Phonon anomalies predict superconducting Tc for AlB2-type structures

We show that the well-known Kohn anomaly predicts Tc for ordered AlB2-type structures. We use ab initio Density Functional Theory to calculate phonon dispersions for Mg1-xAlxB2 compositions and identify a phonon anomaly with magnitude that predicts experimental values of Tc for all x. Key features of these anomalies correlate with the electronic structure of Mg1-xAlxB2. This approach predicts Tc for other known AlB2-type structures as well as new compositions. We predict that Mg0.5Ba0.5B2 will show Tc = 63.6 ± 6.6 K. Other forms of the Mg1-xBaxB2 series will also be superconductors when successfully synthesised. Our calculations predict that the end-member composition, BaB2, is likely to show a Tc significantly higher than currently achieved by other diborides although an applied pressure ~16 GPa may be required to stabilise the structure.

Speed-based diagnostics of aerodynamic and mass imbalance in large wind turbines

Any kind of imbalance in the operation of a wind turbine has adverse effect on the downstream torsional components as well as tower structure. It is crucial to detect imbalance at its very inception. The identification of the type of imbalance is also required so that appropriate measures of fault accommodation can be performed in the control system. In particular, it is important to distinguish between mass and aerodynamic imbalance. While the former is gradually caused by a structural anomaly (e.g. ice deposition, moisture accumulation inside blade), the latter is generally associated to a fault in the pitch control system. This paper proposes a technique for the detection and identification of imbalance fault in large scale wind turbines. Unlike most other existing method it requires only the rotor speed signal which is readily available in existing turbines. Signature frequencies have been proposed in this work to identify imbalance type based on their physical phenomenology. The performance of this technique has been evaluated by simulations using an existing benchmark model. The effectiveness of the proposed method has been confirmed by the simulation results.

Single-walled carbon nanotube-based polymer monoliths for the enantioselective nano-liquid chromatographic separation of racemic pharmaceuticals

Many protocols have been used for extraction of DNA from Thraustochytrids. These generally involve the use of CTAB, phenol/chloroform and ethanol. They also feature mechanical grinding, sonication, N2 freezing or bead beating. However, the resulting chemical and physical damage to extracted DNA reduces its quality. The methods are also unsuitable for large numbers of samples. Commercially-available DNA extraction kits give better quality and yields but are expensive. Therefore, an optimized DNA extraction protocol was developed which is suitable for Thraustochytrids to both minimise expensive and time-consuming steps prior to DNA extraction and also to improve the yield. The most effective method is a combination of single bead in TissueLyser (Qiagen) and Proteinase K. Results were conclusive: both the quality and the yield of extracted DNA were higher than with any other method giving an average yield of 8.5 µg/100 mg biomass.

Detecting rare events using Kullback-Leibler divergence: A weakly supervised approach

Video surveillance infrastructure has been widely installed in public places for security purposes. However, live video feeds are typically monitored by human staff, making the detection of important events as they occur difficult. As such, an expert system that can automatically detect events of interest in surveillance footage is highly desirable. Although a number of approaches have been proposed, they have significant limitations: supervised approaches, which can detect a specific event, ideally require a large number of samples with the event spatially and temporally localised; while unsupervised approaches, which do not require this demanding annotation, can only detect whether an event is abnormal and not specific event types. To overcome these problems, we formulate a weakly-supervised approach using Kullback-Leibler (KL) divergence to detect rare events. The proposed approach leverages the sparse nature of the target events to its advantage, and we show that this data imbalance guarantees the existence of a decision boundary to separate samples that contain the target event from those that do not. This trait, combined with the coarse annotation used by weakly supervised learning (that only indicates approximately when an event occurs), greatly reduces the annotation burden while retaining the ability to detect specific events. Furthermore, the proposed classifier requires only a decision threshold, simplifying its use compared to other weakly supervised approaches. We show that the proposed approach outperforms state-of-the-art methods on a popular real-world traffic surveillance dataset, while preserving real time performance.

Refractory boron compounds for high performance: Theory and experiment

We have shown that novel synthesis methods combined with careful evaluation of DFT phonon calculations provides new insight into boron compounds including a capacity to predict Tc for AlB2-type superconductors.

R1STM: One-class support tensor machine with randomised kernel

Identifying unusual or anomalous patterns in an underlying dataset is an important but challenging task in many applications. The focus of the unsupervised anomaly detection literature has mostly been on vectorised data. However, many applications are more naturally described using higher-order tensor representations. Approaches that vectorise tensorial data can destroy the structural information encoded in the high-dimensional space, and lead to the problem of the curse of dimensionality. In this paper we present the first unsupervised tensorial anomaly detection method, along with a randomised version of our method. Our anomaly detection method, the One-class Support Tensor Machine (1STM), is a generalisation of conventional one-class Support Vector Machines to higher-order spaces. 1STM preserves the multiway structure of tensor data, while achieving significant improvement in accuracy and efficiency over conventional vectorised methods. We then leverage the theory of nonlinear random projections to propose the Randomised 1STM (R1STM). Our empirical analysis on several real and synthetic datasets shows that our R1STM algorithm delivers comparable or better accuracy to a state-of-the-art deep learning method and traditional kernelised approaches for anomaly detection, while being approximately 100 times faster in training and testing.