Papers by Author: Gareth Pierce

Paper TitlePage

Authors: Kenneth Burnham, Gareth Pierce
Abstract: The total electrical generation capacity from wind sources in the International Energy Agency (IEA) Wind Member Countries has increased from 4 GW in 1995 to more than 51 GW in 2005 thus underlining the strategic importance of the resource. In the last year alone the UK increased its wind generation by 447 MW, an increase of 85% over that for the previous year. In 2004, wind generation formed just 0.5% of the national electric demand; this contribution is set to rise over the next few years with some predictions that wind energy will rise to 8% of the total UK demand by 2010. The rotor blades of a wind turbine system are a significant structural component of the overall system, and typically account for 30% of lifecycle costs, and contribute 34% to overall system downtime. Despite their importance, there is currently very little monitoring of the structural integrity of rotor components, and what does exist is limited. We perceive that especially with the current political and technological emphasis on offshore installations, there will be an increase in the perceived need for remote structural monitoring, and there is indeed currently great interest in this area from the wind turbine industry. This work focuses on the applications of acoustic techniques to assess the integrity of typical rotor blade structures. Preliminary results discuss the limiting aspects of acoustic based techniques based on the physics of acoustic wave propagation in typical structural components. Comparisons between acoustic emission approaches and conventional active ultrasound will be considered.
Authors: Graeme Manson, Gareth Pierce, Keith Worden, Daley Chetwynd
Abstract: This paper considers the performance of radial basis function neural networks for the purpose of data classification. The methods are illustrated using a simple two class problem. Two techniques for reducing the rate of misclassifications, via the introduction of an “unable to classify” label, are presented. The first of these considers the imposition of a threshold value on the classifier outputs whilst the second considers the replacement of the crisp network weights with interval ranges. Two network training techniques are investigated and it is found that, although thresholding and uncertain weights give similar results, the level of variability of network performance is dependent upon the training approach
Authors: Gareth Pierce, Keith Worden, Graeme Manson
Abstract: The application of a multilayer perceptron (MLP) neural network to a damage location problem on a GNAT aircraft wing is considered. The problems associated with effective network training and evaluation are discussed, focussing on ensuring good generalisation performance of the network to the classification of new data. Both conventional Maximum Likelihood and Bayesian Evidence based training techniques are considered and a simple thresholding technique is presented to aid in the rejection of poorly regularised network structures. Examples are presented for an artificial simple 2 class problem (drawn from a Gaussian distribution) and a real 9 class problem on the GNAT aircraft wing.
Authors: W.J. Staszewski, Gareth Pierce, Keith Worden, B. Culshaw
Authors: Gabriel Bolu, Gareth Pierce, Anthony Gachagan, Tim Barden, Gerald Harvey
Abstract: Thermosonics is a rapid and potentially cost-saving non-destructive testing (NDT) screening technique that can be applied to the identification of cracks in high pressure compressor turbine blades in turbofan engines. The reliability of the thermosonic technique is not well established for inspecting these complex components; in particular the vibrational energy generated within a component during a thermosonic test is often highly non-uniform, leading to the possibility of missing critical defects. The aim of this study was to develop a methodology, using a combination of vibration measurements and finite element analysis (FEA), to model the vibrational energy within a turbine blade in a typical thermosonic inspection scenario. Using a laser vibrometer, the steady-state vibration response (i.e. frequency response) at several locations on a blade was measured and used to identify the prominent peaks in the frequency spectra. These were then used to generate an excitation function for the finite element modelling approach. Acceptable correlation between the measured and simulated vibration response at a number of specific locations on the blade allowed the forcing function to simulate the vibration response across the whole blade. Finally, the predicted displacement field was used to determine the vibrational energy at every point on the blade which was mapped onto a CAD representation of the blade, thereby highlighting areas on the blade that were below the defect detection threshold.
Authors: James J. Hensman, C.V. Cristodaro, Gareth Pierce, Keith Worden
Abstract: An acoustic emission test was simulated using a three point bend specimen and an artificial AE source. Waveform data was recorded as the sample was cyclically loaded in three point bending, and the cross correlation coefficient of the waveforms was used to measure the repeatability of the test. Results were twofold: the stress state of a specimen affects the ultrasonic propagation therein; and the coupling condition of a transducer may not remain constant during a test.
Showing 1 to 7 of 7 Paper Titles