Probabilistic fault identification using vibration data and neural networks

Bayesian formulated neural networks are implemented using hybrid Monte Carlo method for probabilistic fault identification in cylindrical shells. Each of the 20 nominally identical cylindrical shells is divided into three substructures. Holes of (12±2) mm in diameter are introduced in each of the substructures and vibration data are measured. Modal properties and the Coordinate Modal Assurance Criterion (COMAC) are utilized to train the two modal-property-neural-networks. These COMAC are calculated by taking the natural-frequency-vector to be an additional mode. Modal energies are calculated by determining the integrals of the real and imaginary components of the frequency response functions over bandwidths of 12% of the natural frequencies. The modal energies and the Coordinate Modal Energy Assurance Criterion (COMEAC) are used to train the two frequency-response-function-neural-networks. The averages of the two sets of trained-networks (COMAC and COMEAC as well as modal properties and modal energies) form two committees of networks. The COMEAC and the COMAC are found to be better identification data than using modal properties and modal energies directly. The committee approach is observed to give lower standard deviations than the individual methods. The main advantage of the Bayesian formulation is that it gives identities of damage and their respective confidence intervals.

Neuro-Fuzzy genetic system for selection of construction project managers

Choosing a project manager for a construction project—particularly, large projects—is a critical project decision. The selection process involves different criteria and should be in accordance with company policies and project specifications. Traditionally, potential candidates are interviewed and the most qualified are selected in compliance with company priorities and project conditions. Precise computing models that could take various candidates’ information into consideration and then pinpoint the most qualified person with a high degree of accuracy would be beneficial. On the basis of the opinions of experienced construction company managers, this paper, through presenting a fuzzy system, identifies the important criteria in selecting a project manager. The proposed fuzzy system is based on IF-THEN rules; a genetic algorithm improves the overall accuracy as well as the functions used by the fuzzy system to make initial estimates of the cluster centers for fuzzy c-means clustering. Moreover, a back-propagation neutral network method was used to train the system. The optimal measures of the inference parameters were identified by calculating the system’s output error and propagating this error within the system. After specifying the system parameters, the membership function parameters—which by means of clustering and projection were approximated—were tuned with the genetic algorithm. Results from this system in selecting project managers show its high capability in making high-quality personnel predictions

Policy optimisation of POMDP-based dialogue systems without state space compression

The partially observable Markov decision process (POMDP) has been proposed as a dialogue model that enables automatic improvement of the dialogue policy and robustness to speech understanding errors. It requires, however, a large number of dialogues to train the dialogue policy. Gaussian processes (GP) have recently been applied to POMDP dialogue management optimisation showing an ability to substantially increase the speed of learning. Here, we investigate this further using the Bayesian Update of Dialogue State dialogue manager. We show that it is possible to apply Gaussian processes directly to the belief state, removing the need for a parametric policy representation. In addition, the resulting policy learns significantly faster while maintaining operational performance. © 2012 IEEE.

Gaussian processes for POMDP-based dialogue manager optimization

A partially observable Markov decision process (POMDP) has been proposed as a dialog model that enables automatic optimization of the dialog policy and provides robustness to speech understanding errors. Various approximations allow such a model to be used for building real-world dialog systems. However, they require a large number of dialogs to train the dialog policy and hence they typically rely on the availability of a user simulator. They also require significant designer effort to hand-craft the policy representation. We investigate the use of Gaussian processes (GPs) in policy modeling to overcome these problems. We show that GP policy optimization can be implemented for a real world POMDP dialog manager, and in particular: 1) we examine different formulations of a GP policy to minimize variability in the learning process; 2) we find that the use of GP increases the learning rate by an order of magnitude thereby allowing learning by direct interaction with human users; and 3) we demonstrate that designer effort can be substantially reduced by basing the policy directly on the full belief space thereby avoiding ad hoc feature space modeling. Overall, the GP approach represents an important step forward towards fully automatic dialog policy optimization in real world systems. © 2013 IEEE.

Applications of stochastic context-free grammars using the Inside-Outside algorithm

This paper describes two applications in speech recognition of the use of stochastic context-free grammars (SCFGs) trained automatically via the Inside-Outside Algorithm. First, SCFGs are used to model VQ encoded speech for isolated word recognition and are compared directly to HMMs used for the same task. It is shown that SCFGs can model this low-level VQ data accurately and that a regular grammar based pre-training algorithm is effective both for reducing training time and obtaining robust solutions. Second, an SCFG is inferred from a transcription of the speech used to train a phoneme-based recognizer in an attempt to model phonotactic constraints. When used as a language model, this SCFG gives improved performance over a comparable regular grammar or bigram. © 1991.

The modified Kanerva model for automatic speech recognition

A parallel processing network derived from Kanerva's associative memory theory Kanerva 1984 is shown to be able to train rapidly on connected speech data and recognize further speech data with a label error rate of 0·68%. This modified Kanerva model can be trained substantially faster than other networks with comparable pattern discrimination properties. Kanerva presented his theory of a self-propagating search in 1984, and showed theoretically that large-scale versions of his model would have powerful pattern matching properties. This paper describes how the design for the modified Kanerva model is derived from Kanerva's original theory. Several designs are tested to discover which form may be implemented fastest while still maintaining versatile recognition performance. A method is developed to deal with the time varying nature of the speech signal by recognizing static patterns together with a fixed quantity of contextual information. In order to recognize speech features in different contexts it is necessary for a network to be able to model disjoint pattern classes. This type of modelling cannot be performed by a single layer of links. Network research was once held back by the inability of single-layer networks to solve this sort of problem, and the lack of a training algorithm for multi-layer networks. Rumelhart, Hinton & Williams 1985 provided one solution by demonstrating the "back propagation" training algorithm for multi-layer networks. A second alternative is used in the modified Kanerva model. A non-linear fixed transformation maps the pattern space into a space of higher dimensionality in which the speech features are linearly separable. A single-layer network may then be used to perform the recognition. The advantage of this solution over the other using multi-layer networks lies in the greater power and speed of the single-layer network training algorithm. © 1989.

A three-dimensional tunnel model for calculation of train-induced ground vibration

The frequency range of interest for ground vibration from underground urban railways is approximately 20 to 100 Hz. For typical soils, the wavelengths of ground vibration in this frequency range are of the order of the spacing of train axles, the tunnel diameter and the distance from the tunnel to nearby building foundations. For accurate modelling, the interactions between these entities therefore have to be taken into account. This paper describes an analytical three-dimensional model for the dynamics of a deep underground railway tunnel of circular cross-section. The tunnel is conceptualised as an infinitely long, thin cylindrical shell surrounded by soil of infinite radial extent. The soil is modelled by means of the wave equations for an elastic continuum. The coupled problem is solved in the frequency domain by Fourier decomposition into ring modes circumferentially and a Fourier transform into the wavenumber domain longitudinally. Numerical results for the tunnel and soil responses due to a normal point load applied to the tunnel invert are presented. The tunnel model is suitable for use in combination with track models to calculate the ground vibration due to excitation by running trains and to evaluate different track configurations. © 2006 Elsevier Ltd. All rights reserved.

The effect of a twin tunnel on the propagation of ground-borne vibration from an underground railway

Accurate predictions of ground-borne vibration levels in the vicinity of an underground railway are greatly sought after in modern urban centres. Yet the complexity involved in simulating the underground environment means that it is necessary to make simplifying assumptions about this system. One such commonly made assumption is to ignore the effects of neighbouring tunnels, despite the fact that many underground railway lines consist of twin-bored tunnels, one for the outbound direction and one for the inbound direction. This paper presents a unique model for two tunnels embedded in a homogeneous, elastic fullspace. Each of these tunnels is subject to both known, dynamic train forces and dynamic cavity forces. The net forces acting on the tunnels are written as the sum of those tractions acting on the invert of a single tunnel, and those tractions that represent the motion induced by the neighbouring tunnel. By apportioning the tractions in this way, the vibration response of a two-tunnel system is written as a linear combination of displacement fields produced by a single-tunnel system. Using Fourier decomposition, forces are partitioned into symmetric and antisymmetric modenumber components to minimise computation times. The significance of the interactions between two tunnels is quantified by calculating the insertion gains, in both the vertical and horizontal directions, that result from the existence of a second tunnel. The insertion-gain results are shown to be localised and highly dependent on frequency, tunnel orientation and tunnel thickness. At some locations, the magnitude of these insertion gains is greater than 20 dB. This demonstrates that a high degree of inaccuracy exists in any surface vibration prediction model that includes only one of the two tunnels. This novel two-tunnel solution represents a significant contribution to the existing body of research into vibration from underground railways, as it shows that the second tunnel has a significant influence on the accuracy of vibration predictions for underground railways. © 2011 Elsevier Ltd. All rights reserved.

Characteristics of the spectral dynamics of picosecond modulated AlGaAs/GaAs injection lasers

Breakdown of the optical spectrum of a train of picosecond pulses into components with a distance which exceeds kT (200 cm-1 at λ = 955 nm and T = 300 K) is discovered for the first time in an injection laser. The effect may be caused by combined interaction between photons and phonons, with collective excitations in the degraded electron-hole GaAs plasma, and with the stream of drifting carriers in the active medium of the laser.

Generation of 140 fs pulse train with widely tunable repetition rate through cascaded fibre compression techniques

The authors present numerical simulations of ultrashort pulse generation by a technique of linear spectral broadening in phase modulators and compression in dispersion compensating fibre, followed by a further stage of soliton compression in dispersion shifted fibre. This laser system is predicted to generate pulses of 140 fs duration with a peak power of 1.5 kW over a wide, user selectable repetition rate range while maintaining consistent characteristics of stability and pulse quality. The use of fibre compressors and commercially available modulators is expected to make the system setup compact and cost-effective. © The Institution of Engineering and Technology 2014.