Andreas Kyprianou

Assignment of natural frequencies by an added mass and one or more springs

The problem of assigning natural frequencies to a multi-degree-of-freedom undamped system by an added mass connected by one or more springs is addressed. The added mass and stiffnesses are determined using receptances from the original system. The modifications required to assign a single natural frequency may be obtained by the non-unique solution of a polynomial equation. If more than one frequency is to be assigned, then a system of non-linear multivariate polynomial equations must be solved. Such a modification involves not only an added mass and one or more stiffness terms, but also an added coordinate. The paper presents a methodology, using Groebner bases for the solution of the multivariate polynomials, together with examples of natural frequency assignment. Realistic modifications are found to be bounded within certain frequency ranges. The effect of the modification on the natural frequencies not assigned and the antiresonances is explained.

Wavelet packet denoising for online partial discharge detection in cables and its application to experimental field results

Partial discharge measurements taken online are severely corrupted by noise due to external disturbances. In this paper a powerful noise reduction technique, based on a wavelet packet denoising algorithm, is employed to isolate the signals from the noise. This methodology enables the denoising of partial discharges that are heavily corrupted by noise without assuming any a priori knowledge about the partial discharge features. A brief description of the wavelet packet theory as an extension of the multi-resolution analysis is given. Results of the application of this algorithm to simulated data of low signal-to-noise ratio are presented, demonstrating substantial improvement in signal recovery with minimum shape distortion. Finally, the capability of this technique is highlighted by applying it to experimental field data taken from three-phase 11 kV cables.

Comparison of two partial discharge classification methods

Two signal classification methods have been examined to discover their suitability for the task of partial discharge (PD) identification. An experiment has been designed to artificially mimic signals produced by a range of PD sources that are known to occur within high voltage (HV) items of plant. The bushing tap point of a large Auto-transformer has been highlighted as a possible point on which to attach PD sensing equipment and is utilized in this experiment. Artificial PD signals are injected into the HV electrode of the bushing itself and a high frequency current transformer (HFCT) is used to monitor the current between the tap-point and earth. The experimentally produced data was analyzed using two different signal processing algorithms and their classification performance compared. The signals produced by four different artificial PD sources (surface discharge in air, corona discharge in air, floating discharge in oil and internal discharge in oil) have been processed, then classified using two machine learning techniques, namely the support vector machine (SVM) and probabilistic neural network (PNN). The feature extraction algorithms involve performing wavelet packet analysis on the PD signals recorded over a single power cycle. The dimensionality of the data has been reduced by finding the first four moments of the probability density function (Mean, Standard deviation, Skew and Kurtosis) of the wavelet packet coefficients to produce a suitable feature vector. Initial results indicate that very high identification rates are possible with the SVM able to classify PD signals with a slightly higher accuracy than a PNN.

ARIMA modeling of the performance of different photovoltaic technologies

In this paper, the performance of different technology photovoltaic (PV) systems was modeled using autoregressive integrated moving average (ARIMA) processes. Measurements from mono-crystalline (mono-c-Si), multi-crystalline (multi-c-Si) and amorphous (a-Si) silicon, cadmium telluride (CdTe) and copper indium gallium diselenide (CIGS) systems were used to construct monthly dc performance ratio (PR) time-series, from outdoor measurements. Each PR time-series was modeled a) with multiplicative ARIMA, b) with linear regression and c) with Seasonal-Trend Decomposition by Loess (STL) using the first 4 years of each time-series in order to compare the accuracy of the different methods. The models were used to forecast the PR of the 5th year of the different PV technologies and the results from the aforementioned statistical methods were compared based on the root-mean-square error (RMSE). The results showed that ARIMA produced the lowest RMSE for crystalline silicon (c-Si) technologies, whereas for thin-film technologies, STL was more accurate. The results from ARIMA also showed that thin-film technologies were optimally modeled with identical model orders, whereas for c-Si, each technology required a different optimal model order.

Differential evolution based identification of automotive hydraulic engine mount model parameters

Abstract Hydraulic engine mounts are commonly used in automotive applications, and numerical models exist for performing full-vehicle noise, vibration and harshness (NVH) studies by means of multibody simulation. The parameters of these models are usually determined by the manufacturer from first-principle numerical calculations, or by means of direct testing of the individual components. This paper describes, instead, a four-step identification method developed to determine the parameter values of a specific hydromount numerical model, the Freudenberg hydromount equations, a set of highly non-linear piecewise-continuous differential equations. The identification procedure is based on two concepts, the first being the use of the differential evolution algorithm for determining optimal parameter values, while the second is the use of data obtained from a series of experimental tests of progressively higher displacement amplitude. Identified parameters provide models whose mean square errors between the calculated output force time history and the experimentally measured force time history are typically of the order of 1-2 per cent.

Evaluation of Partial Discharge Denoising using the Wavelet Packets Transform as a Preprocessing Step for Classification

The identification of partial discharges in high voltage equipment has emerged as one of the most effective condition monitoring methods for assessing the integrity of the equipment under test. The fact that the application of PD monitoring methods is being applied online makes the measurements suffer from noise, inevitable at the measurement point, and reduces the sensitivity of the measurements. Signal processing methods to post process the measurements have been utilised, resulting not only in rejection of the noise and improvement of the sensitivity, but also in improved classification of the PD. A powerful noise rejection technique, the wavelet packets transform (WPT) has been extensively employed for the effective extraction of PD signals from noise. This technique is particularly useful in denoising signals which have transient characteristics. It expands the signal into different bases that are chosen adaptively according to a cost function, transforming the signal into a set of wavelet coefficients. The choice of a cost function has a significant effect on the compact representation of the signal. In this paper after the theory of wavelet packets is first briefly presented, and the denoising performance of the various wavelet packets parameters, such as the wavelet function, the thresholding type, and the cost function to be used is studied through the use of data acquired in a laboratory experimental environment for four types of discharges; namely the corona discharge in air, the internal discharge in oil, the floating discharge in oil and the surface discharge in air. The Symmlet wavelet has been compared with the Daubechies wavelet, both with 8 vanishing moments, the hard thresholding rule has been compared with the soft thresholding rule, and three cost functions have been compared as to their suitability for best basis expansion. Using some predefined criteria to assess their denoising performance the Symmlet 8 has been found to outperform the Daubechies 8 wavelet, the hard thresholding rule to yield better performance than the soft thresholding rule and the Shannon entropy cost function to perform better that the log energy and the norm energy cost functions.

Relations between information theory, robustness, and statistical mechanics of stochastic systems

The fundamental question, which will be addressed in this talk are the relations between dissipation, which is a concept of robustness, entropy rate, which is a concept of information theory, and statistical mechanics. Dissipation is a concept which is used in the theory and applications of robustness of filtering and control of uncertain systems. In thermodynamics, when a system is not in equilibrium with its surroundings there exists a potential of producing useful work. Dissipation is the part of this potential that is not transformable to useful work. On the other hand, entropy is fundamental concept on which information theory and in general telecommunication systems are founded on. Entropy rate is a macroscopic property of thermodynamic systems, that quantifies dissipation through the Clausius inequality and irreversible processes. In addition entropy measures the number of microstates, different configurations of the phase space, that correspond to a thermodynamic macrostate of certain entropy value. In this presentation statistical mechanics concepts will be used to bring about the close relationship between entropy and dissipation and in particular, the implication of this relationship, in computing the induced norm associated with disturbance attenuation problems.

Classification of Partial Discharge Signals using Probabilistic Neural Network

Partial Discharge (PD) classification in power cables and high voltage equipment is essential in evaluating the severity of the damage in the insulation. In this paper, the Probabilistic Neural Network (PNN) method is used to classify the PDs. After the algorithm has been trained it uses the input vector, which contains the features that would be used for classification, to calculate the probability density function (pdf) of each class and together with the assignment of a cost for a misclassification the decision that minimizes the expected risk is taken. The maximum likelihood training is employed here. The success of this particular method for classification is asserted. This method has the advantage over Multilayer Neural Network that it gives rapid training speed, guaranteed convergence to a Bayes classifier if enough training examples are provided (i.e. it approaches Bayes optimality), incremental training which is fast (i.e. additionally provided training examples can be incorporated without difficulties) and robustness to noisy examples. The results obtained here (99.3%, 84.3% and 85.5% for the corona, the floating in oil and the internal discharges respectively) are very encouraging for the use of PNN in PD classification.

A Scale-Adaptive Approach for Spatially-Varying Urban Morphology Characterization in Boundary Layer Parametrization Using Multi-Resolution Analysis

Urban morphology characterization is crucial for the parametrization of boundary-layer development over urban areas. One complexity in such a characterization is the three-dimensional variation of the urban canopies and textures, which are customarily reduced to and represented by one-dimensional varying parametrization such as the aerodynamic roughness length