P L Lewin

Partial discharge source discrimination using a support vector machine

Partial discharge (PD) measurements are an important tool for assessing the health of power equipment. Different sources of PD have different effects on the insulation performance of power apparatus. Therefore, discrimination between PD sources is of great interest to both system utilities and equipment manufacturers. This paper investigates the use of a wide bandwidth PD on-line measurement system consisting of a radio frequency current transducer (RFCT) sensor, a digital storage oscilloscope and a high performance personal computer to facilitate automatic PD source identification. Three artificial PD models were used to simulate typical PD sources which may exist within power system apparatus. Wavelet analysis was applied to pre-process measurement data obtained from the wide bandwidth PD sensor. This data was then processed using correlation analysis to cluster the discharges into different groups. A machine learning technique, namely the support vector machine (SVM) was then used to identify between the different PD sources. The SVM is trained to differentiate between the inherent features of each discharge source signal. Laboratory experiments where the trained SVM was tested using measurement data from the RFCT as opposed to conventional measurement data indicate that this approach has a robust performance and has great potential for use with field measurement data.

Comparison of on-line partial discharge detection methods for HV cable joints

The capacitive coupler, acoustic emission (AE) sensor and radio frequency current transducer (RFCT) have been used to detect partial discharge (PD) activity within a 132 kV prefabricated cable joint containing a known defect. Although each of the three methods has been applied individually under different situations, a comprehensive investigation and comparison of these three methods has not yet been made. Results obtained were compared to measurements made using the conventional electrical detection method. The quantification of the capacitive coupler measured signal in mV with a discharge apparent quantity in pC has been investigated and an on-line quantification method proposed. The capacitive coupler has good detection sensitivity and PD location can be realised by studying the time of Right between signals from two sensors. The difference in pulse shape, spectrum and time of flight between an internal discharge and external interference has been investigated. The AE approach has the advantage of being free from electrical interference. However, investigation indicated that AE signals were significantly attenuated within the cable joint. RFCTs were used to detect the discharge current flowing through the cable sheath. Where the detection sensitivity was low, a wavelet de-noising method was applied to process the RFCT signals and proved to tie effective in increasing the measurement signal to noise ratio.

Discrimination of multiple PD sources using wavelet decomposition and principal component analysis

Partial discharge (PD) signals generated within electrical power equipment can be used to assess the condition of the insulation. In practice, testing often results in multiple PD sources. In order to assess the impact of individual PD sources, signals must first be discriminated from one another. This paper presents a procedure for the identification of PD signals generated by multiple sources. Starting with the assumption that different PD sources will display unique signal profiles this will be manifested in the distribution of energies with respect to frequency and time. Therefore the technique presented is based on the comparison of signal energies associated with particular wavelet-decomposition levels. Principal component analysis is adopted to reduce the dimensionality of the data, whilst minimizing lost information in the data concentration step. Physical parameters are extracted from individual PD pulses and projected into 3-dimensional space to allow clustering of data from specific PD sources. The density-based spatial clustering of applications with noise (DBSCAN) clustering algorithm is chosen for its ability to discover clusters of arbitrary shape in n-dimension space. PD data from individual clusters can then be further analyzed by projecting the clustered data with respect to the original phase relationship. Results and analysis of the technique are compared for experimentally measured PD data from a range of sources commonly found in three different types of high voltage (HV) equipment; ac synchronous generators, induction motors and distribution cables. These experiments collect data using varied test arrangements including sensors with different bandwidths to demonstrate the robustness and indicate the potential for wide applicability of the technique to PD analysis for a range of insulation systems.

Feasibility of Iterative Learning Control Mediated by Functional Electrical Stimulation for Reaching After Stroke

Background. An inability to perform tasks involving reaching is a common problem following stroke. Evidence supports the use of robotic therapy and functional electrical stimulation (FES) to reduce upper limb impairments, but current systems may not encourage maximal voluntary contribution from the participant because assistance is not responsive to performance. Objective. This study aimed to investigate whether iterative learning control (ILC) mediated by FES is a feasible intervention in upper limb stroke rehabilitation. Methods. Five hemiparetic participants with reduced upper limb function who were at least 6 months poststroke were recruited from the community. No participants withdrew. Intervention. Participants undertook supported tracking tasks using 27 different trajectories augmented by responsive FES to their triceps brachii muscle, with their hand movement constrained in a 2-dimensional plane by a robot. Eighteen 1-hour treatment sessions were used with 2 participants receiving an additional 7 treatment sessions. Outcome measures. The primary functional outcome measure was the Action Research Arm Test (ARAT). Impairment measures included the upper limb Fugl— Meyer Assessment (FMA), tests of motor control (tracking accuracy), and isometric force. Results. Compliance was excellent and there were no adverse events. Statistically significant improvements were measured (P ≤ .05) in FMA motor score, unassisted tracking for 3 out of 4 trajectories, and in isometric force over 5 out of 6 directions. Changes in ARAT were not statistically significant. Conclusion. This study has demonstrated the feasibility of using ILC mediated by FES for upper limb stroke rehabilitation.

Norm-Optimal Iterative Learning Control Applied to Gantry Robots for Automation Applications

This paper is concerned with the practical implementation of the norm-optimal iterative learning control (NOILC) algorithm. Here, the complexity of this algorithm is first considered with respect to real-time control applications, and a new modified version, fast norm-optimal ILC (F-NOILC), is derived for this application, which potentially allows implementation with a sampling rate three times faster that the original algorithm. A performance index is used to assess the experimental results obtained from applying F-NOILC to an industrial gantry robot system and, in particular, the effects of varying the parameters in the cost function, which is at the heart of the norm-optimal approach

Tracking and surface discharge at the oilߝpressboard interface

A different approach to the study of surface tracking reveals a new view of the oil-pressboard interface and suggests a link between the electric double layer and the boundary layer.

A model of the upper extremity using FES for stroke rehabilitation

A model of the upper extremity is developed in which the forearm is constrained to lie in a horizontal plane and electrical stimulation is applied to the triceps muscle. Identification procedures are described to estimate the unknown parameters using tests that can be performed in a short period of time. Examples of identified parameters obtained experimentally are presented for both stroke patients and unimpaired subjects. A discussion concerning the identifications repeatability, together with results confirming the accuracy of the overall representation, is given. The model has been used during clinical trials in which electrical stimulation is applied to the triceps muscle of a number of stroke patients for the purpose of improving both their performance at reaching tasks and their level of voluntary control over their impaired arm. Its purpose in this context is threefold: Firstly, changes occurring in the levels of stiffness and spasticity in each subjects arm can be monitored by comparing frictional components of models identified at different times during treatment. Secondly, the model is used to calculate the moments applied during tracking tasks that are due to a patients voluntary effort, and it therefore constitutes a useful tool with which to analyze their performance. Thirdly, the model is used to derive the advanced controllers that govern the level of stimulation applied to subjects over the course of the treatment. Details are provided to show how the model is applied in each case, and sample results are shown.

Numerical modelling of negative discharges in air with experimental validation

Axisymmetric finite element models have been developed for the simulation of negative discharges in air without and with the presence of dielectrics. The models are based on the hydrodynamic drift-diffusion approximation. A set of continuity equations accounting for the movement, generation and loss of charge carriers (electrons, positive and negative ions) is coupled with Poissons equation to take into account the effect of space and surface charges on the electric field. The model of a negative corona discharge (without dielectric barriers) in a needle-plane geometry is analysed first. The results obtained show good agreement with experimental observations for various Trichel pulse characteristics. With dielectric barriers introduced into the discharge system, the surface discharge exhibits some similarities and differences to the corona case. The model studies the dynamics of volume charge generation, electric field variations and charge accumulation over the dielectric surface. The predicted surface charge density is consistent with experimental results obtained from the Pockels experiment in terms of distribution form and magnitude.

Partial discharge behavior within a spherical cavity in a solid dielectric material as a function of frequency and amplitude of the applied voltage

Modeling of the partial discharge (PD) process allows a better understanding of the phenomena. In this paper, a simulation model for spherical cavities within a homogeneous dielectric material has been developed. The model is implemented using Finite Element Analysis (FEA) software in parallel with a mathematical package. This method provides many advantages over previous PD models because discharge events can be simulated dynamically and the electric field in the cavity can be calculated numerically. The model has been used to study the effect of different amplitudes and frequencies of the applied voltage and simulation results have been compared with experimental measurement results. It is found that certain model parameters are dependent on the applied stress and parameters that clearly affect PD activity can be readily identified, these parameters include; the electron detrapping time constant, the cavity surface conductivity, the initial electron generation rate and the extinction voltage. The influence of surface charge decay through conduction along the cavity wall on PD activity has also been studied.

Iterative Learning Control for Multiple Point-to-Point Tracking Application

This paper considers a general class of linear iterative learning control (ILC) algorithm applied to tracking tasks which require the plant output to reach given points at predetermined time instants, without the specification of intervening reference points. A framework is developed in the frequency-domain in which the reference is updated between trials. It is shown that superior convergence and robustness properties are obtained compared with those associated with using the original class of ILC algorithm to track a prescribed arbitrary reference trajectory satisfying the point-to-point output constraints. Experimental results using a non-minimum phase test facility are presented to illustrate the theoretical findings.