Estimation of Insulation Life Based on a Dual Temperature Aging Model Weidmann-ACTI Inc. Fifth Annual Technical Conference Albuquerque, NM Nov. 13-15, 2006 Thomas A. Prevost, Hans Peter Gasser Weidmann Electrical Technology Roger Wicks, Brian Glenn, Rick Marek DuPont Presenter – Thomas A. Prevost, Weidmann-ACTI ABSTRACT In order to predict the life of oil filled power transformers it is vital to understand the insulation system. Loading guides as well as diagnostics for power transformers are based on insulation models which predict loss of life based on material characteristics, temperature and time. Up to now the IEEE has defined full-scale test models for distribution transformers but has avoided full scale tests of power transformers due primarily to cost. As an alternative, IEEE is considering a standard model that can be used to analyze insulation systems for power transformers. This model, known as the dual temperature aging model, separately controls the temperature of the conductor with its associated insulation material and the bulk oil. This paper presents preliminary results from a dual temperature aging model study that is currently in progress. Data includes typical diagnostic parameters such as the tensile strength and degree of polymerization (DP) of the cellulose insulation and the furan content present in the oil. This study is also investigating non-thermally upgraded and thermally upgraded conductor insulation to determine the effects of thermal upgrading on insulation aging. The by-products of aging that are the key to diagnostic analysis are also presented. INTRODUCTION It is a common belief that liquid filled power transformers are designed with a life expectancy of approximately 20 years, although there are many transformers in operation today whose life exceeds fifty years. The life of a transformer depends on the life of the insulation system. This life is defined by a chemical process which depends primarily on temperature and time. Other factors such as the presence of oxygen and moisture will accelerate the aging process. With the aging global infrastructure that we have in place today it is becoming increasingly important to be able to accurately estimate the remaining life expectancy of power transformers. IEEE and IEC have refined the loading guides which give the user a basis for estimating life consumption of the transformer based on time and temperature. However, the foundation for these loading guides are insulation life curves which are based on both models and full-scale devices. Such life curves have been published for many years. Two significant changes were made during the last revision of the IEEE guide for loading transformers. (1) The first significant change was that the three loading guides for transformers based on size were combined into one guide. The following three guides: 1 C57.91-1981 IEEE Guide for Loading Mineral-Oil-Immersed Overhead and PadMounted Distribution Transformers Rated 500kVA and Less with 65 °C or 55 °C Average Winding Rise C57.92-1981 IEEE Guide for Loading Mineral-Oil-Immersed Power Transformers Up to and Including 100 MVA with 55 °C or 65 °C Average Winding Rise C57.115-1991 IEEE Guide for Loading Mineral-Oil-Immersed Power Transformers Rated in Excess of 100 MVA (65°C Winding Rise) (Folded into C57.91-1995) (65°C Winding Rise) Were combined into one guide: C57.91-1995 IEEE Guide for Loading Mineral-Oil-Immersed Transformers The second major change was the insulation life criteria suggested for power transformers. In the previous guide, C57.92-1991, the insulation life was based on sealed tube aging tests of insulation models with the end of life criteria being 50% tensile retention. The present loading guide, C57-91-1995, permits the user to select the criteria most applicable to their needs. The end-of-life criteria can be based on % strength, reduction, degree of polymerization, dielectric strength, etc. and uses an insulation-aging factor. While the IEEE loading guide was being revised the standard test procedure for thermal evaluation of liquid-immersed Distribution and Power Transformers IEEE C57.100 was also revised in 1999.(2) C57.100 defined a minimum life expectancy curve for liquidimmersed distribution, power and regulating transformers with a 65°C average winding rise and a 80°C hottest-spot rise. This curve gives a minimum life expectancy of the insulation system of 180,000 hours (20.5 years) at rated load as defined in C57.12.001993. This curve was taken from the loading guide for distribution transformers, C57-911981 based on full-scale models. Figure 1 Minimum Life Expectancy Curve from IEEE C57.91-1995 2 C57.100-1999 defines test procedures for distribution transformers based on full-scale models. For power transformers it is recognized that full-scale models are impracticable. The standard recommends that the transformer manufacturer construct a model which utilizes typical conductors, insulation, and supporting structures that can provide performance which is representative of power transformer thermal endurance. The models are based on the assumption that the conductor turn insulation is subjected to the most critical thermal degradation and is the limiting material. The test procedure for the power transformer models should be similar to that of distribution transformers which involves circulating current through the windings to elevate the temperature to the desired test condition and time and then applying end-point tests on the model. The end-point test for power transformer models, as with full-scale distribution transformer models, is comprised of a short circuit test followed by dielectric test. The IEEE PES transformer committee currently has a project to revise C57.100-1999. There is concern that the requirement for model testing of the insulation system for power transformers is not well defined and could be costly to perform because each power transformer is unique from a design and construction standpoint. The working group is considering standardization of an insulation system model which includes independently controlled conductor and bulk oil insulation which can economically model insulation systems for power transformers. A significant task for the working group will be to establish end-of-life criteria for the insulation system for this model. END-OF-LIFE CRITERIA In order to properly measure and then predict insulation life it is critical to understand the basis from which any reference values are based. Over the years there have been many studies on insulation aging which have generated significant and often conflicting results. (3),(4),(5) It is important to understand the insulation system model which served as the basis for these studies as well as the criteria for determining end-of life. Early work depended on sealed tube models in which insulation materials, oil and other construction materials were placed in a sealed tube and then heated to the desired aging temperature. These models were relatively inexpensive to manufacture so that it was practical to obtain considerable amounts of data. This test, however simple though, was never standardized in regard to material volumes and ratios so that comparing the results of these test is difficult. The early consensus of end-of-life for sealed tube aging models was based on 50% retention of tensile strength. In 1948 Dakin published a paper(6) in which he premised that insulation aging was a chemical rate phenomenon and that deteriation of insulation strength could be modeled by the Arrhenius reaction rate theory with the following form: B ) Per Unit Life = AExp ( Θ + 273 Where: Θ = Winding hottest spot temperature, °C A = Constant B = Constant Equation 1 Arrhenius Reaction Rate 3 The Arrhenius rate theory is still used today to describe the aging process of insulation in transformers. The criteria for end-of-life for insulation models has continued to be debated. It is well known that many transformers that have conductor insulation with less than 50% tensile strength retention continue to operate with no problems. Accordingly, a value of 25% retention has been suggested as end-of life. Another argument has been to use a measurement of the degree of polymerization of the paper rather than the relative mechanical strength.(7) This parameter is more practical when estimating remaining life on equipment in service because it is almost impossible to determine a baseline value for the mechanical strength. As mentioned earlier, another approach for end-of-life is to base failure criteria on dielectric tests following short-circuit tests on transformers or insulation models. Each of these end-of-life criteria give different expected life values for standard 65 °C rise, insulation systems. Table 2 in C57.91-1995 lists these values for various end-of-life criteria. Table 2 - Normal insulation life of a well-dried, oxygen-free 65ºC average winding temperature rise insulation system at the reference temperature of 110ºC Normal insulation life Basis Hours Years 50% retained tensile strength of insulation (former IEEE Std. C57.92-1981 criterion) 65,000 7.42 25% retained tensile strength of insulation 135,000 15.41 200 retained degree of polymerization in insulation 150,000 17.12 Interpretation of distribution transformer functional life test data (former IEEE Std. C57.91-1981 criterion) 180,000 20.55 NOTES 1 - Tensile strength or degree of polymerization (D.P.) retention values were determined by sealed tube aging tests on well-dried insulation samples in oxygen-free oil. 2 - Refer to I.2 in annex I for discussion of the effect of higher values of water and oxygen and also for the discussion on the basis given above. Figure 2 Normal Insulation Life Values from IEEE C57.91-1995 DUAL TEMPERATURE AGING MODEL While IEEE C57.100-1999, Standard Test Procedure for Thermal Evaluation of LiquidImmersed Distribution and Power Transformers, has a practical, cost-effective procedure for thermal evaluation of insulation systems for distribution transformers, many feel that it is not practical for larger power transformers. Furthermore, there is concern that the sealed tube test method is an unrealistic model. For these reasons, a Dual Temperature 4 Aging Model is being developed. In this model the insulated conductor and associated turn spacer insulation can be controlled at one temperature while the bulk oil and its associated barrier type insulation can be controlled at a second temperature. This model has been incorporated into IEC TS 62332, IEC Technical Specification – Electrical Insulation Systems (EIS) – Thermal Evaluation of Combined Liquid and Solid Components. Part 1 – General Requirements. (2005). A study is currently underway to measure several potential end-of-life criteria during aging experiments in order to establish appropriate values for this model. When selecting end-of-life criteria it is important not only to be able to correlate these with actual service conditions but also to be able to measure these properties with established laboratory methods which give repeatable results and are cost effective. The use of retained tensile strength has remained a useful life criterion for this reason. The degree of polymerization has become widely recognized as well. Other diagnostics methods such as furanic compound measurement and DGA are being determined in order to provide correlation information to support in-service diagnostics for estimation of remaining insulation life. Test Program: To verify a base model and then to demonstrate that this model can respond to different insulation materials the following configurations were tested: 1) Non-Thermally upgraded crepe conductor insulation together with high density precompressed kraft pressboard spacer material and low-density pressboard barrier (bulk oil) insulation material. 2) Insuldur thermally upgraded conductor insulation with spacer and bulk oil insulation materials as in (1). These models were aged at different temperatures and at least three different times in order to establish aging curves for each of the models. The results will be compared to existing curves in the IEEE loading guide that establish the baseline for conventional insulation systems. Aging criteria: The criteria for end of life will be dependent on the performance criteria for specific insulation materials. These are: Conductor Insulation: Tensile retention DP (Degree of Polymerization) Other properties that were measured in order to understand the aging mechanisms: Fluid: Gasses Moisture Acid Furans RESULTS: For each paper type and aging temperature a polynomial regression curve was plotted in order to determine the time at which the end-of-life parameter was reached. These values 5 were then plotted on the Arrhenius plots to develop life curves. As an example two of the plots for obtaining 50% tensile retention values are shown below as figures 3 and 4. 176C Upgraded 3.2% N2 100.00 90.00 Percent Tensile 80.00 70.00 60.00 50.00 40.00 30.00 20.00 10.00 0.00 0 500 1000 1500 2000 2500 Aging Hours Figure 3 Plot of tensile retention vs aging time at 176 C 50% Tensile Retention = 692 hours 192 Upgraded 3.2% N2 Percent Tensile 100.00 80.00 60.00 40.00 20.00 0.00 0 100 200 300 400 500 Aging Hours Figure 4 Plot of tensile retention vs aging time at 192 C 50% Tensile Retention = 176 hours The corresponding end-of-life time for each aging temperature was then plotted following the Arrhenius relationship to develop life curves based on the results of the dual temperature aging model. The two properties, which were used to determine end-oflife, were tensile strength retention and degree of polymerization. 6 % Tensile Retention Arrhenius Life Plot End of Life 50% Tensile Retention 3.2% N2 Non-Upgraded Linear (3.2% N2) Linear (Non-Upgraded) 6 5 y = 9354.244x - 17.797 Log Life 4 y = 8084.559x - 15.567 3 2 1 0 0.00205 0.0021 0.00215 0.0022 0.00225 0.0023 0.00235 0.0024 0.00245 1/T Figure 5 Arrhenius Life Plot of 50% Tensile Retention Degree of Polymerization Arrhenius Life Plot End of Life DP = 200 3.2%N2 Non-Upgraded 5 4.5 y = 7836.8x - 14.361 4 2 R = 0.9638 Log Life 3.5 3 2.5 y = 7815.9x - 14.908 2 2 R = 0.8776 1.5 1 0.5 0 0.00205 0.0021 0.00215 0.0022 0.00225 0.0023 1/T Figure 6 Arrhenius Life Plot of 200 DP 7 0.00235 0.0024 0.00245 Arrhenius Life Plot Thermally Upgraded Paper N2=3.2% 50% Tensile C57.92 C57.91 4 DP 200 y = 9354.2x - 17.797 R2 = 0.9295 3.5 y = 6328.8x - 11.269 3 y = 7836.752x - 14.361 2 Log Life R = 0.964 2.5 y = 6972.2x - 13.391 2 1.5 1 0.5 0 0.0021 0.00215 0.0022 0.00225 0.0023 0.00235 1/T Figure 7 Plot of Tensile Retention and DP versus standard aging curves Relationship of Tensile Strength Retention versus DP In this study we found a strong relationship between the tensile strength retention and the degree of polymerization. The influence of thermal upgrading did not affect this relationship. DP vs. Tensile Strength Retention Non-upgraded 3.2% N2 Poly. (Non-upgraded) 1400.00 1200.00 y = 0.002x3 - 0.1914x2 + 11.425x - 68.248 R2 = 0.96 1000.00 DP 800.00 600.00 400.00 200.00 0.00 0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00 90.00 100.00 Tensile Strength (% Retention) Figure 8 Relationship between DP and % Tensile Retention (Note: this includes thermally upgraded and non-upgraded paper) 8 FURANS Background Because it is very difficult to obtain paper samples from in service transformers to determine insulation life, an alternative method has been developed. In this method the insulating oil is analyzed for furan content. Furanic compounds are produced during the breakdown of the cellulose insulation in transformers. There have been many studies which correlate the furan content in the oil to the degree of aging of the cellulose insulation. (8),(9) As the cellulose insulation ages the polymer chain breaks down. Each splitting of the chain liberates a glucose monomer which undergoes further chemical reaction and becomes one of several furanic compounds. Measurement of the furanic compounds, which are partially soluble in oil, can then give an estimate of the DP. Chendong developed a widely accepted relationship between the amount of 2-furaldehyde and DP.(10) Based on field data and laboratory studies he developed the following equation: Log10[2-FALPPM] = 1.51 - .0035 * DP Equation 2 Chendong’s Relationship between 2-Furaldehyde and DP The correlation between furan content and DP has been developed by the analysis of actual data in thousands of transformers. It needs to be noted that most of these studies were done on European transformers which in most cases do not contain thermallyupgraded paper. Subsequent studies done looking at the transformer database based on thermally upgraded insulation and non-upgraded insulation supported a theory that the thermal upgrading chemicals react with the furans breaking them down. (11) Stebbins proposed a modified Chendong equation to be used for transformers with thermally upgraded paper. DP= [Log10(2-FALPPB * 0.88) – 4.51]/-.0035 Equation 3 Stebbin’s Modified Chendong Equation for Thermally Upgraded Kraft When Stebbin’s equation is put into the same form as Chendong’s equation it becomes: Log10[2-FALPPM] = 1.5655 - .0035 * DP Equation 4 Stebbin’s Modified Chendong Equation for Thermally Upgraded Kraft 9 Results of test: Log 2-furaldehyde vs. Paper DP Kraft DP TU Kraft DP Linear (TU Kraft DP) Linear (Kraft DP) 5 4.5 Log (2-FAL - ppb) 4 3.5 3 2.5 2 1.5 1 0.5 0 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 Paper DP Figure 9 Log 2-FAL versus DP for Thermally upgraded and Non-Thermally Upgraded Paper Log 2-furaldehyde vs. Paper DP DuPont-Weidmann Aging Chengdong Aging Stebbins (mod. Chendong) Linear (DuPont-Weidmann Aging) 5 4.5 Log(2-FAL)-ppb 4 3.5 3 2.5 2 1.5 1 0.5 0 0 100 200 300 400 500 600 700 800 900 1000 DP Figure 10 Dual Temperature Aging Study Results versus Chendong and Stebbins Based on the results of this study we found that there was minimal difference between the formation of 2-FAL between non-thermally upgraded kraft paper and thermally upgraded kraft paper. 10 SUMMARY AND CONCLUSION: End-of-Life Criteria The results of the Dual Temperature Aging Model tests to date suggest that the end-oflife for thermally upgraded kraft paper insulation systems are close to the values suggested in the IEEE Loading Guide C57.91-1995 for insulation models. This gives an insulation life of 180,000 hours at a hot spot temperature of 110 °C. This is in contrast to an end-of-life value of 65,000 hours for thermally upgraded kraft paper when aged in a sealed tube at 110 °C. A DP value of 200 correlated very closely with 50% tensile retention. This contradicts earlier studies in which a DP value of 200 was found in samples with % tensile retention values in the range of 25%.(7) Although one study by Schoff in 1985 stated that the DP dropped to a value of 250 when the tensile strength was reduced to 50% of its original value.(5) The authors are collecting more data to support this aging study. The conclusion to date is that end-of-life value for both 50% tensile retention and a DP value of 200 is close to the life curve given in the IEEE Loading guide of 180,000 hours (20.5 years) when operated continuously at rated load. Furans The results of this study supported the theory that there is a relationship between the concentration of 2-Furaldehyde and the DP of the paper. There was no difference in this relationship between thermally upgraded and non-thermally upgraded paper. Measurement of Furans in transformer oil should enable one to predict remaining insulation life. Prediction of Insulation Life The results of the Dual Temperature Aging Study demonstrate that the criteria given in the IEEE Loading Guide, IEEE C57.91-1995 for distribution transformer models is applicable to power transformers as well. The challenge to the asset manager is how to apply this information when attempting to measure remaining life of a transformer in service. Ideally one should take samples of the insulation periodically for measurement of either tensile strength or DP. However, this is nearly impossible since the critical paper is located within the transformer windings at the hot spot. Periodic measurement of furan content in the oil is therefore the most cost-effective diagnostic for prediction of remaining insulation life. The data from this study supports a strong correlation of furans to DP. 11 REFERENCES 1. IEEE C57.92-1995, Guide for Loading Mineral-Oil-Immersed Transformers 2. IEEE C57.100-1999, IEEE Standard Test Procedure for Thermal Evaluation of Liquid-Immersed Distribution and Power Transformers. 3. Montsinger, V.M., “Loading Transformers by Temperature” AIEE Transactions, vol. 49, 1930 pp. 1151-1162 4. McNutt, W.J. “Insulation Thermal Life Considerations for Transformer Loading Guides”, IEEE Transactions on Power Delivery, vol. 7, no. 1 pp. 392-401, Jan. 1992 5. Shroff, D.H. and Stannet, A.W., “A Review of Paper Aging in Transformers”, IEE Proceedings, vol. 132, pt. C no. 6, Nov. 1985 6. Dakin, T., “Electrical Insulation Deteriation Treated as a Chemical Rate Phenomenon”, AIEE Transactions, vol. 67, 1948 pp. 113-122 7. Oommen, T.V. and Arnold, L.N., “Cellulose Insulation Materials Evaluated by Degree Of Polymerization Measurements”, CH1717-8/81/0000-0257 pp. 257261, 1981 8. De Pablo A., “Recent Research Relating to the Usefulness of Furanic Analysis to Transformer Condition Assessment:,CIGRE, paris, 1998, WG 15-01. 9. Griffin, P.J., Lewand, L.R., Finnan, E., Barry, J., “Measurement of Cellulosic Insulation Degradation”, Minutes of the Sixtieth Annual International Conference of Doble Clients, Doble Engineering Company, 1993, Section 10-3. 10. Chendong, I. , “Monitoring Paper Insulation Aging by Measuring Furfural Contents in Oil”, Seventh International Symposium on High Voltage Engineering, Dresden, August 1991. 11. Stebbins, R.D., Myers, D.S., Shkolnik, A.B., “Furanic Compounds in Dielectric Liquid Samples: Review and Update of Diagnostic Interpretation and Estimation of Insulation Ageing”, Proceedings of the 7th International Conference on Properties and Applications of Dielectric Materials, 2003. Volume 3, 1-5 June 2003 Page(s):921 - 926 vol.3 12