Cellulosic paper and oil insulation in a transformer degrade at higher operating temperatures. Degradation is accelerated in the presence of oxygen and moisture. Power transformers being expensive items need to be carefully monitored throughout their operation. Well established time-based maintenance and conservative replacement planning is not feasible in a current market driven electricity industry. Condition based maintenance and online monitoring are now gaining importance. Currently there are varieties of chemical and electrical diagnostic techniques available for insulation condition monitoring of power transformers. This paper presents a description of commonly used chemical diagnostics techniques along with their interpretation schemes. A number of new chemical techniques are also described in this paper. A number of electrical diagnostic techniques have gained exceptional importance to the utility professionals. Among these techniques polarisation/depolarisation current measurement, return voltage measurement and frequency domain dielectric spectroscopy at low frequencies are the most widely used. This paper describes analyses and interpretation of these techniques for transformer insulation condition assessment.

/pdf/review-of-modern-diagnostic-techniques-for-assessing-1y1f1wjvi7.pdf

Review of modern diagnostic techniques for assessing insulation condition in aged transformers

Large amounts of data are increasingly accumulated in the energy sector with the continuous application of sensors, wireless transmission, network communication, and cloud computing technologies. To fulfill the potential of energy big data and obtain insights to achieve smart energy management, we present a comprehensive study of big data driven smart energy management. We first discuss the sources and characteristics of energy big data. Also, a process model of big data driven smart energy management is proposed. Then taking smart grid as the research background, we provide a systematic review of big data analytics for smart energy management. It is discussed from four major aspects, namely power generation side management, microgrid and renewable energy management, asset management and collaborative operation, as well as demand side management (DSM). Afterwards, the industrial development of big data-driven smart energy management is analyzed and discussed. Finally, we point out the challenges of big data-driven smart energy management in IT infrastructure, data collection and governance, data integration and sharing, processing and analysis, security and privacy, and professionals.

Big data driven smart energy management: From big data to big insights

Building energy use is expected to grow by more than 40% in the next 20 years. Electricity remains the largest energy source consumed by buildings, and that demand is growing. To mitigate the impact of the growing demand, strategies are needed to improve buildings' energy efficiency. In residential buildings home appliances, water, and space heating are answerable for the increase of energy use, while space heating and other miscellaneous equipment are behind the increase of energy utilization in non-residential buildings. Building energy management systems support building managers and proprietors to increase energy efficiency in modern and existing buildings, non-residential and residential buildings can benefit from building energy management system to decrease energy use. Base on the type of building, different management strategies can be used to achieve energy savings. This paper presents a review of management strategies for building energy management systems for improving energy efficiency. Different management strategies are investigated in non-residential and residential buildings. Following this, the reviewed researches are discussed in terms of the type of buildings, building systems, and management strategies. Lastly, the paper discusses future challenges for the increase of energy efficiency in building energy management system.

A review of strategies for building energy management system: Model predictive control, demand side management, optimization, and fault detect & diagnosis

Preventive diagnosis and maintenance of transformers have become more and more popular in recent times in order to improve the reliability of electric power systems. Dielectric testing techniques such as return voltage measurement (RVM) and polarization-depolarization current (PDC) measurement are being investigated as potential tools for condition assessment of transformer insulation. A better understanding and analysis of the dielectric test results are only possible with a clear understanding of the physical behavior of the insulation system in response to moisture and aging. A circuit model, which describes the dielectric behavior of the transformer's main insulation system, has been parameterized in this paper. The values of the parameters of the model have been identified from the dielectric tests. A correlation has been developed between the physical condition of the insulation and the equivalent model parameters that enable a clear and transparent interpretation of the dielectric test results.

https://espace.library.uq.edu.au/view/UQ:9899/saha-raj-muller-.pdf

Deriving an equivalent circuit of transformers insulation for understanding the dielectric response measurements

The load data in smart grid contains a lot of valuable knowledge, which is useful for both electricity producers and consumers. Load classification is an important issue in load data mining. A five-stage process model of load classification is constructed based on the summary and analysis of studies about load classification in smart grid environment. Then, the commonly used clustering methods for load classification are summarized and briefly reviewed, and the well-known evaluation methods for load classification are also introduced. Besides, the applications of load classification, including bad data identification and correction, load forecasting and tariff setting, are discussed. Finally, an example of load classification based on Fuzzy c-means (FCM) is presented.

A review of electric load classification in smart grid environment

Building load management using cluster and statistical analyses

The determination of the polarization spectrum in the low frequency region of the dissipation factor, commonly has been used in practice to diagnose insulation conditions. The recovery voltage method (RVM) is one of the methods used to estimate the polarization spectrum. Nevertheless, current analysis of the RVM tests do not use all the information available from the test. In this paper, a mathematical model of the polarization spectrum is proposed. It is shown that, from the derived model, some of the hidden RVM peaks can be identified. Since the difficulty to derive a model from RVM data increases with the number of the peaks to be identified, a simpler model is always preferable. A scheme for determining the parameters of an equivalent circuit is also given, The composite oil/paper insulation is considered which is extensively used in cables and transformers. Cellulose paper, such as the Manila-Kraft type, which is used as insulation, exerts a major influence on the life of the equipment because of its hygroscopic nature. The importance of detecting and eliminating moisture from paper is crucial in oil-impregnated HV insulation. Analysis of the polarization spectrum has been used by several utilities to determine moisture in the paper. Recently, some test results indicated that the polarization spectrum may not clearly show important peaks corresponding to local moisture ingress and thus can mislead experts in the field, The developed model has been validated using RVM data of real power transformers.

Modeling the polarization spectrum in composite oil/paper insulation systems

Fuzzy detection of high impedance faults in radial distribution feeders

A class of hybrid intelligent system for fault diagnosis in electric power systems

International Standards ANSI C84.1-2006 and IEEE std 1250-1995 specify the range of supply voltage to electronics equipment from 0.9 to 1.05 per-unit (pu) of nominal voltage. From this information, it is possible to feed electronics equipment with a voltage level below of nominal voltage without failures. The Conservation Voltage Reduction (CVR) is a technique that aims to achieve the reduction of power consumption by reducing the level of supply voltage, giving at the customer location a lowest possible level of voltage compatible with both nominal values of equipment and the levels stipulated by the regulatory agency. It is essential obtaining a good model that describes the behavior of the loads connected to grid to evaluate the effectiveness of the CVR application. This paper presents an evaluation of three static load model techniques, being two ZIP load models and an exponential model. The model parameters were estimated for samples obtained from measurements performed on residential/commercial loads, through the variation of supply voltage within the operational acceptable range. The data collection was performed at three different periods of times to mitigate the grid's influence on the results. The load models are obtained through an optimization algorithm based on the collected data. The ellipsoidal optimization algorithm was applied to estimate the parameters of each model, searching for the best feasible solution. Some papers in the literature exhibit mathematical solutions for load models without physical meaning. Unlike that, this algorithm is not limited only to find a mathematical solution but also a physical one.

Patrícia Romeiro da Silva Jota

Papers

Building load management using cluster and statistical analyses

Modeling the polarization spectrum in composite oil/paper insulation systems

Fuzzy detection of high impedance faults in radial distribution feeders

A class of hybrid intelligent system for fault diagnosis in electric power systems

Comparison between static models of commercial/residential loads and their effects on Conservation Voltage Reduction