A novel algorithm for frequency dependent modeling of ferrite ring and its optimum design to study VFTO behaviors within GIS

Abstract

One of the fundamental issues in gas insulated substations (GIS) which has destructive effects on GIS equipment is the very fast transient over-voltages (VFTOs). This paper models a 400/230 kV substation in order to study the effects of VFTO extensively implemented on EMTP-RV. In addition, the application of ferrite rings for suppressing VFTOs is assessed thoroughly. The main advantage of this paper is its new proposed algorithm according to the ferrite ring frequency dependent modeling that is validated with experimental results. This paper examines the effects of three compositions of the ferrite ring on VFTO suppression. Moreover, it estimates the dimension of the ferrite ring based on the SF6 gas insulation withstand and the maximum effect of ferrite rings on VFTO suppression constraint with the COMSOL multiphysics software. Furthermore, it gains VFTO attenuated percentages due to the installation of the ferrite ring in different GIS nodes. Finally, it analyzes the offered VFTO amendment technique in various GIS switching scenarios.

Keywords

comsol multiphysics / EMTP-RV / ferrite ring / gas insulated substations (GIS) / transient over-voltage / insulation coordination

Introduction

Very fast transient over-voltages (VFTOs) are produced in gas insulated substations (GIS) due to switching breakers, switchgears, disconnectors and/or earth switches. The operation time of disconnectors contacts (DS) is long, sometimes exceeding 0.6 s. Thus, progressive arcs are produced before switching completion. These arcs can be the main cause of VFTOs with high frequencies [1–3]. It is seen in past researches that these over-voltages can reach up to 2.5 per units (pu) with a range of hundreds of kHz to MHz [4]. It should be mentioned that the VFTO domain has completely depended on the GIS configuration [5,6]. VFTOs are created in micro-seconds.

Hence, VFTOs have high steep wave-fronts. Therefore, protective equipment such as surge-arresters cannot control these over-voltages. The time responses of zinc oxide (ZnO) surge-arresters are longer than the steepness time of VFTO. In recent years, various types of plans have been offered regarding the over-voltage protection issue. Opening and closing resistors, resistor-capacitor (RC) filters and ferrite ring installation are some of the methods that have been studied [7–11]. Ferrite rings can magnetically saturate in very high frequencies. An increase in frequency to more than 100 kHz can raise their losses significantly but they have negligible losses in power frequency ranges.

In 1991, ferrite ring initial models were designed [12], most of which were applied in electronic devices. In 2007, a frequency dependent modeling technique was presented [13], but it was not possible for the authors to design a precise model of electromagnetic equipment in the electromagnetic transients program (EMTP) software. For this reason, data analysis and display (DADiSP) software was applied for the modeling of the ferrite ring to study its effect on VFTO. Previous researches on ferrite ring modeling to suppress the VFTO are tabulated in Table 1 [13–18]. EMTP-RV has a convenient framework which provides a compatible area to implement proposed algorithm in real networks with specific substations. Although the ferrite ring has not been installed in a real network, a novel method was determined in protecting the GIS bus-bar from VFTOs [14,18]. This is the best technique for suppressing VFTO. Additionally, installing ferrite rings will not generate any complexity for the GIS bus-bar to design specific insulation schemes [9,13,15,16]. However, none of these researches have paid attention to sulfur hexafluoride (SF6) gas insulation withstand and the intensity of consistent electric fields that can be produced by ferrite ring installations due to SF6 gas insulation distance reduction.

Presenting a novel algorithm on ferrite ring frequency dependent model

The impedance characteristic of ferromagnetic materials varies considerably with frequency. The equivalent impedance of these materials can be expressed in series or parallel based on the electric-magnetic modeling of these materials [19,20]. This impedance consists of a resistance and an inductance which can be distinguished by their permeability coefficients μ_r(f). It should be mentioned that the series model can better show the frequency dependency of ferromagnetic materials than the parallel model [12]. The frequency dependent impedance (z(jf)) which this ring produced against VFTO in the GIS bus-bar is mainly caused by permeability coefficient μ_r(f) and the inductance of the ring (L) [21] which is expressed as

(1)$$\begin{gathered} z\left(\mathrm{j}f\right)=R_f+\mathrm{2\pi }f{L}_f=\mathrm{j2\pi }f{L}_{\mathrm{e}}=\mathrm{2\pi }f{\mu }_{\mathrm{r}}\left(f\right)L \\ =\mathrm{j2\pi }fL\left[{\mu }_{\mathrm{s}}^{\prime }\left(f\right)-\mathrm{j}{\mu }_{\mathrm{s}}^{\prime }\left(f\right)\right] \\ =\mathrm{2\pi }fL{\mu }_{\mathrm{s}}^{\prime }\left(f\right)+\mathrm{j2\pi }fL{\mu }_{\mathrm{s}}^{\prime }\left(f\right) \\ =R_{\mathrm{s}}\left(f\right)+\mathrm{j}{L}_{\mathrm{s}}\left(f\right), \end{gathered}$$

where L can be expressed as

(2)L={\displaystyle \frac{{\mu }_{\mathrm{0}}d}{\mathrm{2}\mathrm{\pi }}}{L}_{\mathrm{n}}{\displaystyle \frac{R}{r}}.

The impedance model in Eqs. (1) and (2) is related to series electrical models of magnetic materials. For extracting the real and the imaginary parts of magnetic permeability out of Eq. (1), Eq. (3) should be used. Presented parameters of Eq. (3) can be gained using Eqs. (4) and (5). These calculation steps have been implemented for three compositions (MnZn, NiZn, 0.8MnZn+0.2NiZn) in Ref. [22]. The result of this study is represented in Fig. 1. By applying least squares fitting curve methods, the amounts of Eqs. (4) and (5) are calculated, as listed in Table 2. Therefore, ferrite ring frequency dependent modeling can be implemented based on Eqs. (1)–(5) in an EMTP software environment. Figure 2 depicts the proposed ferrite ring frequency dependent algorithm.

(3){\mu }_{\mathrm{r}}\left(f\right)=\mathrm{1}+{\tilde{x}}_{\mathrm{dw}}\left(f\right)+{\tilde{x}}_{\mathrm{rot}}\left(f\right),

(4){\tilde{x}}_{\mathrm{dw}}\left(f\right)={\displaystyle \frac{x_{\mathrm{dw}}{f}_{\mathrm{dw}}^{\mathrm{2}}}{f_{\mathrm{dw}}^{\mathrm{2}}-f^{\mathrm{2}}+\mathrm{j}\beta f/m_{\mathrm{dw}}}},

(5){\tilde{x}}_{\mathrm{rot}}\left(f\right)={\displaystyle \frac{(f_{\mathrm{rot}}+\mathrm{j}\alpha f)x_{\mathrm{rot}}{f}_{\mathrm{rot}}}{{(f_{\mathrm{rot}}+\mathrm{j}\alpha f)}^{\mathrm{2}}-f^{\mathrm{2}}}}.

Validating proposed algorithm

For testing the accuracy and applicability of the proposed algorithm, this paper studies the case mentioned in Ref. [21] in order to compare the simulated outputs of the proposed algorithm with those of the case mentioned in Ref. [21]. A 400 V coaxial is applied in Ref. [21] as an unloaded GIS bus-bar for ferrite ring frequency dependent modeling in order to study the effects of the ferrite ring on VFTO. Transmission line equations and PI model are applied to model the GIS bus-bar in the DADiSP software, but in this paper, the constant parameter (CP) model that is presented in Table 3 is used for modeling the GIS bus-bar in an EMTP-RV software environment. In addition, the arc model is based on Table 3 (R₀ = 1000 Ω, τ = 1 ns and r₀ = 10 Ω). All considerations are based on Fig. 3 that presents the equivalent circuit in Ref. [21]. As demonstrated in Figs. 4 and 5, the proposed algorithm in this paper has accurate similar output waveforms, domains, steepness times and fluctuations to the real experimental test in comparison with the results in Ref. [21]. These similarities are especially in VFTO steepness and fluctuations which cannot be seen in the results in Ref. [21]. As a result, this innovative technique is much more accurate than the method mentioned in Ref. [21].

Case study modeling and determining scenarios for VFTO assessment

This paper takes the Boushehr GIS (400/230 kV) mentioned in Ref. [23] for case study. Samples of equipment destructions have been reported in this substation due to transient over-voltages. Hence, VFTO and ferrite ring control function studies are been thoroughly considered. In addition, this paper focuses on the accurate modeling of the GIS equipment [24–27]. The schematic diagram of this substation is displayed in Fig. 6. The applied models and values in VFTO studies are based on Table 3 [1,7]. Various switching scenarios have been simulated for assessing VFTOs in this substation. Of all simulated situations, five scenarios are selected as the worst VFTO scenarios based on Table 4. The trend of maximum VFTO along the switching way in scenario 1 is depicted in Fig. 9.

After passing through the auto-transformer and reaching the low voltage side, the VFTO domain suppresses. This suppression can be seen more clearly when VFTO passes along the XLPE cable. Moreover, Figs. 8–17 indicate that the VFTO domain is considerably great on one side of the breaker in scenarios 1, 2 and 3, then it suppresses by passing along the GIS route, and finally it reinforces again.

This over-voltage occurs on both sides of the breaker in scenarios 4 and 5 and it rises by spreading along the GIS route extremely. The wave travels in GIS routes without split and/or the breaker increases the VFTO amplitude consistently in all these scenarios so that the voltage domain has an ascending trend in TOFF11 to AS1 auto-transformer and/or in TOFF10 to AS2 auto-transformer. It should be mentioned that if VFTO waves reach AS1 and AS2, they will transfer to the low voltage side with severe suppression.

The XLPE cable effect on wave transmitting process is recuperative which attenuates VFTO amplitudes. The VFTO domain reaches a maximum of 1.6 pu in scenarios 1, 2 and 3. In these scenarios, the maximum over-voltages occur in the high voltage side of auto-transformers. In scenarios 4 and 5, the maximum over-voltages occur on the other side of the breaker. Hence, the maximum over-voltage reaches 3.1 pu in the GTX1 generation terminal of scenario 5.

The optimum design of ferrite ring dimensions based on SF6 gas insulation coordination and maximum VFTO reduction constraint

Ferrite ring is a conductor that can be placed between a bus-bar conductor and a bus-duct wall, which can reduce the SF6 gas insulation distance. The inner (r) and outer (R) radiuses of ferrite rings determine the amount of insulation distance reduction. In general, the SF6 gas insulation withstand level is three times higher than the insulation like the air. A consistent electric field can be calculated with Eq. (6) [28]:

(6)E_{\mathrm{b}}=88.5+{\displaystyle \frac{\mathrm{0.65}}{d_{\mathrm{e}}}}\left(\mathrm{kV}/\mathrm{cm}\right).

Reducing the dimension of the ferrite ring can decrease inductance (L) that will have a less VFTO suppression effect. Hence, a trade between the dimension and inductance should be done in order to find the optimum dimensions for ferrite rings. For this reason, two essential constraints are considered: maximum VFTO suppression and allowed SF6 gas electric field. The dimension of the ferrite ring has been sized proportional to case study the GIS bus-bar at nominal network voltage through COMSOL Multiphysics program. The GIS bus-bars have earthed coaxial aluminum cylinders bus-ducts. The outer cylinder has an inner radius (R₂ = 24.2 cm). The tubular conductor is the kind of bar that has an outer radius (R₁ = 7.62 cm). Three changes of the dimension of the ferrite ring are studied in this paper. First, the inner radius of the ferrite ring is 9 cm but the outer radius varies from 10 to 23 cm (Appendix I, Cases 1 to 7). In this strategy, the maximum generated field occurs between the ring and the GIS bus-bar (1.38 cm distance between r and R₁). The second strategy that allocates 8 to 10 cases of Appendix has an outer radius of 23 cm but the inner radius varies from 10 to 12 cm. Similar to the first mode, the maximum field generates between the ring and the GIS bus-bar.

In the third strategy (Appendix I, Cases 11 to 13), the ring and the GIS bus-bar connects directly to each other to have the same potential. The outer radius varies from 20 to 20.04 cm. Thus, the maximum field is generated between the ring and the GIS earthed bus-duct. After studying all above modes (based on Appendix I), the results show that the best case for considering both SF6 gas withstand limit and maximum VFTO suppression constraints is when the ferrite ring is connected directly to the bus-bar with a 20cm outer radius (Case 13). The electric field strength and voltage changes in the optimum case are presented in Fig. 7.

Ferrite ring effect on VFTO suppression

In this section, the proposed idea of frequency dependent modeling with optimized ferrite ring dimensions has been implemented on an applicable case study. Hence, the compositions of frequency dependent models of MnZn, NiZn and 0.8MnZn+0.2NiZn have been compared with a RL constant model in recent researches. Figures 8–17 illustrate the results of the five scenarios. If ferrite rings are applied in GIS bus-bars, the VFTO domain will suppress considerably in different points of switching route. After comparing different compositions, it is concluded in this paper that MnZn is the best composition that has a maximum effect on VFTO improvement. NiZn and 0.8MnZn+0.2NiZn have healing effects on VFTOs. The differences between ferrite ring frequency dependent technique (presented in three compositions) and parallel RL models can be assessed in the simulations presented. According to Figs. 8–17, the suppressing the VFTO domain of the ferrite ring parallel RL model is similar to that of the 0.8MnZn+0.2NiZn frequency dependent model. Thus, it has less impact on VFTOs in comparison with the MnZn or NiZn models. In scenario 4, existing MnZn and NiZn ferrite rings have limited VFTO near its initial value which has a negative domain. In scenario 1, 2 and 3 (Figs. 9, 11 and 13), VFTO fluctuations have smooth trends along the switching route but Figs. 15 and 17 show that these flat modes cannot be seen in scenarios 4 and 5. By applying ferrite rings in scenario 4, the VFTO suppression trends will be flat on one side of the switch (AL02 to X6). But on the other side (AS1HV to X6), this uniform mode cannot be observed. Moreover, on some of route points (TOFF10 to X6), the VFTO domain increases in comparison with the operational mode without the ferrite ring. Moreover, in TOFF10 to X6, the VFTO domain increases in comparison with the operational mode without the ferrite ring. This unfavorable effect is more tangible when 0.8MnZn+0.2NiZn ferrite rings are applied. Additionally, this ferrite ring effect can be seen in Fig. 17 in X5 point of scenario 5 but this can be negligible in comparison with scenario 4. However, the effect of MnZn and NiZn ferrite rings on both sides of switching locations in scenario 5 can be recognized in Fig.17.

In accordance with the remedial effect of a ferrite ring on GTX1 generator side with 3 pu VFTO, which is better than NiZn, the lowering effect of the MnZn ferrite ring on the other side of the switch can be negligible. In general, the installation of the MnZn ferrite ring is recommended in all scenarios because this composition has a better effect on VFTOs in comparison with other compositions. Table 5 presents suppressed percentages of VFTO domains in different scenarios applying MnZn ferrite ring. On one side of switching locations (GTX1 to X5), the MnZn ferrite ring has a more suppression effect in comparison with NiZn, but on the other side of the switch (AS1HV to X5), NiZn has a better suppression effect.

Conclusions

This paper proposed an adequate framework for frequency dependent modeling of ferrite rings and evaluation of the dimension of the ferrite ring. Hence, a relevant case study based on Ref. [21] was conducted in order to test the accuracy and the applicability of the model proposed. The results obtained using the model proposed are very similar to the measured results. The MnZn, NiZn and 0.8MnZn+0.2NiZn compositions were applied in frequency dependent modeling of the ferrite ring in very fast transient studies. Moreover, the ferrite ring frequency dependent model and the constant RL model, which were considered in recent researches, were distinguished from each other, and this paper proved the necessity of frequency dependent modeling of ferrite rings.

For determining the dimension of the optimum ferrite ring to observe SF6 gas sustainable electric fields and the maximum VFTO suppression, a comprehensive simulation was accomplished using the COMSOL Multiphysics software. When the ferrite ring with an outer radius of 20 cm was connected completely to the bus-bar conductor (and have), it kept generating electric fields of SF6 gas on the admissible range that also has the maximum VFTO suppression power. In this situation, the ferrite ring had the maximum dimension with the higher impedance than in other cases. On the other hand, the network studied in this paper had a better accurate/extensive model than other past researches. The VFTOs reached up to 3 pu in the considered system in simulations using the EMTP-RV software. In this scenario, the installation of the ferrite ring could suppress the VFTO domain up to 78%. Simulated results illustrated that the MnZn ferrite ring had a better effect on VFTO reduction because of its high permeability in comparison with the other two. Therefore, applying MnZn could suppress VFTOs more efficiently.

Appendix: Data of optimum ferrite ring dimension studies

References

[1] Kumar V V, Thomas M J, Naidu M S. Influence of switching conditions on the VFTO magnitudes in a GIS. IEEE Transactions on Power Delivery, 2001, 16(4): 539–544

[2] Lu T C, Zhang B. Calculation of very fast transient over voltages in GIS. In: Proceedings of IEEE/PES Transmission and Distribution Conference & Exhibition: Asia and Pacific. Dalian, China, 2005, 1–5

[3] Ecklin G, Schlicht D, Plessel A. Over voltages in GIS caused by the operation of isolators. In: Ragaller K, ed. Surges in High Voltage Networks. New York: Plenum Press, 1980

[4] Liu Q. Study of protection of transformer from very fast transient over-voltage in 750 kV GIS. In: Proceedings of the Eighth International Conference on Electrical Machines and Systems (ICEMS), 2005, 3: 2153–2156

[5] CIGRE Working Group. Very fast transient phenomena associated with GIS. CIGRE Report No. 33-13, 1988, 1–20

[6] Meppelink J, Diederich K, Feser K, Pfaff W. Very fast transients in GIS. IEEE Transactions on Power Delivery, 1989, 4(1): 223–233

[7] Tavakoli A, Gholami A, Parizad A, Soheilipour H M, Nouri H. Effective factors on the very fast transient currents and voltage in the GIS. In: 2009 IEEE T&D Conference and Exposition: Asia and Pacific. Seoul, South Korea, 2009, 1–6

[8] Xiang Z T, Liu W D, Qian J L, Zhang Y L, Shen Y Z, Wang S P. High voltage simulation tests of suppressing VFTO in GIS by magnetic rings. Transaction of China Electrotechnical Society, 2004, 19(7): 1–3

[9] Jin L J, Liu W D, Qian J L. Research on the suppression of VFTO in GIS by ferrite rings. High Voltage Engineering, 2002, 28(7): 1–3 (in Chinese)

[10] Liu W D, Jin L J, Qian J L, Zhang Y L, Shen Y Z, Wang S P. Possibility of suppressing VFTO in GIS by ferrite rings. Transaction of China Electrotechnical Society, 2002, 17(4): 22–25

[11] Liu W D, Jin L J, Qian J L. Simulation test of suppressing VFT in GIS by ferrite rings. In: Proceedings of International Symposium of Electrical Insulating Material. Himeji, Japan, 2001, 245–247

[12] Blanken P G, van Vlerken J J L M. Modeling of electromagnetic systems. IEEE Transactions on Magnetics, 1991, 27(6): 4509–4515

[13] Thomas D E, Wiggins E M, Salas T M, Nickle F S, Wright S E. Induced transients in substation cables: measurements and models. IEEE Transactions on Power Delivery, 1994, 9(4): 1861–1868

[14] Rama Rao J V G, Amarnath J, Kamakshaiah S. Simulation and experimental method for the suppressing of very fast transient over-voltages in a 245 kV GIS using ferrite rings. In: Proceedings of International Conference on High Voltage Engineering and Application (ICHVE). New Orleans, USA, 2010, 128–131

[15] Li H, Gan L, Li L, Zhou Y, Xu Y. Research on protection measure for very fast transient over-voltage of GIS. In: Proceedings of 2009 Asia-Pacific Power and Energy Engineering Conference. Wuhan, China, 2009

[16] Rama Rao J V G, Amarnath J, Kamakshaiah S. Accurate modeling of very fast transient overvoltages in a 245 kV GIS and research on protection measures. In: Proceedings of the IEEE 2010 Conference on Electrical Insulation and Dielectric Phenomena (CEIDP), 2010

[17] Jin L J, Zheng Y B, Ge P, Zheng Y, Zheng X G. Estimating the size of ferrite for suppressing VFTO in GIS. In: Proceedings of 2006 the 8th International Conference on Properties and applications of Dielectric Materials. Bali, Indonesia, 2006, 388–391

[18] Jin L J, Zhang Y C, Zheng X G, Shen Y Z. Characteristic parameter analysis on the suppressing VFTO in GIS by ferrite. In: Proceedings of 2005 International Symposium on Electrical Insulating Materials. Kitakyushu, Japan, 2005, 825–828

[19] Qian J L, Zhang J R, Gu J Q. Open Shunt Capacity Current and Small Inductance Current by High Voltage Switch. Beijing: Power Publisher, 1999 (in Chinese)

[20] Watson J K, Amoni S. Using parallel complex permeability for ferrite characterization. IEEE Transactions on Magnetics, 1989, 25(5): 4224–4226

[21] Li Q, Wu M. Simulation method for the applications of ferromagnetic materials in suppressing high-frequency transients within GIS. IEEE Transactions on Power Delivery, 2007, 22(3): 1628–1632

[22] Dosoudil R, Usakova M, Slama J, Gruskova A. Frequency variation of complex permeability in dual ferrite filler single polymeric matrix composites. In: Proceedings of CSMAG'07 Conference. Kosice, Slovakia, 2007, 621–624

[23] Majidi M, Javadi H, Oskuoee M. Ferroresonance evaluation at boushehr 230/400 kV GIS substation of Iran’s power network. In: Proceedings of 2011 International Conference on Electric and Electronics (EEIC2011). Nanchang, China, 2011

[24] Rao M M, Thomas M J, Singh B P. Frequency characteristics of very fast transient currents in a 245 kV GIS. IEEE Transactions on Power Delivery, 2005, 20(4): 2450–2457

[25] Flohr T, Pfeiffer W, Zender C, Zimmer V. Diagnosis of prebreakdown phenomena in SF6 in Non-uniform gaps for very fast transient voltage stress. In: Conference Record of the 1994 IEEE International Symposium on Electrical Insulation. Pittsburgh, USA, 1994, 603–606

[26] Amamath J, Paramahamsa D R K, Narasimharao K, Singh B P, Shrivastava K D. Very fast transient over-voltages and transient enclosure voltages in gas insulated substations. In: 2003 Annual Report Conference on Electrical Insulation and Dielectric Phenomena. Albuquerque, USA, 2003, 506–509

[27] Pinches D S, Al-Tai M A. Very fast transient overvoltages generated by gas insulated substations. In: 2008 the 43rd International Universities Power Engineering Conference. Padova, Italy, 2008, 1–5

[28] Kuffel J, Zaengl W S, Kuffel E. High Voltage Engineering: Fundamentals (Second Edition). Newnes (Butterworth-Heinemann Ltd.), UK, 2000

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg