Nonlinear analysis for the pre- and post-yield behaviour of a composite structure with the refined plastic hinge approach

The computational technique of the full ranges of the second-order inelastic behaviour evaluation of steel-concrete composite structure is not always sought forgivingly, and therefore it hinders the development and application of the performance-based design approach for the composite structure. To this end, this paper addresses of the advanced computational technique of the higher-order element with the refined plastic hinges to capture the all-ranges behaviour of an entire steel-concrete composite structure. Moreover, this paper presents the efficient and economical cross-section analysis to evaluate the element section capacity of the non-uniform and arbitrary composite section subjected to the axial and bending interaction. Based on the same single algorithm, it can accurately and effectively evaluate nearly continuous interaction capacity curve from decompression to pure bending technically, which is the important capacity range but highly nonlinear. Hence, this cross-section analysis provides the simple but unique algorithm for the design approach. In summary, the present nonlinear computational technique can simulate both material and geometric nonlinearities of the composite structure in the accurate, efficient and reliable fashion, including partial shear connection and gradual yielding at pre-yield stage, plasticity and strain-hardening effect due to axial and bending interaction at post-yield stage, loading redistribution, second-order P-δ and P-Δ effect, and also the stiffness and strength deterioration. And because of its reliable and accurate behavioural evaluation, the present technique can be extended for the design of the high-strength composite structure and potentially for the fibre-reinforced concrete structure.

Finite element modelling of lamb wave propagation in prestress concrete and effect of the prestress force on the wave’s characteristic

In order to assess the structural reliability of bridges, an accurate and cost effective Non-Destructive Evaluation (NDE) technology is required to ensure their safe and reliable operation. Over 60% of the Australian National Highway System is prestressed concrete (PSC) bridges according to the Bureau of Transport and Communication Economics (1997). Most of the in-service bridges are more than 30 years old and may experience a heavier traffic load than their original intended level. Use of Ultrasonic waves is continuously increasing for (NDE) and Structural Health Monitoring (SHM) in civil, aerospace, electrical, mechanical applications. Ultrasonic Lamb waves are becoming more popular for NDE because it can propagate long distance and reach hidden regions with less energy loses. The purpose of this study is to numerically quantify prestress force (PSF) of (PSC) beam using the fundamental theory of acoustic-elasticity. A three-dimension finite element modelling approach is set up to perform parametric studies in order to better understand how the lamb wave propagation in PSC beam is affected by changing in the PSF level. Results from acoustic-elastic measurement on prestressed beam are presented, showing the feasibility of the lamb wave for PSF evaluation in PSC bridges.

R1SVM: A randomised nonlinear approach to large-scale anomaly detection

The problem of unsupervised anomaly detection arises in a wide variety of practical applications. While one-class support vector machines have demonstrated their effectiveness as an anomaly detection technique, their ability to model large datasets is limited due to their memory and time complexity for training. To address this issue for supervised learning of kernel machines, there has been growing interest in random projection methods as an alternative to the computationally expensive problems of kernel matrix construction and sup-port vector optimisation. In this paper we leverage the theory of nonlinear random projections and propose the Randomised One-class SVM (R1SVM), which is an efficient and scalable anomaly detection technique that can be trained on large-scale datasets. Our empirical analysis on several real-life and synthetic datasets shows that our randomised 1SVM algorithm achieves comparable or better accuracy to deep auto encoder and traditional kernelised approaches for anomaly detection, while being approximately 100 times faster in training and testing.