Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

As transformers age, their internal condition degrades, which increases the risk of failure. To prevent these failures and to maintain transformers in good operating condition is a very important issue for utilities. Traditionally, routine preventative maintenance programs combined with regular testing were used. The change to condition-based maintenance has resulted in the reduction, or even elimination, of routine time-based maintenance. Instead of doing maintenance at a regular interval, maintenance is only carried out if the condition of the equipment requires it. Hence, there is an increasing need for better nonintrusive diagnostic and monitoring tools to assess the internal condition of the transformers. If there is a problem, the transformer can then be repaired or replaced before it fails. An extensive review is given of diagnostic and monitoring tests, and equipment available that assess the condition of power transformers and provide an early warning of potential failure.

Review of condition assessment of power transformers in service

Pattern recognition

Dissolved gas analysis (DGA) is used to assess the condition of power transformers. It uses the concentrations of various gases dissolved in the transformer oil due to decomposition of the oil and paper insulation. DGA has gained worldwide acceptance as a method for the detection of incipient faults in transformers.

A review of dissolved gas analysis measurement and interpretation techniques

A concise review is given of the progress made in the field of partial discharges (PD) at DC voltage. Although ample reference will be made to work of other authors in this field, the paper will concentrate on the progress that was made at Delft University of Technology over a period of 14 years in three PhD projects. In the first project, a start was made with the analysis of the physics of partial discharges at DC voltage and different types of PD were characterized based on parameters like time interval between PD and PD magnitude. In a second project, PD analysis was applied to HVDC apparatus and different means of classification of PD at DC voltage were proposed. In the third project, PD analysis was applied to HVDC mass-impregnated cables and test specifications were proposed. In this paper the work performed in the above three Ph.D. projects is summarized with ample reference to papers of other workers in this field. Attention is given to the mechanism of PD at DC voltage as compared to AC voltage and techniques for measurement and analysis of DC PD patterns. Examples of practical application of DC PD testing are given. Finally, some thoughts on future work are presented.

/pdf/partial-discharges-at-dc-voltage-their-mechanism-detection-2ka0gekh8r.pdf

Partial discharges at DC voltage: their mechanism, detection and analysis

This paper presents early result of an investigation of partial discharge (PD) properties in vegetable oil-impregnated insulating system The laboratory experimental investigation simulated common fault conditions in vegetable oil-impregnated paper and pressboard insulating system PD tests were carried out at elevated temperature to observe possible variation on PD behavior with temperature In order to investigate the possible differences in measured PD properties, PD data was collected using a PD detector with a variable measurement bandwidth The experimental results indicated that temperature changes may result in variation of PD properties The test results also indicated that the measured PD properties also vary with the detector bandwidth

Measurement of partial discharges in vegetable oil-impregnated insulating system

Describe investigations aimed at developing methods of testing SF/sub 6/ gas from GIS (gas-insulated switchgear) for impurities. The method involves observation and analysis of partial discharge (PD) properties of corona produced when the SF/sub 6/ sample is subjected to the nonlinear AC electric field of a point-plane test geometry. The PD properties are quite sensitive to the presence of impurities and, thus, analysis of the discharges should provide information which allows some estimation of the impurity, perhaps even species and concentration. As a preliminary to such development the authors describe a number of experiments involving PD recording and analysis in SF/sub 6/ samples contaminated by moisture. The results of the experiments (recording phase angle and magnitude of discharges) indicate that statistical analysis of the data may provide a possible means of impurity evaluation. >

Investigation of partial discharges in SF/sub 6/ using a computer-based data acquisition system

The need for on-line monitoring of high voltage cables is a matter of prime importance to the utility asset manager because of the potential disruption to consumers if a cable fails. Such monitoring is a difficult proposition and a number of different diagnostic methods have been developed. Partial discharge monitoring, although suffering from possible access problems, is potentially the most useful diagnostic as it can be used both as a monitor of instantaneous condition also as a sensitive monitor for insulation deterioration over time. A partial discharge measurement system has been developed to monitor on-line the insulation condition of high-voltage cables and their accessories. The system measures partial discharge activity using two different high frequency, non-intrusive, sensors. No outage is required for fitting the sensors to cables. Techniques to overcome the problem of significant interference on-site were investigated including choice of detection frequency by filtering, signal gating with an interference-sensitive sensor, common-mode rejection, phase-position windowing, signal discrimination based on waveshapes and detection of signal propagation direction. It was found that the discharge waveshape is a potentially useful indicator of insulation degradation.

On-line Monitoring and Assessment of Power Cable Insulation Condition Using Partial Discharges

The application of the ultra-high frequency (UHF) partial discharge (PD) detection method has proven effective in GIS. This method is now also being applied to detect PDs in oil-filled power transformers. In this paper, the localization of the PD source in transformer is discussed. This is achieved by applying an array of four sensors to capture the PD signals at the same time. To determine the PD location, two methods (first peak and cross-correlation) are used to determine the difference in time of arrival between sensor signals and followed by numerical triangulation. The sensor positions and array size influence the PD waveform patterns captured, and thus the accuracy of the PD localization. Experimental results and analysis indicate the first peak method yields better location accuracy than the cross correlation method. However, the traditional external acoustic detection method performs better than the UHF detection method under the same test configurations.

UHF sensor array for partial discharge location in transformers

The location of a partial discharge (PD) source inside a transformer can be determined from the time differences of arrival (TDOA) between signals that are captured by an array of UHF sensors. From these, the PD location can be found by geometric triangulation which involves solving a set of non-linear equations. This can be achieved using an efficient software realization of the maximum-likelihood estimator. The recorded PD waveforms are affected by the type of sensor used to capture PD signals. In this paper, the accuracy of the PD localization using different sensors is investigated. Two types of sensor, i.e. short monopole and log-spiral are used to capture electromagnetic waves emitted from a PD source. To calculate the TDOA, two methods: first peak and cross-correlation were applied. The localization result shows the monopole sensor produces higher accuracy than the log-spiral and the first peak method achieves better result than the cross-correlation method.

Trevor Blackburn

Papers

Measurement of partial discharges in vegetable oil-impregnated insulating system

Investigation of partial discharges in SF/sub 6/ using a computer-based data acquisition system

On-line Monitoring and Assessment of Power Cable Insulation Condition Using Partial Discharges

UHF sensor array for partial discharge location in transformers

Partial discharge localization in transformers using monopole and log-spiral UHF sensors