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

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

As one step toward the future smart grid, condition monitoring is important to facilitate the reliability of grid asset operation and to save on maintenance cost [1]. Most failures of the power grid are caused by electrical insulation failure, and a key indicator of such electrical failure is the occurrence of partial discharge (PD). Therefore, one focus of condition monitoring is to detect PD, especially in the early stages, to prevent a serious power failure or outage.

An overview of state-of-the-art partial discharge analysis techniques for condition monitoring

As a non-renewable resource, mineral oil has been re-evaluated for its value and suitability as the coolant and insulator in transformers because of its environmental unfriendliness, fire hazards, and operating efficiency dissatisfaction. As a result, alternative dielectric liquids meeting the demands for more reliable, safer, and cleaner energy has been a critical need. Plant-based insulating fluids, non-toxic to the environment and ultimately biodegradable, have been flowing into the mainstream for achieving cost efficiencies, performance advantages, and enhancing safety. With more than 20 years of field experience, plant-based oils are now employed in more than 600,000 transformers worldwide, and have been established an enviable performance track record. Far from being just a mineral oil replacement, but more significantly, plant-based oils have filled a gap in some specific application scenarios where mineral oils fail to satisfy the fire-safety and environmental standards. This article reviews the research status of plant-based insulating fluids for transformer, including both challenges and outlook circa 2020. Plant-based oils have many inherent advantages over mineral oil, but also have numerous and unique challenges.

A critical review of plant-based insulating fluids for transformer: 30-year development

This paper presents a state-of-the-art review on machine learning (ML) based intelligent diagnostics that have been applied for partial discharge (PD) detection, localization, and pattern recognition. ML techniques, particularly those developed in the last five years, are examined and classified as conventional ML or deep learning (DL). Important features of each method, such as types of input signal, sampling rate, core methodology, and accuracy, are summarized and compared in detail. Advantages and disadvantages of different ML algorithms are discussed. Moreover, technical roadblocks preventing intelligent PD diagnostics from being applied to industry are identified, such as insufficient/imbalanced dataset, data inconsistency, and difficulties in cost-effective real-time deployment. Finally, potential solutions are proposed, and future research directions are suggested.

Condition Monitoring Based on Partial Discharge Diagnostics Using Machine Learning Methods: A Comprehensive State-of-the-Art Review

Dissolved gas-in-oil analysis (DGA) is a sensitive and reliable technique for the detection of incipient fault condition within oil-immersed transformers. The presence of certain key gases is monitored and quantified. There are a number of methods developed for analyzing these gases and interpreting their significance: Key Gas, Rogers Ratio, Doernenburg, Logarithmic Nomograph, IEC Ratio and Duval Triangle. This paper investigates the accuracy and consistency of these methods in interpreting the transformer condition. The evaluation is carried out on DGA data obtained from the local power utilities and from published papers. The data consists of 92 different cases. The key gases considered are hydrogen, methane, ethane, ethylene and acetylene. A MATLAB program was developed to automate the evaluation of the methods.

Comparative Study and Analysis of DGA Methods for Transformer Mineral Oil

Innovations in computer technology have made possible continuous on-line monitoring of partial discharge (PD) activities. The power industry aims to assess the condition of power system equipment through on-line monitoring of PD activities. This involves long-term continuous data recording and it is very difficult to extract useful information from such a large amount of raw data, particularly if it is done manually. Instead, data mining can be applied in solving this problem. Data mining can be categorized into predictive modelling and descriptive modelling. In this paper, work was mainly focused on predictive data mining, which is classification of PD. The back propagation neural network (BPN), self-organizing map (SOM) and support vector machine (SVM) were used for classification and compared. Results indicate SVM is the best method in terms of classification accuracy and processing speed.

Application of data mining on partial discharge part I: predictive modelling classification

Polarization and Depolarization Current (PDC) testing is a non-destructive dielectric testing method to determine the conductivity and moisture content of transformer insulations. This research was carried out to find the dielectric responsive function and maximum conductivities of biodegradable and mineral transformer oils for comparative analysis. Each insulation oil with different moisture levels (dried, normal, or wet) was tested. The dried condition was achieved by processing the oil in a dry-out plant to evacuate the moisture. Normal condition refers to the oil as taken directly from the drum while wet condition is the oil after 200ppms of moisture have been added. The polarization and depolarization current at test DC voltage of 1000V was recorded for 10,000 seconds. Each test was repeated five to ten times to collect the most consistent PDC data and to check the repeatability of each PDC pattern. In the analysis, the PDC curve was plotted and the DC conductivity of the oil was calculated. It was found that the trends of the PDC response for the two transformer oils are different.

Polarization and Depolarization Current (PDC) tests on biodegradable and mineral transformer oils at different moisture levels

Partial discharge (PD) is a common phenomenon which occurs in insulation of high voltage equipments such as transformers and has a damaging effect to the insulation. In this paper, the application of descriptive data mining on PD occuring in insulating system is shown. Experiments were set up to create three basic types of PD: corona, surface discharges and internal discharges. Partial discharge data were analysed using phase resolved analysis and pulse height analysis. Descriptive data mining was applied on the collected data using decision tree with genetic algorithm (GA) to mine the rules/relationships which can be used to differentiate the PD. These extracted rules are useful as input to predictive data mining such as fuzzy logics.

Descriptive data mining of partial discharge using decision tree with genetic algorithm

This paper presents results of experimental tests performed on laboratory models of transformer windings immersed in bio-degradable oil. The investigations covered three common types of transformer faults: partial discharge (PD), arcing and overheating. Dissolved hydro-carbon gas products generated by these three fault types in bio-degradable oil are analyzed and compared with existing dissolved gas analysis (DGA) interpretation methods for mineral oil insulation. The PD and arcing tests were performed in small test cells using mineral and vegetable oils, with and without insulated pressboards. The overheating test was performed by heating the oils in the oven. At the end of each test, an oil sample was extracted for dissolved gas and moisture measurement. The results show that existing DGA interpretation techniques need to be modified if they are to be used for insulation assessment of bio-degradable oil.

K.X. Lai

Papers

Comparative Study and Analysis of DGA Methods for Transformer Mineral Oil

Application of data mining on partial discharge part I: predictive modelling classification

Polarization and Depolarization Current (PDC) tests on biodegradable and mineral transformer oils at different moisture levels

Descriptive data mining of partial discharge using decision tree with genetic algorithm

Dissolved gas analysis of faults in biodegradable oil transformer insulating systems