A bird’s-eye view of the rise of transformer electrical, dielectric testing

Electrical Tester – 8 April 2020

Author: Jill Duplessis 

Transformers are a marvel of physics and engineering. They allow one continuous energy path to meet the needs of many different segments; they provide a link, for example, between transmission circuits that operate most efficiently at high voltages and distribution circuits that necessarily operate at medium voltage. Transformers do this within an impressively small footprint and with remarkably low intrinsic losses.

Megger, a true pioneer in dielectric testing, has been associated with transformers from almost the very beginning. The first power transformers appeared around the middle of the 1880s and this was soon followed by the first practical insulation test set, invented in 1895 by Megger’s founder Sydney Evershed. This test set produced a DC test voltage that was high enough to allow the direct measurement of insulation resistance values in the megohm range. This groundbreaking ‘megohm meter’ was soon being described as a Megger, and this name became a registered trademark on May 25, 1903. 

Fast forward almost a century and a quarter, and Megger now offers one of the broadest ranges of transformer test equipment out there, including single-function, multifunction and comprehensive vehicle-based solutions. The company’s instruments cover both screening and diagnostic tests for transformers and demonstrate Megger’s commitment to helping ensure that these important assets deliver the best possible performance and reliability. 

Offering such a wide range of solutions positions the company in a unique place of neutrality that allows us to share without prejudice the strengths and weaknesses of each type of test, and how the tests can best be used in combination. This underpins exceptional insight into transformers, which is immensely important as the electrical industry moves more and more toward integrating data from multiple test sources (e.g., electrical/oil/infrared, deenergized/energized/continuous, etc.) and must know how to manage this data responsibly. 

Megger test instruments are invaluable at many stages in the lifecycle of a transformer, including, for example: 

  • As a component test solution prior to the transformer’s factory assembly 
  • After factory assembly and during dryout, to efficiently identify the dryout completion point 
  • During manufacturer testing 
  • Prior to despatch from the manufacturer’s site and again upon arrival at its destination 
  • After field assembly and prior to energization (acceptance tests) 
  • During planned outages throughout the transformer service life (screening tests) 
  • When problems arise (diagnostic tests) 
  • After a failure, to identify the cause (forensic tests) 

Screening and diagnostic tests provide information about all three of the transformer’s health indices: 

  • Dielectric – the ability of the transformer’s insulating materials to support an electric field and thereby provide electrical isolation of energized parts from other parts that are operating at different voltages or that are maintained at ground potential 
  • Mechanical – this relates to the ability of a transformer to carry current, and therefore also encompasses magnetic health 
  • Thermal – the ability of the transformer to dissipate heat safely and effectively 

A screening test is a first level test which is performed routinely as a precautionary measure to verify that one (or more) of the transformer’s health indices is within acceptable limits. The intention is to catch problems early in their development before other, more overt symptoms develop. Early detection of problems provides opportunities to manage the asset from a more informed perspective. This, in turn, creates the possibility of minimising the impact of undesirable situations by performing maintenance or repairs or by operating the asset at reduced capacity. 

In many cases, screening tests are also performed as commissioning tests to establish benchmark records for the asset when it is new, for the benefit of future comparative reference, and to screen for unexpected problems introduced during manufacture or installation. 

A diagnostic test (pinpointing test) is typically conclusive – that is, it indicates what is wrong rather than providing a general indication that something is wrong. When screening tests are performed routinely, it is usually not necessary to routinely carry out the corresponding diagnostic tests. 

Dielectric frequency response (DFR) is a good example of a diagnostic test. It provides data about the percentage of moisture in the paper, the oil conductivity and the insulation thermal characteristics. DFR testing is used in conjunction with a database and computer-aided assessment, which are specific to the DFR test instrument purchased. 

Accurate knowledge of the moisture content and oil conductivity is, however, only needed when either or both are at unacceptable levels. This makes the alternative technique of narrowband dielectric frequency response (NB DFR) an excellent choice for screening purposes, as it has excellent problem detection capabilities and provides results that can be analysed simply by inspection. A NB DFR test, for example, will alert users to unacceptable levels of moisture and/or oil conductivity, which may indicate the need for further investigation using a full DFR test, with which one can then further pinpoint the offending contaminant and quantify levels present. 

As Figure 1 shows, today’s electrical, dielectric screening tests for transformers and accessories are far more interesting than the traditional ‘tale of two tests’ – DC insulation resistance testing and power factor/dissipation factor/capacitance testing. 

DC insulation resistance testing was the world’s first dielectric screening test. It was, however, closely followed by power factor/dissipation factor (PF/DF) and capacitance testing which was in use in cable manufacturers’ laboratories by the early 1900s. PF/DF and capacitance tests grew in popularity with the introduction of capacitance-graded bushings. The construction of these bushings poses challenges to testing with DC sources, whereas PF/DF is successful in finding localized problems with this type of bushing, which previously could not be detected. 

In around 1920 came the transformer turns ratio (TTR) test. This is very much a functional test, providing reassurance that the output voltage of a transformer is as expected and, when used together with excitation current tests, is also an important dielectric test for assessing the turn-to-turn insulation of the windings. No other dielectric screening test, except the excitation current test, stresses the turn-to-turn insulation. 

Even today, the Megger hand-crank TTR test instrument, which was introduced in 1949, is regarded by a large contingent of the most experienced testers in North America as the gold standard for TTR test instruments. When another TTR instrument produces results that seem dubious, the hand-crank is the instrument they use to finally determine whether or not there is a problem. 

Although the theory of the TTR test is often seen as simple – using the assumption that the ratio of the transformer terminal voltages is equal to the turns ratio – this is an oversimplification that can lead to errors because of the effects of the core permeability, winding configuration, excitation losses, test voltage and more. Much has been learned in recent years about sources of error in TTR measurement and new approaches have been developed to compensate for these. 

For example, field test results show that simultaneous three-phase excitation and a step-up method of testing, as opposed to the better known step-down method, greatly improve the accuracy of TTR test results. Step-up TTR testing allows better coupling between windings, generates more flux and reduces voltage dependence when compared with the step-down approach. Megger’s newest transformer test instrument, the TTRU3, incorporates all of the most recent developments to deliver the most consistently accurate ratio tests since the hand crank.

Toward the end of the twentieth century, there was a growing feeling that PF/DF and capacitance tests could provide more information. In response, by 1996, Dr Peter Werelius and Björn (Bengtsson) Jernström successfully brought dielectric frequency response (DFR) testing out of the laboratory and into the field, introducing the first field-portable DFR test instrument, which is now known as the Megger IDAX. 

This offers a reliable way to estimate the moisture content in a transformer’s cellulose/paper insulation, its oil conductivity and its thermal behaviour. The presence of water in the solid part of the insulation, even in small concentrations, increases its ageing rate and lowers the admissible hot-spot temperature of the transformer, which means the loading profile of the transformer in emergency situations may need to be reviewed. It also increases the risk of bubble formation and subsequent dielectric failure, as well as reducing the dielectric strength of transformer oil and the inception level of partial discharge activity. 

Since it is a diagnostic test, DFR is not intended to be performed on a routine basis, provided that a suitable screening test is used to flag up potential moisture problems. It is important to note, however, that a traditional PF/ DF and capacitance test is not a good screening tool for moisture contamination. On average, a line-frequency PF/DF test may not detect a problem until the moisture in paper is nearly 3%. To put this in perspective, at a constant operating temperature of 90 ºC, cellulose with 1% moisture has a life expectancy of approximately 12 years; at 3% moisture, the life expectancy is only 3 years. 

A PF/DF and capacitance test has other deficiencies as shown in Figure 2. Because of these, in the early years of this century, NB DFR testing emerged as an important supplement to PF/DF testing. Essentially, in an NB DFR test, PF/DF tests are repeated multiple times at different frequencies. Many PF/DF instruments have variable frequency capabilities and can potentially be used for NB DFR testing. However, to achieve the full benefits of this test, it is important to test at frequencies down to 1 Hz or lower, and up to at least 500 Hz (see the notes in Figure 2). Megger’s dedicated PF/DF and capacitance test instrument, the Delta, provides this functionality, as does the company’s TRAX multifunction substation test instrument. 

One significant milestone in the decades of DFR research is Megger’s development of the individual temperature correction (ITC) algorithm in 2010, which makes the interpretation of PF/DF and NB DFR results significantly more reliable. Any change in PF/DF may indicate a problem, but the electrical characteristics of insulating materials change with temperature. 

Since screening tests on insulation rely on detecting changes in the electrical characteristics of materials, only data that has been obtained at the same temperature can be compared. This way, any change in the measured electrical parameters can be attributed to a change in the state of the material and not a difference in temperature. Since it is not practical to ensure that assets are at the same temperature every time they are tested, some way of determining the insulation’s equivalent behavior at a baseline temperature (usually 20 ºC) is needed. 

Network frequency (50/60 Hz) power factor measurements have traditionally been corrected to 20 °C using an approximate correction factor. However, it is well-known that the true correction factor depends on insulation status, and since a look-up table can’t account for this, using a generic correction may introduce significant errors. Fortunately, by using knowledge about DFR interpretation and remembering that modern power factor instruments can sweep over a range of frequencies, it is now possible to determine a much more accurate temperature correction factor, with the aid of the ITC algorithm. This is done by carrying out a limited frequency sweep on the actual test object. 

Of course, dielectric screening is just one aspect of transformer health assessment and, in future articles, I am planning to look at some other key aspects. In the meantime, I hope this article has provided a useful bird’s-eye overview of the way in which transformer dielectric testing has evolved over the years, and how today’s test technology can provide valuable, dependable and easily acquired insights into the condition of these valuable and ubiquitous assets.