Anomalous Observations Research Articles

Spinodal decomposition during isothermal gas-solid equilibration – Its effects and implications

The possibility for a system to sample non-classical chemical pathways while attempting to attain thermodynamic equilibrium under conditions of restricted kinetics has been elucidated in this work through the analogy of gas phase nitriding of iron based alloys. Detailed thermodynamic assessment has for the first time shown evidence for the occurrence of in-situ isothermal spinodal decomposition during controlled nitridation of ferritic Fe-Cr alloys. Recognition of this phenomenon has led to the explanation of certain reported anomalous experimental observations: dependence of nitridation kinetics of ferritic Fe-Cr alloys on N diffusion and the paradox of Cr-nitrides precipitation sequence not following the predictions of the phase diagram in nitrided Fe-Cr alloys. Finally, the nitridation kinetics of Fe-Al alloys being controlled by Al diffusion has been attributed to the absence of a possibility for spinodal decomposition during nitriding of Fe-Al alloys.

Improving Data Quality for the Australian High Frequency Ocean Radar Network through Real-Time and Delayed-Mode Quality-Control Procedures

Quality-control procedures and their impact on data quality are described for the High-Frequency Ocean Radar (HFR) network in Australia, in particular for the commercial phased-array (WERA) HFR type. Threshold-based quality-control procedures were used to obtain radial velocity and signal-to-noise ratio (SNR), however, values were set through quantitative analyses with independent measurements available within the HFR coverage, when available, or from long-term data statistics. An artifact removal procedure was also applied to the spatial distribution of SNR for the first-order Bragg peaks, under the assumption the SNR is a valid proxy for radial velocity quality and that SNR decays with range from the receiver. The proposed iterative procedure was specially designed to remove anomalous observations associated with strong SNR peaks caused by the 50 Hz sources. The procedure iteratively fits a polynomial along the radial beam (1-D case) or a surface (2-D case) to the SNR associated with the radial velocity. Observations that exceed a detection threshold were then identified and flagged. After removing suspect data, new iterations were run with updated detection thresholds until no additional spikes were found or a maximum number of iterations was reached.

Numerical simulation of anomalous observations from an in-situ long-term sorption diffusion experiment in a rock matrix

At the in-situ Long Term Sorption Diffusion Experiment (LTDE-SD), performed in the Äspö Hard Rock Laboratory (Sweden), anomalous penetration profiles of tracers were observed. The concentration of sorbing tracers, particularly 137Cs, sharply decreased around the rock surface but was still quite high in the distance, which was known that it is difficult to model with the classic Fick’s law. To understand the mechanism of matrix diffusion and sorption of solute by reproducing the observed penetration profiles in the LTDE-SD test, three cases for the matrix were proposed and evaluated: (1) homogeneous medium with a disturbed zone; (2) heterogeneous medium considering mineral grains with intragranular porosities and their intergranular pores; and (3) heterogeneous medium considering vein and microfractures as well as porous mineral grains and intergranular pores. The modelling domains for Cases 2 and 3 were established using 2-D microstructure characterization results from analyses including chemical staining of minerals and 14C-PMMA autoradiography. The numerical models for Cases 1–3 were calibrated by minimizing the discrepancies between simulations and observations of the penetration profiles and the changes in concentration in the reservoir for the tracers, 137Cs and 22Na. The results show that only the calibrated transport parameter values for Case 3 were within the reported ranges, and we concluded that the sharp decrease in concentrations in the near-surface zone in the observed penetration profiles was likely to result from disturbed biotite grains and the long tails in the observed penetration profiles might originate from diffusion along the vein and microfractures transecting minerals.

Distribution of Vapor and Condensate in a Hydrothermal System: Insights From Self‐Potential Inversion at Mount Tongariro, New Zealand

AbstractInversion of self‐potential data for source current density, js, in complex volcanic settings, yields hydrological information without the need for a prior groundwater flow model; js contains information about pH, pore saturation, and permeability, from which we infer the distribution of liquid and vapor phases. To understand the hydrothermal flow dynamics and hydraulic connectivity between surface thermal features at Mount Tongariro volcano, New Zealand, we undertook a reconnaissance scale self‐potential survey and developed an inversion routine for js, constrained by an existing 3‐D conductivity model from magnetotelluric measurements. The 3‐D distribution of js at Mount Tongariro reveals a discontinuous zero js zone interpreted as vapor or residually saturated pore space, surrounded by low to moderate js interpreted as circulating condensate liquid. Bounding faults act as conduits for down flowing groundwater or condensate, as well as barriers for the hydrothermal system. Localized small‐scale circulation associated with individual surface thermal features, rather than a single circulating system, accounts for the lack of widespread anomalous geochemical observations prior to the 2012 Te Maari eruption.

Time series anomaly detection based on shapelet learning

We consider the problem of learning to detect anomalous time series from an unlabeled data set, possibly contaminated with anomalies in the training data. This scenario is important for applications in medicine, economics, or industrial quality control, in which labeling is difficult and requires expensive expert knowledge, and anomalous data is difficult to obtain. This article presents a novel method for unsupervised anomaly detection based on the shapelet transformation for time series. Our approach learns representative features that describe the shape of time series stemming from the normal class, and simultaneously learns to accurately detect anomalous time series. An objective function is proposed that encourages learning of a feature representation in which the normal time series lie within a compact hypersphere of the feature space, whereas anomalous observations will lie outside of a decision boundary. This objective is optimized by a block-coordinate descent procedure. Our method can efficiently detect anomalous time series in unseen test data without retraining the model by reusing the learned feature representation. We demonstrate on multiple benchmark data sets that our approach reliably detects anomalous time series, and is more robust than competing methods when the training instances contain anomalous time series.

Temperature dependent Raman investigation of multiwall carbon nanotubes

We report anomalous observations in our investigations of the temperature dependent Raman spectroscopic measurement of multiwall carbon nanotubes. The Micro-Raman spectra were recorded with the laser source having 514.5 nm wavelength and within the temperature range of 80–440 K. The major Raman bands, the G and D band, are observed at 1584 and 1348 cm−1, respectively, at ambient. The absence of the radial breathing mode confirms the multiwall nature of carbon nanotubes. It has been observed that with an increase in the temperature above 120 K, there is a shift in Raman bands towards the higher wave-number region. However, a drop in the G and D bands is observed from 80 to 120 K which was not observed for the second order band. Thereafter, all Raman modes exhibited mode hardening up to about 320 K followed by mild softening of the phonon modes. Linear temperature coefficients were found to have higher contribution to mode hardening as compared to higher order terms. Total anharmonicity estimation shows a predominant effect of the quasi-harmonic term as compared to the true anharmonic term.

Proliferation of twinning in hexagonal close-packed metals: Application to magnesium

Plastic deformation of metallic alloys usually takes place through slip, but occasionally involves twinning. In particular, twinning is important in hexagonal close packed (HCP) materials where the easy slip systems are insufficient to accommodate arbitrary deformations. While deformation by slip mechanisms is reasonably well understood, comparatively less is known about deformation by twinning. Indeed, the identification of relevant twinning modes remains an art. In this paper, we develop a framework combining a fundamental kinematic definition of twins with large-scale atomistic calculations to predict twinning modes of crystalline materials. We apply this framework to magnesium where there are two accepted twin modes, tension and compression, but a number of anomalous observations. Remarkably, our framework shows that there is a very large number of twinning modes that are important in magnesium. Thus, in contrast to the traditional view that plastic deformation is kinematically partitioned between a few modes, our results suggest that deformation in HCP materials is the result of an energetic and kinetic competition between numerous possibilities. Consequently, our findings suggest that the commonly used models of deformation need to be extended in order to take into account a broader and richer variety of twin modes, which, in turn, opens up new avenues for improving the mechanical properties.

Understanding evolution of vortex rings in viscous fluids

The evolution of vortex rings in isodensity and isoviscosity fluid has been studied analytically using a novel mathematical model. The model predicts the spatiotemporal variation in peak vorticity, circulation, vortex size and spacing based on instantaneous vortex parameters. This proposed model is quantitatively verified using experimental measurements. Experiments are conducted using high-speed particle image velocimetry (PIV) and laser induced fluorescence (LIF) techniques. Non-buoyant vortex rings are generated from a nozzle using a constant hydrostatic pressure tank. The vortex Reynolds number based on circulation $(\unicode[STIX]{x1D6E4}/\unicode[STIX]{x1D708})$ is varied in the range 100–1500 to account for a large range of operating conditions. Experimental results show good agreement with theoretical predictions. However, it is observed that neither Saffman’s thin-core model nor the thick-core equations could correctly explain vortex evolution for all initial conditions. Therefore, a transitional theory is framed using force balance equations which seamlessly integrate short- and long-time asymptotic theories. It is found that the parameter $A=(a/\unicode[STIX]{x1D70E})^{2}$, where $a$ is the vortex half-spacing and $\unicode[STIX]{x1D70E}$ denotes the standard deviation of the Gaussian vorticity profile, governs the regime of vortex evolution. For higher values of $A$, evolution follows short-time behaviour, while for $A=O(1)$, long-time behaviour is prominent. Using this theory, many reported anomalous observations have been explained.

Fault detection in operating helicopter drivetrain components based on support vector data description

The objective of the paper is to develop a vibration-based automated procedure dealing with early detection of mechanical degradation of helicopter drive train components using Health and Usage Monitoring Systems (HUMS) data. An anomaly-detection method devoted to the quantification of the degree of deviation of the mechanical state of a component from its nominal condition is developed. This method is based on an Anomaly Score (AS) formed by a combination of a set of statistical features correlated with specific damages, also known as Condition Indicators (CI), thus the operational variability is implicitly included in the model through the CI correlation. The problem of fault detection is then recast as a one-class classification problem in the space spanned by a set of CI, with the aim of a global differentiation between normal and anomalous observations, respectively related to healthy and supposedly faulty components. In this paper, a procedure based on an efficient one-class classification method that does not require any assumption on the data distribution, is used. The core of such an approach is the Support Vector Data Description (SVDD), that allows an efficient data description without the need of a significant amount of statistical data. Several analyses have been carried out in order to validate the proposed procedure, using flight vibration data collected from a H135, formerly known as EC135, servicing helicopter, for which micro-pitting damage on a gear was detected by HUMS and assessed through visual inspection. The capability of the proposed approach of providing better trade-off between false alarm rates and missed detection rates with respect to individual CI and to the AS obtained assuming jointly-Gaussian-distributed CI has been also analysed.

A clustering approach to detect multiple outliers in linear functional relationship model for circular data

ABSTRACTOutlier detection has been used extensively in data analysis to detect anomalous observation in data. It has important applications such as in fraud detection and robust analysis, among others. In this paper, we propose a method in detecting multiple outliers in linear functional relationship model for circular variables. Using the residual values of the Caires and Wyatt model, we applied the hierarchical clustering approach. With the use of a tree diagram, we illustrate the detection of outliers graphically. A Monte Carlo simulation study is done to verify the accuracy of the proposed method. Low probability of masking and swamping effects indicate the validity of the proposed approach. Also, the illustrations to two sets of real data are given to show its practical applicability.

Determinants of house buyers’ expected holding periods in boom and bust markets in California

PurposeThis study aims to investigate the determinants of homeowners’ planned holding periods. Real estate market is known for displaying buying and selling behavior that does not conform to traditional economic theories such as rational expectation or expected utility. Mounting evidence of anomalous observations appear to be supported by other theories, such as prospect theory, which in particular helps explain the disposition effect – sellers are too quick to sell when prices are climbing and hold on to properties longer when prices are plummeting. While this evidence is widely documented in housing studies based on data on realized holding periods (i.e. ex post), this study explores factors that may motivate homeowners to alter their expected holding horizons (i.e. ex ante) to form new preferred holding periods that may be shorter or longer than those planned during house search.Design/methodology/approachThe empirical study uses data collected from two cross-section surveys of recent homebuyers in rising and declining housing markets in Southern California in 2004-2005 and 2007-2008, respectively.FindingsThe empirical results demonstrate that in addition to the financial characteristics of the recent homebuyer, the characteristics of the buying experience – non-monetary, such as the realized search duration, and monetary, such as perception of negative or positive premium paid for the house relative to its market value – have a statistically significant effect on the holding horizon revision. The data strongly indicate that the perception of having overpaid increases the likelihood of upward revision of the original holding horizon. This effect is stronger in the declining than in the rising market – a crucial finding that mirrors the disposition effect.Originality/valueThis study sheds new light on what may contribute to the disposition effect in housing markets that has not yet been investigated in past literature. The novel approach here is to look at how different house price environments may affect homeowners’ holding periods ex ante when they begin, rather than ex post when already realized.

A framework for scalable real‐time anomaly detection over voluminous, geospatial data streams

SummaryThis study presents a framework to enable distributed detection, storage, and analysis of anomalies in voluminous data streams. Individual observations within these streams are multidimensional, with each dimension corresponding to a feature of interest. We consider time‐series geospatial datasets generated by remote and in situ observational devices. Three aspects make this problem particularly challenging: (1) the cumulative volume and rates of data arrivals, (2) evolution of the datasets over time, and (3) spatiotemporal correlations associated with the data. Further, solutions must minimize user intervention and be amenable to distributed processing to ensure scalability.Our approach achieves accurate, high‐throughput classifications in real time, which we demonstrate with our reference anomaly detector implementations. We also provide interfaces that allow new implementations to be developed and parallelized automatically, ensuring applicability across problem domains. To help quantify the magnitude of anomalous observations, detector implementations provide a corresponding degree of irregularity. We have incorporated these algorithms into our distributed storage platform, Galileo, and profiled their suitability through empirical analysis that demonstrates high throughput (10 000 observations per‐second, per‐node) on a real‐worl Petabyte dataset.

Innovation paradoxes: a review and typology of explanations

ABSTRACT The concept of innovation paradoxes refers to a family of anomalous observations demonstrating that relatively high or outstanding innovation efforts lead to either insignificant or undesirable outcomes. While researchers have long been busy studying the nature and causes of innovation paradoxes, they have yet to assess the fruits of their research efforts. This paper addresses this neglect, in particular by identifying and reviewing the literature of two innovation paradoxes – the European innovation paradox and the Swedish innovation paradox. The findings show that research on both paradoxes has proceeded along similar lines, leading to the development of a working explanatory typology of innovation paradoxes. The paper ends with a discussion of key observations, findings and suggestions.

Anomaly detection in earth dam and levee passive seismic data using support vector machines and automatic feature selection

We investigate techniques for earth dam and levee health monitoring and automatic detection of anomalous events in passive seismic data. We have developed a novel data-driven workflow specific to our domain, which could be generalized for monitoring other systems with time series data. We use machine learning and geophysical data collected from sensors located on the surface of the levee to identify internal erosion events. In this paper, we describe our research experiments with two-class and one-class support vector machines (SVMs). We use two different data sets from experimental laboratory earth embankments (each having approximately 80% normal and 20% anomalies) to ensure our workflow is robust enough to work with multiple data sets and different types of anomalous events (e.g., cracks and piping). We apply wavelet-denoising techniques and extract nine spectral features from decomposed segments of the time series data. The two-class SVM with 10-fold cross validation achieved over 94% overall accuracy and 96% F1-score. Experiments with the one-class SVM (no labeled data for anomalies) using the top features selected by our automatic feature selection algorithm increase our overall results from 83% accuracy and 89% F1-score to over 91% accuracy and 95% F1-score. Results show that we can successfully separate normal from anomalous data observations.

Detection of anomalous behavior in a robot system based on deep learning elements

The preprocessing procedure for anomalous behavior of robot system elements is proposed in the paper. It uses a special kind of a neural network called an autoencoder to solve two problems. The first problem is to decrease the dimensionality of the training data using the autoencoder to calculate the Mahalanobis distance, which can be viewed as one of the best metrics to detect the anomalous behavior of robots or sensors in the robot systems. The second problem is to apply the autoencoder to transfer learning. The autoencoder is trained by means of the target data which corresponds to the extreme operational conditions of the robot system. The source data containing the normal and anomalous observations derived from the normal operation conditions is reconstructed to the target data using the trained autoencoder. The reconstructed source data is used to define a optimal threshold for making decision on the anomaly of the observation based on the Mahalanobis distance.

Understanding the vibrational mode-specific polarization effects in femtosecond Raman-induced Kerr-effect spectroscopy.

Optically heterodyne-detected femtosecond Raman-induced Kerr-effect spectroscopy (OHD-FRIKES) was observed in neat cyclohexane. In this Letter, an examination of the effect of the Raman pump ellipticity on the multiplex OHD-FRIKES spectra is discussed. The Raman pump ellipticity scanned OHD-FRIKES results reproduce anomalous observables from previous OHD-FRIKES experiments and suggest new methods of tracking transient vibrational mode polarization in complex systems.

Investigation of barrier inhomogeneities and interface state density in Au/MgZnO: Ga Schottky contact

The electrical characteristics of Au/MgZnO:Ga (GMZO) Schottky contact, which is fabricated using a dual ion beam sputtering system, have been investigated using current–voltage (I–V) and capacitance–voltage (C–V) measurements over a wide temperature range of 80 to 300 K. The apparent Schottky barrier height (SBH) and ideality factor obtained from the I–V measurements are observed to increase and reduce, respectively, with increasing measurement temperature. This anomalous observation in the behaviour of the SBH is in good agreement with the predictions of a double Gaussian distribution (DGD) of the inhomogeneous SBH at a metal–semiconductor (MS) interface. The values of the SBH as determined from C–V measurements are expectedly higher than those extracted from I–V measurements. The DGD model is observed to fit with experimentally obtained data for the temperature-dependent SBH with mean values of the SBH of 0.95 and 0.54 eV and standard deviations of 0.131 and 0.072 eV in the temperature range of 160–300 K and 80–160 K, respectively. The larger value of the SBH standard deviation confirms more SBH inhomogeneity at the MS interface, and these inhomogeneities are attributed to the presence of deep level or surface level interface states. The calculated interface states density is seen to vary from 6.46 × 1014 eV−1 cm−2 at EC-0.27 eV to 1.58 × 1014 eV−1 cm−2 at EC-0.74 eV, where EC is the bottom of a conduction band at 300 K.

Enhanced nucleation of primary silicon in Al-20 wt. % Si alloy with Ni-Si inoculation

The inoculation effect of a novel Ni-32.5 wt. % Si master alloy on Al-20 wt. % Si melt was investigated. With the addition of 0.5 wt. % Ni-Si master alloy, primary Si is refined from ∼350 μm to ∼150 μm and the corresponding nucleation undercooling is reduced by 4.5 °C. Moreover, its morphologies become more regular with higher proportion of octahedrons. TEM studies reveal that inoculation introduces a high number density of nano-particles in Si grain. Careful structural analysis suggests that these nano-particles are different from any known crystal compounds containing Ni and Si. Both the high number density and the unknown crystal structures of these nano-particles are different from the heterogeneous nucleation theory. These anomalous observations imply that pre-nucleation clusters assisted nucleation mechanism is active here. The Ni-32.5 wt. % Si master alloy decomposes into local ordering clusters and triggers the formation of pre-nucleation clusters in Al-Si melt. The pre-nucleation clusters accelerate crystal nucleation and transform into the nano-particles.

Anytime algorithm for frequent pattern outlier detection

Outlier detection consists in detecting anomalous observations from data. During the past decade, outlier detection methods were proposed using the concept of frequent patterns. Basically such methods require to mine all frequent patterns for computing the outlier factor of each transaction. This approach remains too expensive despite recent progress in pattern mining field to provide results within a short response time of only a few seconds. In this paper, we provide the first anytime method for calculating the frequent pattern outlier factor (FPOF). This method which can be interrupted at anytime by the end-user accurately approximates FPOF by mining a sample of patterns. It also computes the maximum error on the estimated FPOF for helping the user to stop the process at the right time. Experiments show the interest of this method for very large datasets where exhaustive mining fails to provide good approximate solutions. The accuracy of our anytime approximate method outperforms the baseline approach for a same budget in number of patterns.

Laser fabrication of W-reinforced Cu layers: I. Corrosion behavior in 3.5% NaCl solution and synthetic acid rain

Surface modification of commercially pure copper (cp Cu) with W powder was achieved by using a high power diode laser (HPDL) resulting in a microstructure of micron-sized W particles in a Cu matrix. In both 3.5 wt% NaCl solution and synthetic acid rain (SAR), the laser-fabricated specimens possess lower corrosion current density (Icorr) than that of cp Cu despite the active shift in open-circuit potential (OCP). Although Cu is cathodic to W in 3.5 wt% NaCl solution, the Cu phase in the laser-fabricated specimens dominantly dissolves during the immersion test. This anomalous observation is possible because the OCP values of the laser-fabricated specimens, cp Cu and cp W are quite near to each other, while the anodic and cathodic current densities are all not negligible for the Cu and W phases at OCP. As the anodic current density of the Cu phase is greater than anodic current density of the W phase, the Cu phase corrodes more quickly than the W phase. For the laser-fabricated specimens immersed in SAR, the Cu phase is anodic and is selectively attacked while the W phase is passive due to the formation of WO3.