50175 research outputs found
Sort by
Discovering frequent chain episodes
No full textFrequent episode discovery is a popular framework in temporal data mining with many applications. An episode is a partially ordered set of nodes with each node associated with an event-type. The episodes literature has seen different notions of frequency and a variety of associated discovery algorithms under these different frequencies when the associated partial order is total (serial episode) or trivial (parallel episode). Recently an apriori-based discovery algorithm for mining episodes where the associated partial order has no restriction but the node to event-type association is one-one (general injective episodes) was proposed based on the non-overlapped frequency measure. This work pointed out that frequency alone is not a sufficient indicator of interestingness in the context of episodes with general partial orders and introduced a new measure of interestingness called bidirectional evidence (BE) to address this issue. This algorithm discovers episodes by incorporating both frequency and BE thresholds in the level-wise procedure. In this paper, we extend this BE-based algorithm to a much larger class of episodes that we call chain episodes. This class encompasses all serial and parallel episodes (injective or otherwise) and also many other non-injective episodes with unrestricted partial orders. We first discuss how the BE measure can be generalized to chain episodes and prove the monotonicity property it satisfies in this general context. We then describe our candidate generation step (with correctness proofs) which nicely exploits this new monotonicity property. We further describe the frequency counting (with correctness proofs) and BE computation steps for chain episodes. The experimental results demonstrate the effectiveness of our algorithms
Analytical Expressions to Link SCNF and OCNF of Transformer Windings to Their Inductances and Capacitances for 1-Phi, 3-Phi Y and Delta Configurations
No full textRecently, the present authors' research group derived an analytical expression to link the harmonic sum of the squares of short circuit natural frequencies (SCNFs) of a single, isolated winding to its elementary inductances and capacitances. Also, it was demonstrated how this expression could be practically used for locating radial and axial displacements, along with assessing its severity on a single, isolated, actual transformer winding. The next logical task was to extend the derived expression to multi-phase windings, viz., star/delta connections and for any terminal condition of the neutral. Actually, it would be desirable to have a similar expression for open circuit natural frequency (OCNF) as well. A major highlight of this paper is that it presents a single, unified approach (using salient features of coefficients of the numerator and denominator polynomial of driving-point impedance) by which compact analytical expressions can be simultaneously derived for both SCNF and OCNF, for any multi-phase transformer configuration, and for any condition of the neutral terminal. Complete derivation details of the proposed method are presented, and then, the expressions for SCNF and OCNF are derived for 1-Phi isolated windings as well as 3-Phi (Y, Delta) configurations. Finally, simulation results for all the cases are reported
The Madden-Julian Oscillation in an Aquaplanet-Like General Circulation Model With and Without Continents
No full textWe have conducted experiments with the Community Atmospheric Model version 5.1 configured as an aquaplanet, with the sun at March equinox, in order to study how the model Madden-Julian oscillation (MJO) responds to (i) the introduction of continents and (ii) realistic sea surface temperature (SST) distribution. Model results are compared with ERA-Interim reanalysis and analyzed in terms of the moist static energy (MSE) budget to study the growth and propagation of MJO. With idealized zonally symmetric SST, the aquaplanet model produces a double Intertropical Convergence Zone (ITCZ) and an MJO-like mode with variance at intraseasonal (30- to 96-day) periods and zonal wave number 1. When we introduce continents with realistic orography and interactive surface temperature, soil moisture, and albedo, the variance of MJO is mainly confined to zonal wave numbers 1 and 2 but reduced by a factor of 3 due to weaker boundary layer moisture convergence. With prescribed climatological January SST boundary condition (which includes Indo-west Pacific warm pool) in the presence of continents, model MJO variance is enhanced and distributed across zonal wave numbers 1 to 5, in closer agreement with observation. Thus, the presence of land by itself is not enough, but realistic SST distribution is necessary for improved space-time characteristics of model MJO. Both in simulations and ERA-Interim data, meridional and vertical advection of MSE promotes eastward movement of MJO. In the model experiments, meridional advection of low-level MSE anomaly is most significant in the vicinity of the ITCZ, indicating the importance of processes that determine the location of (single or double) ITCZ
Thermal, rheological, wettability and hard machining performance of MoS2 and CaF2 based minimum quantity hybrid nano-green cutting fluids
No full textMinimum quantity cutting fluids (MQCF) is an alternative for dry machining and flood cooling. However, use of MQCF is limited to mild machining conditions due to high heat generation during machining of hard materials. The applicability of MQCF can be extended in aggressive machining conditions by using vegetable-based green cutting fluids (GCF) with solid lubricant nanoparticles as potential additives. In the present study, hybrid nano green cutting fluids (HN-GCFs) were prepared by mixing different volume concentration of calcium fluoride (CaF2) and molybdenum disulphide (MoS2) solid lubricants in indigenously developed vegetable based GCF. Dispersion stability, thermal conductivity, specific volumetric heat capacity, viscosities and wetting angle of HN-GCFs were measured. In-house fabricated MQCF setup is used to apply minimum HN-GCF across the machining region. Later, the experimentally optimised concentration of both CaF2 and MoS2 based HN-GCFs were selected for machining experiments using MQCF setup. Hard machining experiments were carried out using four different cutting fluids: commercial mineral oil (MO), GCF, 0.3% concentration of CaF2 based HN-GCF (HN-GCF-0.3C) and 0.3% concentration of MoS2 based HN-GCF (HN-GCF-0.3M). Comparative tool wear, workpiece material adhesion and workpiece surface morphology were also investigated. Results show that 0.3% concentration of HN-GCF-0.3M performs better concerning tool-workpiece interface coefficient of friction, cutting force, feed force, and workpiece surface roughness compared to other cutting fluids. During machining with HN-GCF-0.3M, cutting tool experiences 17% and 28% lower cutting as well as feed force, respectively as compared with machining using MO. Also, 11% reduction in tool-chip interface coefficient of friction was observed during machining using HN-GCF-0.3M, and it also produces 37% better workpiece surface finish. Elemental analysis of cutting tool confirms less work material adhesion on the cutting tool rake face using HN-GCF-0.3M
CNN Fixations: An Unraveling Approach to Visualize the Discriminative Image Regions
No full textDeep convolutional neural networks (CNNs) have revolutionized the computer vision research and have seen unprecedented adoption for multiple tasks, such as classification, detection, and caption generation. However, they offer little transparency into their inner workings and are often treated as black boxes that deliver excellent performance. In this paper, we aim at alleviating this opaqueness of CNNs by providing visual explanations for the network's predictions. Our approach can analyze a variety of CNN-based models trained for computer vision applications, such as object recognition and caption generation. Unlike the existing methods, we achieve this via unraveling the forward pass operation. The proposed method exploits feature dependencies across the layer hierarchy and uncovers the discriminative image locations that guide the network's predictions. We name these locations CNN fixations, loosely analogous to human eye fixations. Our approach is a generic method that requires no architectural changes, additional training, or gradient computation, and computes the important image locations (CNN fixations). We demonstrate through a variety of applications that our approach is able to localize the discriminative image locations across different network architectures, diverse vision tasks, and data modalities
Flow investigations of reverse running volute pumps with backward vanes in comparison to forward type turbine vanes
No full textThe use of centrifugal pumps as turbines in the recent years has come as a boon to small and micro power application given its simplicity and robustness. However, attempts are continuously being made to improve the performance by modifying the geometry, and yet to retain its simpleness. This paper proposes a new design to the impeller of an existing pump with forward-shaped vanes in an unchanged volute that is in complete contrast to the conventional backward vanes. Three methodologies are involved in the analysis starting with classical theory, experiment and simulations. The theory entails the focus on optimizing the nozzle shape of the impeller to reduce viscous and eddies. The forward vane having shorter nozzle length has proved to be more efficient compared to the longer backward vane impeller, with an efficiency increase of nearly 5%. The experimental and CFD analysis to study the internal flow saw similitude in the streamline change in Euler moment. It also showed that there were not only increased viscous effects but also enhanced flow separation in the backward vanes at overload flows. The study also found radial clearance losses to be unacceptable for both the shapes. The overall conclusion was to move ahead with the forward design and convince the industry to adopt them for there have been greater strides in cost-effective manufacturing processes. The paper also recommends more study of intermediate blade angles since there was still persistence of small degree of vorticities in the forward blade nozzle. The optimization of volute-impeller interface along with the influence of non-flow zone would be other areas for future investigation. Synergy of academia and industry is also welcome to lend improved understanding in pumps as turbines and better translation to praxis
Fibonacci steady states in a driven integrable quantum system
No full textWe study an integrable system that is reducible to free fermions by a Jordan-Wigner transformation which is subjected to a Fibonacci driving protocol based on two noncommuting Hamiltonians. In the high-frequency limit omega -> infinity, we show that the system reaches a nonequilibrium steady state, up to some small fluctuations which can be quantified. For each momentum k, the trajectory of the stroboscopically observed state lies between two concentric circles on the Bloch sphere; the circles represent the boundaries of the small fluctuations. The residual energy is found to oscillate in a quasiperiodic way between two values which correspond to the two Hamiltonians that define the Fibonacci protocol. These results can be understood in terms of an effective Hamiltonian which simulates the dynamics of the system in the high-frequency limit
Evidence for exchange bias coupling at the perovskite/brownmillerite interface in spontaneously stabilized SrCoO3-delta/SrCoO2.5 bilayers
No full textInterface effect in complex oxide thin-film heterostructures lies at the vanguard of current research to design technologically relevant functionality and explore emergent physical phenomena. While most of the previous works focus on the perovskite/perovskite heterostructures, the study of perovskite/brownmillerite interfaces remains in its infancy. Here, we investigate spontaneously stabilized perovskite-ferromagnet (SrCoO3-delta)/brownmillerite-antiferromagnet (SrCoO2.5) bilayer with T-N > T-C and discover an unconventional interfacial magnetic exchange bias effect. From magnetometry investigations, it is rationalized that the observed effect stems from the interfacial ferromagnet/antiferromagnet coupling. The possibility for coupled ferromagnet/spin-glass interface engendering such effect is ruled out. Strikingly, a finite coercive field persists in the paramagnetic state of SrCoO3-delta,whereas the exchange bias field vanishes at T-C . We conjecture the observed effect to be due to the effective external quenched staggered field provided by the antiferromagnetic layer for the ferromagnetic spins at the interface. Our results not only unveil a paradigm to tailor the interfacial magnetic properties in oxide heterostructures without altering the cations at the interface, but also provide a purview to delve into the fundamental aspects of exchange bias in such unusual systems, paving a big step forward in thin-film magnetism
Ferromagnetism in the multiferroic alloy systems BiFeO3-BaTiO3 and BiFeO3-SrTiO3: Intrinsic or extrinsic?
No full textAmong the different bulk alloys of the room temperature multiferroic compound BiFeO3, alloying with ATiO(3) (A = Ba, Sr, Pb) is interesting as they have been reported to induce ferromagnetism under certain conditions. While this makes them interesting as potential multiferroic magnetoelectric materials, there is a lack of clarity on the origin of ferromagnetism in these systems. We have performed a detailed magneto-structural analysis on two alloys, namely, BiFeO3-BaTiO3 and BiFeO3-SrTiO3, and found the systems' propensity for the formation of ferrimagnetic hexaferrite (BaFe12O19 and SrFe12O19) phases. Though insignificantly small to the extent of being undetectable by x-ray powder diffraction, we prove that the ferromagnetic character of the specimen is entirely due to spontaneous precipitation of the hexaferrite grains. While our results suggest that care should be exercised before attributing ferromagnetism in such alloy systems as intrinsic to the perovskite phase, the propensity for the spontaneous precipitation of the hexaferrite phase in these multiferroic alloy systems is encouraging as it provides an opportunity for designing selfgrown ferroelectric-ferromagnetic composites with good magnetoelectric coupling
