# Comparison of different metal oxide surge arrester models.

1 INTRODUCTIONThe nonlinear feature of the metal oxide surge arrester requires careful consideration of how to incorporate this nonlinearity into a surge arrester model accurately. The most frequently used surge arrester model is that proposed by the IEEE Working Group 3.4.11 [1]. This model was further tested, validated and compared to measurements from laboratory tests by various scientists [2, 3].

Although the IEEE model provided an accurate approximation of the surge arrester behavior, its main draw-back was the complexity of using this model to approximate the surge arrester characteristic using data provided by the manufacturer. To counter this, a simplified model has been proposed by Pinceti and Giannettoni [4, 5] which makes the modeling of the surge arrester considerably easier. Since surge arrester models are essential in a variety of power sys-tem analyses [5-11], it is important to determine the accuracy degree of the models developed.

To achieve this, the IEEE model and the simplified model proposed by Pinceti and Giannettoni will be implemented into the program package EMTP-RV where their accuracy will be analyzed for three surge arresters of different rated voltage. In addition, the process of model implementation into the EMTP-RV program package will be described in detail.

2 MODEL RECOMMENDED BY IEEE WORKING GROUP 3.4.11

The model recommended by the IEEE Working Group 3.4.11 [1] is a frequency--dependent model which combines the characteristics of two nonlinear resistors [A.sub.0] and [A.sub.1] (Fig. 1) to approximate the data provided by the manufacturer effectively.

Other elements that are used in the IEEE model are inductance [L.sub.0] associated with magnetic fields in the immediate vicinity of the surge arrester, stabilizing element [R.sub.0] used to circumvent possible numerical instabilities, the filter between the two nonlinear resistances consisting of [L.sub.1] and [R.sub.1] and, finally, terminal--to--terminal capacitance C of the surge arrester [1].

[FIGURE 1 OMITTED]

These elements are computed from:

[R.sub.0] = 100 x d/n[[OMEGA]], (1)

[L.sub.0] = 0.2 x d/n[[mu]H], (2)

[R.sub.1] = 65 x d/n[[OMEGA]], (3)

[L.sub.1] = 15 x d/n[[mu]H], (4)

C = 100 x n/d[pF]. (5)

where d is the estimated height of the surge arrester in meters and n is the number of parallel columns of metal oxide in the arrester [1].

The [R.sub.1] - [L.sub.1] filter has a low impedance for slow front surges which implies that the two nonlinear resistances [A.sub.0] and [A.sub.1] are connected essentially in parallel. On the other hand, for fast front surges, this filter has a high impedance which implies that the majority of current will pass through the [A.sub.0] nonlinear resistance.

The nonlinear characteristics of the [A.sub.0] and [A.sub.1] resistors in per unit (p.u.) values are presented in Table I [1].

2.1 IEEE Model in EMTP-RV

The implementation of the IEEE surge arrester model into the EMTP-RV environment (Fig. 2) requires the following steps:

* formation of the [A.sub.0] and [A.sub.1] nonlinear resistances (Table 1) using the EMTP-RV functions ZnO and ZnO Data Function,

* calculation of element values given by equations (1-5) and connection of the elements according to Fig. 1,

* adjustment of the referent voltage ([V.sub.ref]) in [A.sub.0] and [A.sub.1] nonlinear resistances to match the switching surge voltages (wave shapes such as 30/60 ([micro]s),

* adjustment of [L.sub.1] to match the residual voltage at 10 kA current surge with a 8/20 [micro]s shape.

In the ZnO Data Function the following values must be entered: a) Data Voltage Rating must equal 1, b) [V.sub.ref] must equal the residual voltage at 10 kA current surge with 8/20 [micro]s shape, and c) Desired Voltage Rating must be half the value of [V.sub.ref]. Initially, [V.sub.ref] in [A.sub.0] and [A.sub.1] nonlinear resistances must equal twice the value of the residual voltage for a 10 kA current surge with 8/20 [micro]s shape.

The procedure for implementing the IEEE model is relatively complicated and time consuming compared to the Pinceti and Giannettoni simplified model [4, 5].

[FIGURE 2 OMITTED]

3 SIMPLIFIED MODEL RECOMMENDED BY PINCETI AND GIANNETTONI

The model recommended by Pinceti and Giannettoni [4, 5] is a simplified version of the IEEE model (Fig. 3). In this simplified version the capacitance C has been removed due to its negligible influence [4]. Furthermore, the resistances [R.sub.0] and [R.sub.1] have been replaced by one resistance R to counter possible numerical instabilities.

[FIGURE 3 OMITTED]

The characteristics of the [A.sub.0] and [A.sub.1] nonlinear resistors are identical to that of the IEEE model (Table 1). The parameters [L.sub.0] and [L.sub.1] of this simplified surge arrester model are computed from:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (6)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (7)

where [U.sub.r] represents the surge arrester rated voltage, [U.sub.rl/T2] is the residual voltage for a 10 kA fast front current surge (1/[T.sub.2]) and [U.sub.r8/20] is the residual voltage for a 10 kA current surge with 8/20 [micro]s shape. As mentioned, the resistance R = 1 M[ohm] is introduced to avoid numerical instabilities

3.1 Simplified Model in EMTP-RV

The implementation of the simplified surge arrester into the EMTP-RV environment (Fig. 4) requires considerably less effort than the IEEE model. The following steps are required:

* formation of the [A.sub.0] and [A.sub.1] nonlinear resistances (Table 1) using the functions ZnO and ZnO Data Function,

* calculation of elements values given by equations (6-7) and connecting the elements according to Fig. 3.

In [A.sub.0] and [A.sub.1], [V.sub.ref] is twice the value of the residual voltage for a 10 kA current surge with 8/20 [micro]s shape.

[FIGURE 4 OMITTED]

4 EXAMPLE OF PARAMETER SELECTION FOR METAL OXIDE SURGE ARRESTERS

In this section the accuracy of the presented models will be analyzed relative to the manufacturer's data [12]. Three metal oxide surge arresters of different rated voltages will be observed. The manufacturer's data of these surge arresters are presented in Table 2, Table 3 and Table 4 [12]. The number of parallel columns n = 1.

4.1 Comparison for surge arrester [U.sub.r] = 10 kV

For the IEEE model, introduction of d and n into equations (1-5) yields: [R.sub.0] = 18.7 [ohm], [L.sub.0] = 0.0374 [micro]H, [R.sub.1] = 12.155 [ohm], [L.sub.1] = 2.805 [micro]H and C = 534.7594 pF. The value of [V.sub.ref] in [A.sub.0] and [A.sub.1] was iteratively changed in order to obtain the best approximation of the residual voltage for the slow front wave shape 30/60 [micro]s for 1 kA current amplitude. This resulted in [V.sub.ref] = 48.025 kV. As mentioned, to adequately approximate fast front surges, the parameter [L.sub.1] must be altered in order to obtain a better approximation. The residual voltages produced by the model are then compared to the residual voltages at 10 kA current surge with an 8/20 [micro]s shape. The best agreement was found for [L.sub.1] = 1.2 [micro]H (Table 5).

As for the simplified model, substituting the manufacturer's data from Table 2 into equations (6-7) the following values of inductances were found: [L.sub.0] = 0.149051 [micro]H and [L.sub.1] = 0.447154 [micro]H. Again, it is important to emphasize that unlike the IEEE model, the computation of the required parameters for the simplified model is a great deal easier.

By connecting a current source in EMTP - RV to the surge arrester model in order to simulate the various current surge waves (Table 2) the residual voltages for both models were obtained and are presented in Table 5.

Furthermore, the percent errors (p.e.) of the two presented models are computed relative to the manufacturer's data and are also given in Table 5.

4.2 Comparison for surge arrester [U.sub.r] = 30 kV

For the IEEE model, introduction of d and n into equations (1-5) yields: [R.sub.0] = 34.7 [ohm], [L.sub.0] = 0.0694 [micro]H, [R.sub.1] = 22.555 [ohm], [L.sub.1] = 5.205 [micro]H and C = 288.1844 pF. The value of [V.sub.ref] in the [A.sub.0] and [A.sub.1] was found to be 144.05 kV whereas the parameter [L.sub.1] was altered to obtain the best agreement to [L.sub.1] = 3.57 [micro]H (Table 6).

As for the simplified model, substituting the manufacturer's data from Table 2 into equations (6-7) the following values of inductances were found: [L.sub.0] = 0.447761 [micro]H and [L.sub.1] = 1.343284 [micro]H. As previously, the residual voltages as well as percent errors for this surge arrester is presented in Table 6.

4.3 Comparison for surge arrester [U.sub.r] = 50 kV

For the IEEE model, introduction of d and n into equations (1-5) yields: [R.sub.0] = 50.7 [ohm], [L.sub.0] = 0.1014 [micro]H, [R.sub.1] = 32.955 [ohm], [L.sub.1] = 7.605 [micro]H and = 197.239 pF. The value of [V.sub.ref] in [A.sub.0] and [A.sub.1] was found to be 240.4 kV whereas the parameter [L.sub.1] was altered to obtain the best agreement to [L.sub.1] = 5.775 [micro]H (Table 7).

As for the simplified model, substituting the manufacturer's data from Table 2 into equations (6-7) the following values of inductances were found: [L.sub.0] = 2.23941 [micro]H and [L.sub.1] = 0.74647 [micro]H. As previously, the residual voltages as well as percent errors for this surge arrester is presented in Table 7.

4.4 Accuracy analysis

For fast front surges of a 1/5 [micro]s wave shape, it is evident that the simplified model yields more accurate results then the IEEE model in all cases.

For fast front surges of an 8/20 [micro]s wave shape, both models display similar accuracy degrees although, for these wave shapes, the simplified model has the highest percent error always for the 20 kA amplitude.

For slow front surges of a 30/60 [micro]s wave shape, the IEEE model displays a greater accuracy degree than the simplified model for all cases.

The highest percent errors of 5.53 % ([U.sub.r] = 10 kV), 5.38 % ([U.sub.r] = 30 kV) and 5.47 % ([U.sub.r] = 50 kV) is displayed by the simplified model always at a 0.125 kA amplitude of the 30/60 [micro]s current wave shape.

5 CONCLUSION

In this paper, two models were compared with respect to the accuracy of their approximation of the manufacturer's data. The comparison was performed in the program package EMTP-RV. The accuracy of the two models was analyzed on three surge arresters of different rated voltage.

In summary, the simplified model proposed by Pinceti and Giannettoni [4] yields highly satisfactory results especially taking into consideration a much easier implementation of the model in the EMTP-RV environment.

REFERENCES

[1.] IEEE Working Group 3.4.11: "Modeling of metal oxide surge arresters", IEEE Transactions on Power Delivery 1992; 7(1):302-309.

[2.] Darveniza. M, Roby. D, Tumma. LR, "Laboratory and analytical studies of the effects of multipulse lightning current on metal oxide arresters", IEEE Transactions on Power Delivery 1994; 9:764-771.

[3.] Darveniza. M, Tumma. LR, Richter. B, Roby. D, "Multipulse lightning currents and metal-oxide arresters", IEEE Transactions on Power Delivery 1997; 12:1168-1175.

[4.] Pinceti. P, Giannettoni. M,: "A simplified model for zinc oxide surge arresters", IEEE Transactions on Power Delivery 1999; 14(2):393-398.

[5.] Magro. MC, Giannettoni. M, Pinceti. P, "Validation of ZnO surge arresters model for overvoltage studies", IEEE Transactions on Power Delivery 2004; 19:1692-1695.

[6.] Goedde. GL, Kojovic. LA, Woodworth. JJ, "Surge arrester characteristics that provide reliable overvoltage protection in distribution and low-voltage systems", in proceedings of IEEE Power Engineering Society Summer Meeting 2000; 4:2375-2380.

[7.] Hayashi. T, Mizuno. Y, Naito. K, "Study on transmission-line arresters for tower with high footing resistance", IEEE Trans-actions on Power Delivery 2008; 23:2456-2460.

[8.] He. J, Hu. J, Gu. S, Zhang. B, Zeng. R, "Analysis and improvement of potential distribution of 1000-kV ultra-high-voltage metal-oxide arrester", IEEE Transactions on Power Delivery 2009; 24:1225-1233.

[9.] Kannus. K, Lahti. K, "Evaluation of the operational condition and reliability of surge arresters used on medium voltage net-works", IEEE Transactions on Power Delivery 2005, 20:745-750.

[10.] Tarasiewicz. EJ, Rimmer. F, Morched. AS, "Transmission line arrester energy, cost, and risk of failure analysis for partially shielded transmission lines", IEEE Transactions on Power Delivery 2000; 15:919-924.

[11.] Tarchini. JA, Gimenez. W, "Line surge arrester selection to improve lightning performance of transmission lines", in proceedings of IEEE Bologna Power Tech Conference Proceedings 2003; 832-837.

[12.] ABB MWK Surge Arrester Datasheet, Medium Voltage Products & Systems.

Dino Lovric, Slavko Vujevic, Tonci Modric

Rudera Boskovica 32, HR-21000, Croatia, dlovric@fesb.hr, vujevic@fesb.hr, tmodric@fesb.hr

Table 1. U-I characteristics for [A.sub.0] and [A.sub.1] U (p.u.) I (kA) [A.sub.0] [A.sub.1] 0.01 0.875 - 0.1 0.963 0.769 1 1.050 0.850 2 1.088 0.894 4 1.125 0.925 6 1.138 0.938 8 1.169 0.956 10 1.188 0.969 12 1.206 0.975 14 1.231 0.988 16 1.250 0.994 18 1.281 1.000 20 1.313 1.006 Table 2. Manufacturer's data of the metal oxide surge arrester [U.sub.r] = 10 kV Wave 1/5 [micro]s I [kA] 1 5 10 U [kV] 29 25.6 29 [U.sub.r] = 10 kV Wave 8/20 [micro]s I [kA] 1 5 10 20 d = 0.187 m U [kV] 20.8 23.2 24.6 27.2 Wave 30/60 [micro]s I [kA] 0.125 0.25 0.5 1 U [kV] 18.1 19 19.7 20.5 Table 3. Manufacturer's data of the metal oxide surge arrester [U.sub.r] = 30 kV Wave 1/5 [micro]s I [kA] 1 5 10 U [kV] 62.9 76.9 86.9 [U.sub.r] = 30 kV Wave 8/20 [micro]s d = 0.347 m I [kA] 1 5 10 20 U [kV] 62.4 69.6 73.7 81.6 Wave 30/60 [micro]s I [kA] 0.125 0.25 0.5 1 U [kV] 54.3 57.1 59.1 61.4 Table 4. Manufacturer's data of the metal oxide surge arrester [U.sub.r] = 50 kV Wave 1/5 [micro]s I [kA] 1 5 10 U [kV] 104.8 128.2 144.8 [U.sub.r] = 50 kV Wave 8/20 [micro]s d = 0.507 m I [kA] 1 5 10 20 U [kV] 104 116 122.8 135.9 Wave 30/60 [micro]s I [kA] 0.125 0.25 0.5 1 U [kV] 90.4 95.2 98.4 102.4 Table 5. Residual voltages and percent errors of the IEEE and the simplified model for surge arrester [U.sub.r] = 10 kV IEEE Pinceti and Giannettoni I (kA) U [kV] p.e. [%] U [kV] p.e. [%] 1/5 [micro]s 1 21.6098 2.90 21.495 2.36 5 26.0781 1.87 25.6742 0.29 10 27.6525 -4.65 28.42 -2.00 8/20 [micro]s 1 20.5591 -1.16 20.9794 0.86 5 23.129 -0.31 23.2907 0.39 10 24.5991 0.00 24.4913 -0.44 20 26.6279 -2.10 26.0806 -4.12 30/60 [micro]s 0.125 18.647 3.02 19.1005 5.53 0.25 19.2226 1.17 19.6874 3.62 0.5 19.8207 0.61 20.2946 3.02 1 20.447 -0.26 20.9252 2.07 Table 6. Residual voltages and percent errors of the IEEE and the simplified model for surge arrester [U.sub.r] = 30 kV IEEE Pinceti and Giannettoni I (kA) U [kV] p.e. [%] U [kV] p.e. [%] 1/5 [micro]s 1 64.752 2.94 64.4022 2.39 5 77.8944 1.29 76.9373 0.05 10 82.4676 -5.10 85.1711 -1.99 8/20 [micro]s 1 61.6575 -1.19 62.8534 0.73 5 69.3307 -0.39 69.7804 0.26 10 73.7028 0.00 73.3799 -0.43 20 79.7204 -2.30 78.2531 -4.10 30/60 [micro]s 0.125 55.9309 3.00 57.223 5.38 0.25 57.6572 0.98 58.9818 3.30 0.5 59.4507 0.59 60.8012 2.88 1 61.328 -0.12 62.6905 2.10 Table 7. Residual voltages and percent errors of the IEEE and the simplified model for surge arrester [U.sub.r] = 50 kV IEEE Pinceti and Giannettoni I (kA) U [kV] p.e. [%] U [kV] p.e. [%] 1/5 [micro]s 1 107.879 2.94 107.309 2.39 5 129.678 1.15 128.2 0.00 10 137.393 -5.12 141.922 -1.99 8/20 [micro]s 1 102.874 -1.08 104.727 0.70 5 115.585 -0.36 116.27 0.23 10 122.799 0.00 122.269 -0.43 20 132.735 -2.33 130.39 -4.05 30/60 [micro]s 0.125 93.3403 3.25 95.3446 5.47 0.25 96.2204 1.07 98.2756 3.23 0.5 99.2118 0.82 101.308 2.96 1 102.342 -0.06 104.456 2.01

Printer friendly Cite/link Email Feedback | |

Author: | Lovric, Dino; Vujevic, Slavko; Modric, Tonci |
---|---|

Publication: | International Journal of Emerging Sciences |

Article Type: | Report |

Geographic Code: | 4EXCR |

Date: | Dec 1, 2011 |

Words: | 3025 |

Previous Article: | The complete 3-D modeling and real-time simulation of the Tesla's boat. |

Next Article: | Immunity against voltage dips--main recommendations to stakeholders of the CIGRE/CIRED/UIE joint working group C4.110. |

Topics: |