1

Effects of Terminal Connections, Noise and DC Magnetism on Sweep Frequency Response Analysis of Transformers
Esam Al Murawwi TRANSCO UAE Redy Mardiana The Petroleum Institution UAE Charles Q. Su Charling Technology Ltd Australia Braham Barkat The Petroleum Institution UAE

Abstract- A transformer is a very important component in any electrical network and losing it can cause serious problems. To ensure its healthiness, the transformer should be tested in its different stages starting from the moment it is manufactured until the moment it is connected to the network. To ensure that, analysts use different tests, Sweep Frequency Response Analysis (SFRA) test is one of the main tests to be done on the transformer. This paper presents some case studies on the effects of terminal connections, noise and DC Magnetism on SFRA results. Those case studies are taken from different 40 MVA power transformers used in Abu Dhabi Transmission and Dispatch Company (TRANSCO). Terminal connections may lead to confusing SFRA results which could be misinterpreted in the assessment of transformer winding conditions. The effects of grounding the tertiary winding, opening the tertiary winding and connecting the neutral point in the terminal connections are highlighted in this work. Also, the effect of noise and residual magnetism of the core has to be taken into consideration while performing the SFRA test and interoperating the results.

3. 4.

check its condition as part of in-service maintenance, and check its condition after an earthquake or other natural disaster, or after a lightning strike.

The frequency response can be divided into four regions as depicted in Figure 1 [5]. These regions are: 1. The lower frequency region, dominated by the core 2. The middle frequency region, dominated by the interactions between the windings 3. The higher frequency region, dominated by the internal connections of the individual winding structures, and 4. The highest frequency region (usually above 1MHz), dominated by the measurement set-up and connection leads. Note that the boundaries between the different regions can vary from transformer to transformer.

I. INTRODUCTION Power transformers are considered major components in the electrical network and losing them can cause serious power outage. Moreover, a lot of problems can arise in power transformers starting from day one at the manufacturing stage until the day of energizing them [1, 2]. To ensure the healthiness of those transformers, engineers use different tests procedures and an example of the tests is Sweep Frequency Response Analysis (SFRA). SFRA is used by many industries and manufacturers to assess the mechanical integrity of the core, winding, and clamping structure of power transformers. It is basically an injection of a low voltage signal at one of the transformers terminal, over the frequency range 20 Hz 2 MHz, and measurement of the response at another terminal [1]. SFRA is used at several phases during the life-time of a transformer. It is used to: 1. test a new transformer at the factory, in order to create a reference fingerprint 2. check its condition after transportation to the site where it is to be commissioned

Figure 1 Four sweep frequency response analysis (SFRA) spectral regions, and the associated dominant parts of the transformer.

Analysis of transformer condition is commonly based on frequency spectrum comparison [1], [19]. There are mainly three strategies to do so which are: 1. Time-based (a comparison between current SFRA results and previous results for the same unit)

2 2. Type-based (a comparison between SFRA results for one transformer with the corresponding results for another transformer of the same type) Design-based (a comparison between the three phases of the same transformer)

3.

The test engineer should start with the time-based comparison, if previous SFRA results are available. Note that the terminal connections must be the same as in the previous measurement, if a meaningful comparison is to be obtained. Thus photographs of the connections are essential. If previous results are not available, a type-based comparison is recommended. If neither a time-based nor a type-based comparison is possible, a design-based comparison should be carried out. Two or all three types of comparison can be combined in order to achieve better condition assessment. This paper will cover the effects of the terminal connections, noise and DC magnetism on SFRA results.

Figure 3 SFRA injection into the HV winding, with the LV winding (except the neutral) and the tertiary winding short-circuited [HV LV (SC)]

II. EFFECT OF TERMINAL CONNECTIONS Differences between SFRA traces, due to inconsistent terminal connections, can be as big as those due to a fault. Therefore, special care must be taken with terminal connections. Commonly, four SFRA tests with different terminal connections are carried out on a three-winding transformer: 1. Injection into the HV winding, with the LV winding open and the tertiary winding shortcircuited (Figure 2) 2. Injection into the HV winding with the LV winding (except the neutral) is short-circuited and the tertiary winding short-circuited (Figure 3) 3. Injection into the LV winding with the HV winding open and the tertiary winding shortcircuited (Figure 4) 4. Injection into the tertiary winding with the HV and LV windings open (Figure 5) Figures 2 to 5 show these four cases. It is important that the connections be made accurately for each case.

Figure 4 SFRA injection into the LV winding, with the HV winding open and the tertiary winding short-circuited [LV HV (OC)]

Figure 5 SFRA injection into the tertiary winding, with the HV and LV windings open [TV LV and HV (OC)]

Three cases will be covered in this section that shows the effect of terminal connection. The different cases are taken from 40 MVA power transformers with 132/11 KV. Two cases will cover the connections for the tertiary winding. Figure 6 shows the three different connections for the tertiary busing. The tertiary may be kept open (Figure 6-a), closed (Figure 6-b), or closed and grounded (Figure 6-c). Moreover, the last case will highlight the effect of including the neutral in the shorting for the connection in Figure 3.
3u 3w 3u 3w 3u 3w

a.

b.

c.

Figure 2 SFRA injection into the HV winding, with the LV winding open and the tertiary winding short-circuited [HV LV (OC)]

Figure 6 Different connection for the tertiary winding: a. tertiary opened; b. tertiary closed; c. tertiary grounded

3 A. Grounding the Tertiary Winding

Normal practice is to close the tertiary winding when performing SFRA measurements. However, if the tertiary winding is grounded, very different data may be obtained. Figures 7 to 9 present results for three different configurations of SFRA testing, in each case (a) with the tertiary winding closed, (b) with the tertiary winding closed and grounded.
0 -20

-10

magnitude (dB)

-20 -30 -40 -50 10 2u 2v 2w

10

10

10

10

frequency (Hz)
magnitude (dB)

a.
0

-40 -60
-10

-80 -100

1U 1V 1W 10
2

magnitude (dB)

-20 -30 -40 -50 10 2u 2v 2w

10

10 frequency (Hz)

10

10

a.
0 -20

10

10

10

10

frequency (Hz)

magnitude (dB)

-40 -60 -80 -100 10 1U 1V 1W

b. Figure 9 Injection into the LV winding, with the HV winding open circuited: a. tertiary winding short-circuited, b. tertiary winding shortcircuited and grounded.

10

10

10

10

frequency (Hz)

b.
Figure 7 Injection into the HV winding, with the LV winding open circuit: a. tertiary winding short-circuited, b. tertiary winding shortcircuited and grounded.
0 -10

Figures 7 to 9 show that the grounding of the tertiary winding has a larger effect on the measurements when the voltage is injected into the LV winding, rather than into the HV winding. The reason is that the LV winding is closer to the tertiary winding than is the HV winding, so that the capacitive coupling to the LV winding is greater. B. Keeping the Tertiary Winding Open The tertiary winding may also be open-circuited during measurements. Two cases were investigated, again on a 40 MVA transformer: 1. Injection into the HV winding, with the LV winding open-circuited, tertiary winding shortcircuited or open-circuited (Figure 10) 2. Injection into the LV winding, with the HV winding open-circuited, tertiary winding shortcircuited or open-circuited (Figure 11)
-10 -20 Tertiary short-circuited Tertiary opened

magnitude (dB)

-20 -30 -40 -50 -60 -70 10 1U 1V 1W

10

10 frequency (Hz)

10

10

a.
0

magnitude (dB)

-20

-30

magnitude (dB)
2

-40 -50 -60 -70 -80

-40

-60

1U 1V 1W 10 10
3

-80

10 frequency (Hz)

10

10

-90 10

10

10

10

10

b. Figure 8 Injection into the HV winding, with the LV winding short circuited: a. tertiary winding short-circuited, b. tertiary winding shortcircuited and grounded.

frequency (Hz)

Figure 10 Injection into the HV winding, with the LV winding opencircuited, tertiary winding short-circuited or open-circuited

4
0

III. EFFECT OF NOISE ON SFRA RESULTS To ensure good measurement and repeatability of SFRA testing, the transformer should be isolated. Also, any interference should be avoided. Figure 13 shows that a noisy signal came from a drilling activity occurring next to the transformer area. This drilling activity affected the SFRA curves as illustrated in Figure 13 in the low frequency region. After stopping the drilling, normal results were found. It is also suggested while repeating SFRA tests in future measurement that noise or interference must be avoided to get good repeatable results for easy comparison.
0

-10

magnitude (dB)

-20

-30

-40 Tertiary short-circuited Tertiary opened -50 10

10

10

10

10

frequency (Hz)

Figure 11 Injection into the LV winding, with the HV winding opencircuited, tertiary winding short-circuited or open-circuited.

-20
magnitude (dB)

Apart from the differences in the high frequency region seen in Figure 11, in the mid-frequency region a double peak (indicated by a circle) appears when the tertiary is open- circuited, but a single peak when it is short-circuited (closed). The same result was observed for each of the three phases of the transformer. The double peak distinguishes the tertiary winding open-circuited case from the shortcircuited and short-circuited and grounded cases. C. Including the Neutral in the Terminal Connection In normal SFRA practice, when the voltage is injected into the HV winding side and the LV winding is shortcircuited, the LV terminals are short-circuited but not connected to the neutral point.
0 -10 Without neutral With neutral

-40 -60 -80 -100 10

10

10

10

10

frequency (Hz)
Figure 13 Noisy signal of SFRA signal

IV. EFFECT OF DC MAGNETISM ON SFRA RESULTS Residual Magnetism is a phenomenon where the transformers core is being magnetized due to an applied DC bias. There are some transformer testing that requires injection of a DC bias such as winding resistance test. SFRA test should be done before any DC injection during factory stage or during commissioning stage. During maintenance stage or other future SFRA testing, documents and identification of any DC injection should be mentioned and recorded. The effect of DC injection is mainly shown in the low frequency region where there is a shift in the graph as illustrated in Figure 14.
0 -20
magnitude (dB)

magnitude (dB)

-20 -30 -40 -50 -60 -70 10

10

10

10

10

frequency (Hz)

Figure 12 Injection into the HV winding, LV winding short- circuited: LV terminals connected or not connected to the neutral point.

Before shifting (without DC injection) After shifting (with DC injection)

-40 -60 -80 -100 10

Figure 12 shows the difference between connecting and not connecting the LV terminals to the neutral point. Consistent practice is essential in this point, if meaningful conclusions are to be drawn from the measurements. It is therefore again recommended that photographs of the terminal connections be taken and retained for future reference.

10

10

frequency (Hz)
Figure 14 Effect of DC injection

5 V. PROCEDURES TO ACHIEVE RELIABILITY The following precautions and recommendations should be followed while performing the SFRA test to achieve reliability [1], [17], [18], [19]: 1. Measurements should be made under the same transformer conditions, e.g., assembled, oil filled, same tap changer, etc. 2. The status of the transformer should be noted (e.g., transformer age, number of failures, type of failures, etc). 3. If possible the same measuring instrument should be used for future tests. Otherwise, the use of an instrument with the same frequency range, accuracy range, settling time, and applied voltage is recommended. 4. Three shielded high frequency cables should be used (source, reference and measurement). 5. The test leads should be as short as possible. 6. Ground extensions of the test leads should have low inductance. 7. Ground extensions of the measuring cable shields should not make electrical contact with the terminals. 8. Electrical noise should be eliminated as far as possible. 9. SFRA measurements should be carried out before and after moving the transformer. 10. If reference data are available, they should be available in both hardcopy and softcopy in case they need to be uploaded for easy comparison during future tests. 11. Photos of the SFRA terminal connections should be taken and made available for any future SFRA tests. In such tests the same connections should be used. 12. The delta TV winding should be closed when testing HV and LV windings, and it should be noted in the test report whether the delta TV winding was earthed at one point or left floating.
[1]

REFERENCES
Kraetge, M. Kruger, M.Viljoen, and A. Dierks, Aspects of Practical Applications of Sweep Frequency Response Analysis (SFRA) on Power Transformers. 6th Southern Africa Regional Conference, CIGRE 2009. M. de Nigris, R. Passglia, R. Berti, L. Bergonzi, and R. Maggi, Application of Modern Technologies for the Condition Assessment of Power Transformers. CIGRE 2004. The Electric Power Industry Standard of Peoples Republic of China, Frequency Response Analysis on Winding Deformation of Power Transformers, DL Standard, DL/T911-2004, ICS 27.100, F24, Document No. 15182-2005, June 2005. CIGRE Working Group A2/26, Mechanical condition Assessment of Transformer Windings using Frequency Response Analysis (FRA). Paris, April 2008. IEC 60076-18 Ed.1: Power Transformers-Part 18: Measurement of Frequency Response. 2009 (not yet published). IEEE PC57.149TM/D4 Draft, Trial-Use Guide for the Application and Interpretation of Frequency Response Analysis for Oil Immersed Transformers, 2007 draft version (not yet published). L. Satish, and A. Saravanakumar, Identification of terminal connection and system function for sensitive frequency response measurement on transformers, IEEE Trans. Power Deliv., vol.23, no.2, April 2008. N.Abeywickrama, Y.V.Serdyuk, and S.M.Gubanki, High-frequency Modeling of Power Transformers for use in Frequency Response Analysis (FRA), IEEE Trans. Power Deliv., Vol. 23.No.4, October 2008. R. Ragavan and L. Satish, An Effective method to compute transfer function of a Transformer from its equivalent circuit, IEEE Trans. Power Deliv., vol.20, no2, pp.780-788, April 2005. M. Florkowski and J. Furgal, Detection of transformer winding deformations based on the transfer function Measurements and simulations, Meas. Sci. Technol., vol.14, pp. 1986-92, Sep. 2003. M. Wang, A. J. Vandermaar, and K. D. Srivastava, Improved Detection of Power Transformer Winding Movement by Extending the FRA High Frequency Range, IEEE Trans. Power Deliv., vol. 20 . no.3, July 2007. A. Shintemirov, W. H. Tang, Q. H. Wu, A Hybrid Winding Model of Disc-Type Power Transformers for Frequency Response Analysis, IEEE Trans. Power Deliv., Vol. 24.No.2, April 2009. S. Ryder, Diagnosing a wider range of transformer faults using frequency response analysis, presented at the 13th Int. Symp. High Voltage Engineering, 2003. S. A. Ryder, Diagnosis transformer faults using frequency response analysis, Electr. Insul. Mag., vol. 19, no. 2, pp. 16-22, 2003. S. A. Ryder, Transformer diagnosis using frequency response analysis: results from fault simulation, Power Engineering Society Summer Meeting, vol. 1, pp. 399-404, 2002. K.G.N.B. Abeywickrama, Y.V. Serdyuk, and S.M. Gubanski, Exploring possibilities for characterization of power transformer insulation by frequency response analysis (FRA), IEEE Trans. Power Deliv., vol.22, no.3, pp.1375-1382, July 2006. M. Wang, John Vandermaar, and K.D. Srivastava, Transformer Winding Movement Monitoring in Service- Key Factors Affecting FRA measurements, IEEE Electrical Insulation Magazine, vol. 20, no. 5. September/October 2004. Mats Karlstorm, Peter Werelius, Matz Ohlen, Lars Adeen, and Eddie Brynjebo, Considerations to Ensure Measurement Repeatability when using SFRA on Transformers E. Al Murawwi, R. Mardiana, Q. Su, Effect of terminal connections on Sweep Frequency Response Analysis of transformers, IEEE insulation magazine, May/Jun, 2012

[2]

[3]

[4]

[5] [6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14] [15]

[16]

[17]

VI. CONCLUSION The effects of terminal connections on SFRA test results for transformers with three windings have been presented and discussed. Different terminal connections may yield different SFRA responses, and may lead to incorrect diagnosis of the transformer condition. It is essential to make identical terminal connections for a given test type. Moreover, other factors such as noise and DC magnetism may affect SFRA results as well. Special considerations have to be made to avoid the above mentioned factors while doing the SFRA measurements to get reliable results.
[18]

[19]