Você está na página 1de 7

DISSOLVED GAS ANALYSIS OF LOAD TAP CHANGERS FIELD GUIDE AND SUBCOMMITTEE REPORT Rick G.

Asche, PE Portland General Electric


ABSTRACT
The Doble Client Committee on Transformers formed a subcommittee in 2004 to prepare a guide for analyzing DGA results from Load Tap Changers. The resulting guide is the compilation of DGA test results for several major tapchangers designs used mostly in the United States and Canada. The guide establishes normal gassing tendencies.

INTRODUCTION
Research on gassing levels in load tapchangers done by Cinergy Energy was presented to the Doble Conference in 1993. As a result of this presentation other utilities began pilot programs to analyze the dissolved gases inside of load tapchangers (LTC). Experience began to show normal patterns of gassing levels, but there were large variances between manufacturers and styles of LTCs. Sometimes a unit with elevated gas levels was entered for inspection with no problem found. Discussion at the Doble Client committee in 1995 favored the formation of a Subcommittee to pull together client experience and develop analytical tools to help determine when a problem existed in the LTC. As the Subcommittee project progressed, DGA of LTCs grew from a few utilities with pilot projects to almost all major utilities adopting it as a mainstream maintenance practice. Feedback from users of the technique only confirmed its cost effectiveness and proactive nature. LTC-DGA Subcommittee The main purpose of the LTC-DGA Subcommittee was to promote the use of DGA on Load Tap Changers (LTC) and develop a field guide useful in the interpretation of results. During the development process we focused on the major tapchangers types used by the client groups in the United States and Canada due to the large population base. Most participants were just starting LTC-DGA programs, but a handful of clients had many years of test data. Data was collected after the significant data values were defined. At this point of the project, a total of 44,800 samples have been collected for about 1060 different transformers. Most of the data was from units with no identified problem. As the project progressed, limits were set for normal operation. This method was used rather than an upper boundary concern limit. This distinguishes this method from most other laboratory recommendations that attempt to define the internal problem. Tapchanger designs are similar in the materials used; bare copper or silver plated sliding contacts, tungsten arcing contacts that were sliding or button contacts, phenolic or epoxy composite supports and very little cellulose. Vacuum tapchangers use similar materials but suppress the arcing by putting the arcing contacts in a vacuum. The major difference then is the speed of operation and the duration of the arcing period. The normal gases generated in an LTCs are those produced by arcing in oil. (Vacuum LTCs should not have any arcing gases) A developing problem starts overheating contacts and then the hot metal gases are generated. How does LTC-DGA differ from DGA? Both types of DGA analysis are based on trending the results over a period of time and a number of samples. In conventional DGA analysis, there are normal levels of gases that are very low or non existent. The appearance of any acetylene, ethane or ethylene is cause for alarm in most transformers. On a highly loaded unit, there will be the paper degradation gases carbon monoxide, carbon dioxide. Increasing amounts of any of these gases would be cause for alarm. In contrast, gases are generated in a LTC during normal operation. These are acetylene, ethylene, ethane and hydrogen. Some of these accumulate until equilibrium is reached. For this reason, original attempts to use DGA in Tapchangers and circuit breakers shortly after DGA was developed were not continued industry wide. Industry maintenance practice on

2010 Doble Engineering Company -77th Annual International Doble Client Conference All Rights Reserved

tapchangers was fairly conservative and more frequent than it is today. Some utilities performed maintenance on annual intervals based on experience. Utilities used time or operations as the criteria for maintenance. LTC failures were common, especially in certain specific types of LTCs and clients would shorten their maintenance intervals to prevent these failures. This helped, but failures would still occur. Enter the 1990s and the push for Just -In-Time or Condition Based Maintenance. Reliability Centered Maintenance (RCM) became the new buzz-word for maintenance. A key component of RCM was determining the condition of equipment prior to maintenance to determine if it was required. The application of test methods both online and offline was emphasized to determine the internal conditions of apparatus. One of the most cost effective benefits of RCM was the re-introduction of Dissolved Gas Analysis (DGA) on tapchangers to determine internal conditions. Off course, data interpretation also had to follow along as experience was gained. Trend analysis is key to determine individual performance of a tapchanger, but aggregating test data over a fleet of the same design is beneficial to establish general operating parameters. The aggregate may have a wider range of normal data.

Sampling Intervals
Trend analysis for incipient problems that were tracked showed rapid increases in DGA gases. The sample interval is an important parameter. Most utilities have adopted an annual sample interval for most LTCs, but some of the poor performing designs are sampled more frequently. Experience of the subcommittee favors 6 months for the worst performing units and annually for the more reliable units as a practical interval. Ideally a continuous monitor is desired, but there are no commercial units offered to the market. With periodic sampling a few LTCs failures are not detected prior to failure. Data reported from more frequent DGA sampling was insufficient to determine a better recommended interval. Once a problem is detected, most clients would resample immediately and then resample periodically to establish gassing rates. If the level was high enough, units would be scheduled for maintenance.

KEY GASES
The Key gases used for LTC-DGA analysis are Ethylene and Acetylene. Ethane and Hydrogen are also present but seem to always be follower gases. Tracking just the Ethylene and Acetylene gas levels have been used by most clients to determine when a problem is developing inside the tapchanger. The field guide lists the statistically normal operating gas level. A tapchanger with a normal stable level above these levels may still be normal if the gas levels remain stable.

RATIOS
The use of gas ratios was pioneered early in DGA analysis. Rodgers Ratios is a well known example of gas ratios used by to analysis DGA data for problem identification. Rather than use the absolute values, the relationship between the amounts of gas was tied to the temperature of the generating surface. Hot Metal Gases are generated depending on the temperature of the surface. As you move away from the surface, the specific gases changed as well as the amounts being generated. Using the relative amounts of these Hot Metal Gases could tell the maximum temperatures being reached. DGA data from our study of normally operating LTCs have a ratio of Ethylene to Acetylene of less than 1.0. As wear occurs within the LTC or a problem starts to develop, the Ethylene production tends to increase and overtake the acetylene production. Some have suggested that should the contacts have significant carbon buildup, arcing may not occur and only heating takes place. The E/A ratio now becomes greater than 1. This was compared across all the types of tapchangers to see if it can be applied universally in lieu of limits specifically applied to a specific design. In the guide, we assigned a ratio of 1.0 as the concern level if no specific ratio had previously been suggested. Applying ratios to small quantities of gas is not reliable since small changes can have a huge impact on the ratio. Thus a minimum amount of 100PPM should be considered before apply the ratio.

2010 Doble Engineering Company -77th Annual International Doble Client Conference All Rights Reserved

Other laboratories and Utilities have used ratios both lower and greater than 1.0. Reasons for the difference are experience based.

Duval Triangle
The Duval Triangle was proposed as a tool to assist with analysis of LTC-DGA data. In 2009, Dr. Duval published a paper at the spring Doble Conference validating the application of the method to LTCs and the equations he found useful. Again, analysis is trending, but single data points can be used to identify the temperatures inside the tapchanger. It will prove useful for those who wish to use this form of presentation. It again is a ratio based analysis tool.

OTHER METHODS TO IMPROVE LTC MAINTENANCE


There are other methods to deal with tapchanger maintenance problems and to improve the service life and the reliability.

Oil Filtration
Recognition of the fact that not all failures will be prevented unless very frequent sampling is employed led some pioneering companies to look at other techniques to extend tapchanger maintenance. When tapchangers are opened for maintenance, the oil is usually very sludged up, with a large amount of carbon present. This is the result of oil breakdown due to arcing and heating. There seems to be an exponential acceleration in oil degradation as contacts become worn. If the oil could be kept clean, could the maintenance interval be extended? This was answered with success with the use of oil filters being applied to LTCs online. Since 2000 when first introduced to the client group, this has now become a common practice for both retrofitting and as OEM equipment on new units. Oil filters do affect the gas generation rates by lowering them. As filters become more common place and more experience is gained, new baselines may be established. In lieu of the baseline values, use of the ratio is still appropriate at this time. Oil filters have been so effective at reducing coking problems in transformers, that clients have been able to extend maintenance intervals to reasonable levels again.

Free Venting
Another issue that became apparent during the project was the difference between gas accumulations in sealed units versus free breathing units of the same design. Some manufacturers had sealed their LTCs to prevent the intrusion of moisture and if maintenance was frequent, the accumulating gas levels were not harmful. This philosophy has now changed to make all LTCs free breathers by either direct venting or venting through a desiccant breather. My utility ran a little experiment a number of years ago when we were working on online gas monitors. We installed one of these 3 gas units on a free breathing tapchanger and established a normal baseline. We then plugged the free breather for 30 days. The gas levels rose as would be predicted and after this period we removed the plug. The gas levels including acetylene again returned to the normal range. Clients have reported efforts to retrofit free breathers on tapchangers that previously had been sealed.

Tracer Elements
Research has been continuing on using wear indicators in contact materials. These efforts are still proceeding, but the requirement to manufacture contacts with tracer inserts prevents the methods widespread use. It also introduces new compounds to detect in oil analysis.

Delta Temperature
Monitoring systems were developed to measure the tapchanger oil temperature and main tank oil temperature. A simple delta calculation is easy to perform. An alarm would be created if the tapchanger temperature was greater than 5 C warmer than
2010 Doble Engineering Company -77th Annual International Doble Client Conference All Rights Reserved

the main tank. This alarm point was selected based on experience that showed the tapchanger should be cooler than the main tank. When a problem developed inside the LTC compartment, the temperature rises and this should be seen as an alarm. Although the stand alone monitoring systems were not widely applied as retrofit devices, the concept was and is now being incorporated into new transformer temperature monitors.

Online Gas Monitors


One vendor of Transformer On-line Gas monitoring equipment is continuing research into applying a 3 gas monitoring system on tapchangers. My utility has participated in this research passively. The promise of this method to provide timely warning of impending problem is offset by the high capital cost of the monitor, the ongoing maintenance costs and the reliability of the system. Still, this research could help answer the question as to how fast a tapchanger deteriorates.

Tapchanger Replacement
Another result of this study is identification of poorly designed tapchangers. Most of these units were purchased many years ago, and are no longer manufactured, but as the owner of such a unit, the client is faced with a number of options: . Run the unit to failure . Replace the transformer and scrap the old unit . Continue monitoring and repair in kind . Upgrade the LTC unit . Replace the LTC unit Obviously, the first two options have very high financial impact. Failures of an LTC will sometime cause failure of a transformer, some even catastrophically. Our very job function as maintenance personnel is to prevent this. Replacing an entire transformer because of a poor performing LTC may be necessary, but what do you then do with the old transformer? Most of us do the third option of monitoring and repairing in-kind. Its been said that the definition of insanity is doing the same thing over and over again, expecting different results. Still this may be the only option a maintenance group is allowed. The last two choices must be analyzed for cost effectiveness. If the transformer has useful life left, it may be beneficial to perform either a major upgrade or an entire replacement.

Types of Problems Found


Over the history of the Subcommittee we have seen a reduction in the number of tapchanger failures and an increase in the number of saves. Some of this was reported in the 2007 Client Conference with a symposium on tapchanger saves as the result of early detection. Problems could be classed into: 1. Problem with arcing contacts (diversion, transfer, etc) that interrupt load current. 2. Problem with selector switch (includes main and moving contacts, slip ring contacts) 3. Problem with reversing switch contacts 4. Problem with studs 5. Problem with resistor in UZ style LTCs 6. Other The detection of any problem inside of the tapchanger almost always means that the transformer must be taken out of service, the tapchanger opened and something must be fixed. Because of this, determining the exact location of the problem is only marginally beneficial. If the utility carries an extensive selection of repair parts, then timely repair is achieved. This is why the field guide did not try to determine the exact condition code to distinguish between a selector switch versus a reversing switch or arcing switch problem.

2010 Doble Engineering Company -77th Annual International Doble Client Conference All Rights Reserved

LTC-DGA FIELD GUIDE


The heart of the Field Guide is the table from the guide which presents the normal gassing levels and the ratios the subcommittee agreed upon as most indicative of normal operation. This is shown in Table 1. For this report we have extended a default ratio of 1 to all arcing type tapchangers unless a prior ratio had been determined. We have also divided FPE types and GE types in discreet categories due to these have differing levels of gas among the various models. TABLE 1 Table of Normal Dissolved Gas Levels of Load Tap Changers
FB => Free Breather,
Manufacturer

DB => Desiccant Breather S => Sealed, ppm => Parts Per Million.
Type(s) Tank Type Hydrogen H2 (ppm) Methane CH4 (ppm) Acetylene C2H2 (ppm) Ethylene C2H4 (ppm) Ethane C2H6 (ppm) Ratio Ethylene/Acetylene E/A

A-BB, ASEA, RTE, RTE- ASEA, Magnetek Waukesha, Allis-Chalmers Siemens Cooper McGrawEdison Pennsylvania FPE General Electric

UZD, UZE, UZF

DB

<600

< 0.33

TLS, TLH - 8, 10 TLS, TLH -20, 21 550 FB

<50 <100

<150 <100

<50 <50 <2000

0.16 < E/A <2 0.25 < E/A <1 <1

Maloney Reinhausen Various Westinghouse

TC-251, TC525, TC-546 LR, LRS, LRT68, 72 LR-65 LR-65 LR, LRS, LRT45, 59, 68, 72 MA/MB MR Vacuum URS

FB S S FB FB FB S See Note 2 FB S Filter S FB S <5 <2000 <5000 <1000 <10,000 <100 <2,000

<2000 <500 <5000 <500 <100 (no online filter) <2500 <7500 See Note 2 <40 <500 <1000 <500 <3000 <250 <5000

<1 <0.4 <0.5 <0.5 <13 <1 <1 <2 <13 <13 <13 <13 <13 <13 <13

See Note 2

URT UTT

<50 <1000

Interpretation of Table Data: If no entry is made for the load tap changer type, this gas is not considered significant for trending. Not all load tap changers will follow the patterns shown. Key for the superscripts and notes used in the Table

2010 Doble Engineering Company -77th Annual International Doble Client Conference All Rights Reserved

1.

A service advisory, FS-TC-013, LTC Tap Gear Breather Modification, Revision 1: December 4, 1983, original issue: June, 1976 exists for the TC-25 load tap changer. The modification includes the method to convert the tank from a sealed to a free breather system. C2H4 + CH4 + C2H6 < 500 ppm The ethylene/acetylene ratio limit of <1 for this LTC type is suggested from general experience of the subcommittee members but was NOT studied in detail.

1. 2.

CONCLUSION
LTC-DGA has been shown to be a cost effective maintenance tool to detect abnormal operation of a load tapchanger. The guide published by the LTC-DGA Subcommittee pools the client groups experience and data to form operating norms. The method was validated by case histories demonstrating the prevention of LTC failures when gas data showed deviation from the guidelines established in the guide. The use of the Acetylene / Ethylene gas ratio is accepted as a general tool to indicate incipient faults. Not all LTC failures can be detected in time due to the infrequent sample interval and the rate at which problems develop. Still clients report successful detection of incipient problems at almost every Utility that has applied the technique. Until better, more cost effective on-line methods prove themselves; LTC-DGA will continue to be a mainstay program for maintenance organizations.

ACKNOWLEDGEMENTS
The author wishes to acknowledge the man hours of labor that have been expended over the life of this project by the committee secretary in data manipulations, the Client members of the subcommittee for contributing their gas data and faithfulness in attending committee meetings, for the contributions in preparing the guide, and for the field personnel that actually take the oil samples. It is rare that a new maintenance activity is so readily accepted by the Utility Industry as a whole and that is so effective at reducing maintenance costs. The author especially wishes to thank the recent pioneers of this method that had the persistence to overcome historical prejudice against this method being used for LTCs.

REFERENCES
[1] R. Youngblood, Cinergy Co, F. Jakob, T.Haupert Application of DGA to Detection of Hot Spots in Load Tapchangers Sixtieth Doble Client Conference Minutes 1993 section 6-4.1 [2] R. Youngblood, Cinergy Co. An Update on Load Tapchanger Hot Spot Detection Through the use of DGA Set", SixtyFirst Doble Client Conference Minutes 1994 section 6-14.1 [3] Duval M.-Hydro Quebec (IREQ), "The Duval Triangle for LTCs, Alternative Fluids and Other Applications, 76th Annual International Doble Client Conference Minutes 2009 [4] R.G. Asche, Portland General Electric, Dissolved Gas Analysis of Load Tapchanger Subcommittee Project Report, 697th Annual International Doble Client Conference Minutes 2002, section 13E [5] Doble Client Transformer Committee Sub-Committee Report on Transformer Load Tap Changer Dissolved Gas Analysis November 18, 2009 Revision

BIOGRAPHY
2010 Doble Engineering Company -77th Annual International Doble Client Conference All Rights Reserved

Rick G. Asche, PE has chaired the subcommittee of LTC-DGA since its inception in 1995. He is a Senior Engineer with the substation maintenance group of Portland General Electric in Portland Oregon providing expert support for special testing, protection system maintenance and equipment commissioning. He has been employed at PGE since 1976. He has served as chairman of the ACCA and IM Doble Client committees. He has also authored and presented several papers at the Doble Spring Conference and WEI Hands on Relay School. He is a 1976 BSEE graduate of Oregon State University and is a Registered Engineer in the State of Oregon.

2010 Doble Engineering Company -77th Annual International Doble Client Conference All Rights Reserved

Você também pode gostar