Impressively, Fe-GQDs exhibited good selectivity for As3+ ions over a wide pH range. This turn on sensing of As3+ ions can be explained by the fact that Fe-GQDs aggregates were formed upon the addition of As3+ ions, which restricted the intra-molecular vibrational motion and consequently made the entire system emissive. A combination of techniques, such as time-correlated single-photon counting (TCSPC) experiments, Raman spectroscopy, XPS, DLS and zeta potential analysis, was employed to understand the aggregation-induced enhanced emission (AIEE) mechanism. The resultant limit of detection (LOD) value for as-prepared fluorescent Fe-GQDs was 5.1 ppb, which is well below the permissible limit for arsenic in drinking water. A reactive oxygen species (ROS) study in the presence and absence of UV light revealed the potential of the proposed quantum dots to act as an antibacterial agent. Overall, the low LOD value, the selectivity of Fe-GQDs over a wide pH range and the ROS study render our work of practical significance
Nitrogen doping as a fundamental way to enhance the EMI shielding behavior of cobalt particle-embedded carbonaceous nanostructures
No full textThe influence of nitrogen doping in pyrolysis-derived carbonaceous nanostructures with embedded Co-nanoparticles (Co@C) for electromagnetic (EM) absorption at microwave frequencies is explored. The synthesized Co-nanoparticles were found to be encapsulated by a graphitic carbon layer forming core-shell nanostructures. Interestingly, we observed that nitrogen (N) doping helps in the formation of smaller sized Co-nanoparticles embedded in the carbonaceous matrix along with a plethora of defects in the carbon layer of the Co@C sample. These defects in the carbon layer help to enhance the scattering of microwave radiation. We demonstrate that the scattering of EM waves, due to the presence of these defects, is advantageous in electromagnetic interference shielding. On the other hand, the smaller Co-nanoparticles predominantly acquire the highly magnetic hcp-phase, which helps in enhancing the EMI shielding through absorption of microwaves. Analysis of the complex permittivity and permeability suggests the enhancement of scattering at the defects and subsequent absorption of microwaves through the dispersed metallic Co nanoparticles and by the conducting graphitic C layer. The value of the shielding effectiveness was enhanced from similar to-24 dB for the (undoped) Co@C sample to similar to-33 dB for the N-doped Co@CN sample. Moreover, the shielding effectiveness due to absorption of microwaves also was found to be enhanced drastically. Hence, our results demonstrate that the effective EM shielding can be enhanced through the enhancement of microwave absorption by creating a defect-rich carbon framework via N-doping
On the effect of Re addition on microstructural evolution of a CoNi-based superalloy
No full textIn this study, the effect of rhenium (Re) addition on microstructural evolution of a new low-density Co-Ni-Al-Mo-Nb based superalloy is presented. Addition of Re significantly influences the gamma' precipitate morphology, the gamma/gamma' lattice misfit and the gamma/gamma' microstructural stability during long term aging. An addition of 2 at.% Re to a Co-30Ni-10Al-5Mo-2Nb (all in at.%) alloy, aged at 900 degrees C for 50 h, reduces the gamma/gamma' lattice misfit by similar to 40% (from +0.32% to +0.19%, measured at room temperature) and hence alters the gamma' morphology from cuboidal to round-cornered cuboidal precipitates. The composition profiles across the gamma/gamma' interface by atom probe tomography (APT) reveal Re partitions to the gamma phase (K-Re = 0.34) and also results in the partitioning reversal of Mo to the gamma phase (K-Mo = 0.90) from the gamma' precipitate. An inhomogeneous distribution of Gibbsian interfacial excess for the solute Re (Gamma(Re), ranging from 0.8 to 9.6 atom.nm(-2)) has been observed at the gamma/gamma' interface. A coarsening study at 900 degrees C (up to 1000 h) suggests that the coarsening of gamma' precipitates occurs solely by evaporation-condensation (EC) mechanism. This is contrary to that observed in the Co-30Ni-10Al-5Mo-2Nb alloy as well as in some of the Ni-Al based and high mass density Co-Al-W based superalloys, where gamma' precipitates coarsen by coagulation/coalescence mechanism with an extensive alignment of gamma' along <100> directions as a sign of microstructural instability. The gamma' coarsening rate constant (K-r) and gamma/gamma' interfacial energy are estimated to be 1.13 x 10(-27) m(3)/s and 8.4 mJ/m(2), which are comparable and lower than Co-Al-W based superalloys
Balancing Stragglers Against Staleness in Distributed Deep Learning
No full textSynchronous SGD is frequently the algorithm of choice for training deep learning models on compute clusters within reasonable time frames. However, even if a large number of workers (CPUs or GPUs) are at disposal for training, hetero-geneity of compute nodes and unreliability of the interconnecting network frequently pose a bottleneck to the training speed. Since the workers have to wait for each other at every model update step, even a single straggler/slow worker can derail the whole training performance. In this paper, we propose a novel approach to mitigate the straggler problem in large compute clusters. We cluster the compute nodes into multiple groups where each group updates the model synchronously stored in its own parameter server. The parameter servers of the different groups update the model in a central parameter server in an asynchronous manner. Few stragglers in the same group (or even separate groups) have little effect on the computational performance. The staleness of the asynchronous updates can be controlled by limiting the number of groups. Our method, in essence, provides a mechanism to move seamlessly between a pure synchronous and a pure asynchronous setting, thereby balancing between the computational overhead of synchronous SGD and the accuracy degradation of a pure asynchronous SGD. We empirically show that with increasing delay from straggler nodes (more than 300 delay in a node), progressive grouping of available workers still finishes the training within 20 of the no-delay case, with the limit to the number of groups governed by the permissible degradation in accuracy (� 2.5 compared to the no-delay case). © 2018 IEEE
Direct estimation and experimental validation of the acoustic source characteristics of turbocharged diesel engine exhaust system
No full textThe present study aims at application of the novel direct method developed by Kumar and Munjal for naturally aspirated engines Applied Acoustics 135, 70-84 (2018)] to the turbocharged engine's exhaust system for estimation of acoustic source characteristics. In this paper, transfer matrices of the turbine and compressor of an automotive turbocharger are derived. Mass continuity across the turbine and compressor along with the respective static pressure ratios are used for deriving their transfer matrices. The acoustic source characteristics of each cylinder for the source-load junction downstream of the exhaust valve/s is computed assuming that the cylinder discharges to the ideal pressure release boundary condition with the constant pressure being product of the turbine pressure ratio and mean pressure in the exhaust pipe downstream of the turbine. Source characteristics of each cylinder are used to estimate the source characteristics downstream of the exhaust manifold. Using the derived transfer matrix of the turbine, the source characteristics are further estimated at the source-load junction downstream of the turbine. Finally, the estimated source characteristics of the engine downstream of the turbine are used to predict the unmuffled sound pressure level (SPL) spectrum which is shown to compare reasonably well with the experimentally measured values. Importance of acoustic modelling of the turbine is highlighted by comparing the SPL spectra predicted using the estimated source characteristics with and without the turbine. The estimated source characteristics and unmuffled approximate SPL spectrum of a turbocharged engine being thus known at the engine's inception stage, the present study can be used for the integration of design and analysis of the engine and muffler. (C) 2019 Elsevier Ltd. All rights reserved
A Unified Controller for Utility-Interactive Uninterruptible Power Converters for Grid Connected and Autonomous Operations
No full textThis paper proposes a unified control based utility-interactive uninterruptible power converter (UIUPC). The proposed control enables the UIUPC to inherently transfer from a p - q control strategy in grid-connected mode to a voltage-frequency control strategy in autonomous mode and vice-versa. This happens seamlessly with the same control topology. The need to switch between two separate control architectures is thus eliminated. Critical islanding detection and synchronizing mechanisms are also not needed in the proposed control method. The unified controller proposed in this paper is derived from the concept of controlling the perturbations in the magnitude and speed of the point of common coupling (PCC) space vector. The PCC space vector is compared with a reference space vector applied in the direction of the PCC space vector on an instantaneous basis to extract the perturbations. The reference system of the proposed control is linked to the dynamics of the PCC space vector. This offers the ability of parallel operation of similar unified control UIUPC. Thus, in addition to the above said advantages, the proposed control also puts forth an alternative to paralleling approaches like conventional droop and master-slave configurations. The effectiveness of the proposed control is validated by simulation and experimentation
Factorization theorems for classical group characters, with applications to alternating sign matrices and plane partitions
No full textWe show that, for a certain class of partitions and an even number of variables of which half are reciprocals of the other half, Schur polynomials can be factorized into products of odd and even orthogonal characters. We also obtain related factorizations involving sums of two Schur polynomials, and certain odd-sized sets of variables. Our results generalize the factorization identities proved by Ciucu and Krattenthaler (2009) 14] for partitions of rectangular shape. We observe that if, in some of the results, the partitions are taken to have rectangular or double-staircase shapes and all of the variables are set to 1, then factorization identities for numbers of certain plane partitions, alternating sign matrices and related combinatorial objects are obtained. (C) 2019 Elsevier Inc. All rights reserved
Discovering structural similarities among rgas in Indian Art Music: a computational approach
No full textIndian Art Music has a huge variety of rgas. The similarity across rgas has traditionally been approached from various musicological viewpoints. This work aims at discovering structural similarities among renditions of rgas using a data-driven approach. Starting from melodic contours, we obtain the descriptive note-level transcription of each rendition. Repetitive note patterns of variable and fixed lengths are derived using stochastic models. We propose a latent variable approach for raga distinction based on statistics of these patterns. The posterior probability of the latent variable is shown to capture similarities across raga renditions. We show that it is possible to visualize the similarities in a low-dimensional embedded space. Experiments show that it is possible to compare and contrast relations and distances between ragas in the embedded space with the musicological knowledge of the same for both Hindustani and Carnatic music forms. The proposed approach also shows robustness to duration of rendition
