Просмотр полной версии : Transformer diagnostics

16.07.2012, 13:33

Long term transformer operation has revealed typical forms of damage, their signs, possible causes and methods of detection.

The core.

Defects in the interlayer insulation may cause overheating due to local currents or currents in short-circuited loops, occurring as a result of insulation breach between massive body components and core steel. If moisture condenses on the surface of the oil, it appears on the upper yoke, enters the gaps between plates of core steel as oil emulsion, destroys interleave insulation and causes steel corrosion. Oil quality suffers (flashpoint is reduced, acidity is increased) and the no-losses are increased.


The most common type of damage in the coils are turn-to-turn short-circuits. These may be caused by insulation degradation due to natural wear or long transformer overloads with insufficient cooling of the coils. Inter-coil insulation can also be degraded by mechanical damage during short-circuits. Signs of turn-to-turn short-circuits are tripping of gas protection, increased heating, variation in phase impedance to direct current etc.

Transformers of 1000 kVA and above feature a gas relay; the relay is tripped by gases emitted by decomposing oil as a result of the abovementioned damage. Analysis of the gas accumulated in the relay allows to assess the reasons for protection tripping and the nature of damage; this allows to detect the damage in the early stage and, in some cases, remedy it quickly.

The described methods of integral control of transformer’s insulation (insulation impedance, absorption ratio, tan delta etc.) do not allow to determine nature and extent of partial damage to insulation in the early stages. Therefore, one of the most accurate damage detection methods is timely gas chromatography analysis of gases dissolved in the oil.

At this time, the following connections have been made between the gases emitted into the oil and the causes of their generation. Thus, hydrogen indicates sparking and arcing partial discharges; acetylene indicates arcing and sparking; ethylene indicates local oil and cellulose insulation heating above 873oK; methane indicates local insulation heating in the range of 673 – 873oK or partial discharges that go with the heating; ethane indicates local oil and insulation heating to 573 – 673oK; carbon monoxide and dioxide indicate aging and moisturizing of oil and solid insulation; carbon dioxide indicates heating of solid insulation.

Apart from the mentioned gases, the oil may contain oxygen (air), which means that there is a breach in the transformer, and water, which degrades insulation performance of transformer oil.

Methods for taking oil samples and analysis of water and solved gases are numerous. However, the main drawback of the gas chromatography method is the inability to run online tests: ittakes time to transport the sample to the gas chromatographer, while the test itself along with preparation of the chromatographer takes 2 to 3 hours.

Principle of current continuous diagnostic units is based on continuous measurement of all dissolved gas volume or determination of its volumetric impedance. In air tight transformers and high voltage oil filled taps, oil pressure and temperature measurement systems are used.

Such systems are manufactured, for example, by Syprotec Corporation, USA. Their line of Hydran systems of various models are connected directly to the transformer. They measure the total concentration of volatile gas and recalculate it in hydrogen equivalent. The mathematical programming of the systems allows analyzing the data and forecasting development of damage which may lead to transformer failure.

A subsystem based on the mathematical model of the transformer’s loading capacity, which does not require installation of sensors inside the transformer, can be a part of the diagnostics system. This device requires transformer load data (usually, the existing measurement system has the transformer phase power sensors already installed), voltage of the transformer and ambient temperature. Also, no-load losses and short-circuit losses and nominal coil and upper oil layer temperatures are required.

This integral insulation wear assessment subsystem supplies real time data on insulation wear and forecasts transformer life time. This information, combined with scheduled insulation parameters checks (insulation impedance, absorption coefficient etc.) allow to make repairs depending on the actual wear of transformer insulation.