Emmanouil Styvaktakis

Expert System for Classification and Analysis of Power System Events

This paper presents an expert system that is able to classify different types of power system events to the underlying causes (i.e., events) and offer useful information in terms of power quality. The expert system uses the voltage waveforms and distinguishes the different types of voltage dips (fault-induced, transformer saturation, induction motor starting), as well as interruptions (nonfault, fault induced). A method for event-based classification is used, where a segmentation algorithm is first applied to divide waveforms into several possible events. The expert system is tested using real measurements and the results show that the system enables fast and accurate analysis of data from power quality monitors.

Categorization and analysis of power system transients

Power system transients are power-quality disturbances that can be harmful to electronic equipment. This paper contributes and provides some solutions to the following issues: 1) to introduce a new way to identify different categories of power system transients based on their underlying causes; 2) to propose a model and analysis tool for oscillatory transients, where emphasis is on finding phenomena and characteristics associated with the underlying causes of transients. A model-based approach, ESPRIT, is applied to a number of simulated voltage waveforms to extract the parameters of oscillatory transients, and the results may be used for identifying and understanding the causes of transients by correlating the major components in transients with the phenomena that may appear in different types of transients and some a priori knowledge of power system settings.

Classification of underlying causes of power quality disturbances: deterministic versus statistical methods

This paper presents the two main types of classification methods for power quality disturbances based on underlying causes: deterministic classification, giving an expert system as an example, and statistical classification, with support vector machines (a novel method) as an example. An expert system is suitable when one has limited amount of data and sufficient power system expert knowledge; however, its application requires a set of threshold values. Statistical methods are suitable when large amount of data is available for training. Two important issues to guarantee the effectiveness of a classifier, data segmentation, and feature extraction are discussed. Segmentation of a sequence of data recording is preprocessing to partition the data into segments each representing a duration containing either an event or a transition between two events. Extraction of features is applied to each segment individually. Some useful features and their effectiveness are then discussed. Some experimental results are included for demonstrating the effectiveness of both systems. Finally, conclusions are given together with the discussion of some future research directions.

Bridge the gap: signal processing for power quality applications

Abstract With an increasing amount of measurement data, automating power quality characterization and classification of disturbances is desirable. This will require combined efforts and knowledge from both electric power systems and signal processing. This paper focuses on several major power quality problems, and some up-to-date signal processing techniques that may offer good solutions to the problems. Some examples and results are also included. Finally, some future research directions are discussed.

Automatic classification of power system events using RMS voltage measurements

Power quality monitors in the occasion of a disturbance can either save the actual voltage waveform that contains the event or the corresponding RMS. The latter option reduces significantly the memory that is needed for saving the event. This paper shows that even with this type of monitoring, analysis of the measurements can be in depth. The paper proposes a method for automatic classification of power system events using RMS voltage measurements. The system is tested with measurements from a distribution network and the results show that classification is possible for the considered types of events. Finally, the limitations of this type of monitoring are shown.

Voltage dip detection and power system transients

Recently developed power quality mitigation equipment, like the static transfer switch, needs methods for fast and reliable detection of voltage dips. Such a detection scheme must be able to detect a voltage dip as fast as possible and be immune to other types of disturbances. In this paper, the authors address the problem of voltage dip detection regarding Kalman filtering, the characteristics of fault-induced voltage dips and other power system disturbances. They investigate how the voltage dip characteristics influence the speed of detection and show that disturbances other than fault-induced dips could trigger a detection scheme. Special attention is given to trans former-related events. Their characteristics are presented using measurements. Kalman filtering modelling issues are discussed. Statistics on the characteristics of fault-induced dips and transformer events are presented from medium voltage networks.

Classification of power system events: voltage dips

The extensive monitoring programs that are run by power utilities enable new insights of the power system operation and new characterisation methods must be used for the classification and analysis of the recordings. This paper focuses on events that cause a temporary decrease in the fundamental frequency voltage magnitude (voltage dip). The analysis of the recordings from surveys in medium and low voltage networks shows that new classes of voltage dips should be introduced in order to characterise the events that are captured by the power quality monitors. The aim of the paper is to show the distinctive characteristics of each class and give the guidelines for the automatic processing of the recordings. Finally, a large number of recordings from a medium voltage network are classified using these characteristics and the results are presented.

Expert system for voltage dip classification and analysis

The increasing amount of data obtained by power quality monitors and the need for better understanding of power system disturbances require new analysis tools. This paper presents an expert system that is able to classify different types of voltage dips according to the underlying causes (i.e. events) and offer useful information in terms of power quality. The expert system uses the voltage waveforms and distinguishes the different types of voltage dips (fault-induced, transformer saturation, induction motor starting), explains the changes in the voltage dip magnitude (change in the system, change in the fault type, transformer saturation, motor load influence) and separates interruptions into non-fault and fault-induced. A method is proposed for event-based classification, where a segmentation algorithm is first applied to divide waveforms into several possible events. Kalman filtering is employed to model the waveforms and the residuals of the model are used for segmentation. The expert system is tested using real measurements and the results show that the system enables fast and accurate analysis of data from power quality monitors.

A statistical-based sequential method for fast online detection of fault-induced voltage dips

This paper addresses the problem of detecting voltage dips regarding measurements consisting of fault events, transformer saturation events, and capacitor-switching events. A novel statistical-based sequential detection method is proposed for online classification of these events. The detector is based on the Neyman-Pearson criterion that maximizes the detection rate of fault-induced dips with constrained false alarm rate of the other two types of event. The sequential detector is able to give an earliest possible event discrimination together with the estimated confidence at the time instant ranging from 1/8,1/4,1/2, to 3/4 cycle of the fundamental frequency after detecting an initial voltage drop at 0.95 p.u. The performance of the proposed scheme is evaluated using measurements from medium voltage networks.

A fault location technique using high frequency fault clearing transients

This letter suggests that a voltage recorder, placed next to a circuit breaker not as usual on the side of the substation, but on the side of the transmission line, may reveal the location of permanent faults. Two different approaches to estimating fault location are presented here: spectrum estimation and wavelet analysis. The methods are tested and compared with simulations of typical transmission systems using the Electromagnetic Transients Program (EMTP).