LARGE POWER TRANSFORMERS AND SHUNT REACTORS HAVE FAILED in part because of corrosive sulfur in the electrical insulating mineral oils. This has occurred even though the oils, in most or all cases, passed internationally recognized industry specifications.

The problem occurs because the corrosive sulfur reacts with copper conductors (Fig. 1) and silver contacts to form metal sulfides that contaminate the insulating paper. Since the metal sulfides are conductive, the dielectric breakdown strength of the paper is reduced. Under some conditions, a breakdown occurs through the insulation between conductor strands on a disk or between disks (Fig. 2). This ultimately leads to the failure of the apparatus. Although the number of failures has been small (double digits), the assets lost have been substantial. Examples have included the failure of a 500-kV shunt reactor and a 450-MVA autotransformer. The problem, which has been reported by many countries, tends to develop undetected in the apparatus for several years before failure occurs. As a result, there is some concern that the number of failures could grow considerably unless the industry takes action quickly.

The problem with corrosive sulfur is both time and temperature dependent. The longer an apparatus operates at higher temperatures, the greater the corrosion and formation of metal sulfides. Sealed transformers are more susceptible to the corrosive-sulfur problem, as oxygen reacts with copper and organo-metallic compounds competing for reaction sites with the corrosive sulfur. This slows down the formation of the conductive metal sulfides but does not stop their formation. Lower-voltage apparatus have an advantage if the copper conductors are coated with enamel insulation. The enamel creates a barrier preventing reaction with the corrosive sulfur.

Corrosive-sulfur problems were first recognized in the early 1900s in the United States. Methods to detect corrosive sulfur in oil were improved at that time and have been effective for many decades. It appears that in recent years changes occurred that have gone undetected, as oils continued to meet specifications. Yet the margin or temperature at which corrosive sulfur begins to form has been reduced for some oils.

PREVENTION

Worldwide there are specifications for corrosive sulfur in electrical insulating mineral oils. This is because mineral oils contain corrosive-sulfur compounds that must be removed to a high degree in the refining process for manufacture of electrical insulating oils. The effectiveness of the removal of corrosive-sulfur compounds should be verified and hence is specified. The corrosive-sulfur problem is of sufficient magnitude that it cannot be treated as an unusual occurrence. Standards organizations need to revise present specifications for a more rigorous evaluation for corrosive sulfur.

At Doble Engineering Co., we have revised our insulating mineral oil specification requirements for corrosive sulfur by modifying the present ASTM method D 1275. The modifications include extending the duration of the test from 19 to 48 hours and increasing the test temperature from 140°C to 150°C (284°F to 302°F). Some added precautions are taken to minimize the oxygen content in the test oil. In testing to date, we have found that the oils from failed transformers with evidence of copper-sulfide formation did not pass this modified test, but did pass the present ASTM method. As additional research is performed, it is likely that new test methods will be developed. However, to help prevent this problem from becoming widespread, immediate use of this more rigorous testing is recommended.

Corrosive sulfur is not unique to transformer mineral oils. Materials used in electric apparatus or to fill electric apparatus with oil may contain sulfur compounds, some of which may be corrosive. This includes hoses, gaskets, some water-based glues, copper and paper insulation. Care is required in the selection of materials for compatibility with mineral oils. Materials in contact with the oil should not add corrosive sulfur in amounts that would make the oil “corrosive” as measured by appropriate tests.

DETECTION OF EXISTING PROBLEMS

To date, the failures have occurred without prior evidence of an incipient fault. This makes the problem difficult to detect and manage. Failures have occurred after several years of apparatus being in service. The corrosion process appears to take this time to form critical amounts of conductive sulfides. Dissolved gas-in-oil tests are routinely used to detect developing problems from overheating, partial-discharge activity or arcing. Thus far, there has been no indication of partial-discharge activity or arcing in the units using dissolved gas-in-oil analysis, even when a sample was taken one day before the failure. Although there can be an increase in the insulation power factor, this test has not been found to routinely detect this problem when performed in the standard fashion. Laboratory investigations using power factor have been able to detect copper sulfide in paper under some conditions. More research is needed to determine if this can be applied to apparatus.

Currently, the best means for detecting corrosive-sulfur problems has been by internal investigation. Testing the oil can determine if the apparatus could develop a problem, but it is the internal inspection that has revealed the evidence of copper-sulfide formation. Often, the most visible evidence is on copper surfaces. If there is exposed copper, the tarnish will be visible. As the tarnish worsens, it turns gray and can be mistaken for carbon from overheating or a failure. Unwrapping the copper conductor can reveal further tarnish. The amount of corrosion is not uniform in a disk (Fig. 1) or throughout the windings.

In some cases, the copper sulfide can be visibly seen on the paper insulation and ranges considerably in coloration depending upon severity, closeness to the conductor and other factors. The copper sulfide is also not uniformly distributed in the paper insulation. Figures 3 and 4 show analysis of paper surfaces as performed by scanning electron microscopy (SEM) and energy dispersive x-ray (EDX) analysis. The SEM achieves high magnification and EDX is used to determine the elemental composition of any particles. Figure 3 shows normal paper, with the individual fibers clearly visible in both the Kraft and crepe papers, and with no signs of foreign material present. Even the crimping of the crepe paper is clearly visible. When copper/sulfur or other metal/sulfur contamination of the paper surface occurs, the results can be profound (Fig. 4). It can be seen from the SEM photomicrographs at a 200 times magnification that copper-sulfur contamination is clearly on the paper insulation. Figure 4a shows all the fiber surfaces and the gaps between the fibers encrusted with the metal sulfides. Figure 4b shows an earlier stage with what appears to be globs of metal sulfides and sulfates.

RESEARCH EFFORTS AND INTERIM SOLUTIONS

It is likely that there is a substantial amount of oil in service in large power transformers that can cause the sulfur-corrosion problems. Although these oils with excessive amounts of corrosive sulfur represent a very small percentage of the total oil volume in service, it is important as copper corrosion and sulfide formation cannot be reversed. Once a corrosive-sulfur problem is detected, it might be possible to mitigate against further significant corrosion. Possible methods include removal of the corrosive-sulfur compounds, oil replacement, partial oil replacement and passivation using metal deactivators. No simple means for removing corrosive-sulfur compounds from the oil has been developed. Depending on the size of the problem, this may be an area for future research. Preliminary studies at the Doble Laboratories have shown that mineral oils can have considerable differences in the temperatures at which they begin to form copper sulfide under test conditions. Mixing of higher corrosive-sulfur-content oils with those of very low sulfur contents can significantly improve the former's characteristics well beyond acceptable limits. Another method that has been employed is the use of a passivator that binds up some of the active sites on the metal surface retarding reactions with corrosive sulfur. While some companies have already started using these approaches, more research is needed to better understand how well these methods will work and the long-term benefits. For the immediate future, however, these methods appear to be promising and could help mitigate developing problems.

THE INDUSTRY RESPONDS

Recently, the electric power industry has seen an old problem resurface, corrosive sulfur in oil. Enough problems have occurred around the world that the industry needs to respond. This includes changing specifications, identifying where problems could exist with appropriate tests and developing methods for mitigation. Doble Engineering Co. has already changed its specification and is undertaking a collaborative study with industry participants to help solve this problem. International standards organizations already have started the process to modify existing methods and/or develop new ones. With a quick response, this problem can be managed to minimize its impact on the industry.

Paul Griffin has been with Doble Engineering Co. since 1978 and held the position of laboratory manager before becoming vice president of Laboratory Services. Since joining Doble, Griffin has published more than 50 technical papers pertaining to testing of electrical insulating materials and laboratory diagnostics. pgriffin@doble.com

Lance Lewand is the laboratory manager for the Doble Materials Laboratory and is also the product manager for the Doble DOMINO, which is a moisture-in-oil sensor. The Materials Laboratory is responsible for routine and investigative analyses of liquid and solid dielectrics for electric apparatus. Since joining Doble in 1992, Lewand has published numerous technical papers pertaining to testing and sampling of electrical insulating materials and laboratory diagnostics. Lewand was formerly manager of Transformer Fluid Test Laboratory and PCB and Oil Field Services at MET Electrical Testing Co. (Baltimore, Maryland, U.S.). His years of field service experience in this capacity provide a unique perspective, coupling laboratory analysis and field service work. Lewand received his BS degree from St. Mary's College of Maryland. He has been actively involved in professional organizations such as ASTM D-27 since 1989 and is a subcommittee chair. He also is the secretary of the Doble Committee on Insulating Materials. llewand@doble.com