**2. Gas-filled surge arresters (GFSA)**

The basic design of gas-filled surge arresters is a simple configuration of two or three electrodes encased in a glass or ceramic enclosure filled with insulating gas (**Figure 1**) [9]. The common insulation medium of choice is noble gases [10]. The form of the electrodes is such that it provides a pseudo-homogeneous macro component of the electric field [11]. The strength of the electric field impacts the ratio of free electron generation and loss in the insulating gas. If the field is strong enough, a self-sustained avalanche process will lead to the electrical breakdown of the gas and discharge of the overvoltage [12, 13]. The performance of GFSA is best described by its pulse shape characteristic. The narrower the pulse characteristic is, and the smaller the slope is, the better protective characteristics of GFSA are [6, 7].

Solutions to improve the performance of GFSA in response time include the use of hollow cathodes, thus avoiding radioactive materials or implementing internal radioactive sources [14, 15]. The differences in regulations and standards and the lack of clear strategy for their clear storage and disposal make radioactive sources a challenging solution in the context of radiation protection and potential environmental contamination. Still, there are some natural (cosmic radiation, solar flares, coronal mass ejection) and man-made phenomena (nuclear explosions, especially at high altitude) where the power grid, electrical machinery, and electronic components can experience voltage transients under ionizing radiation.

In order to describe the electrical breakdown of the insulating gas, avalanche coefficients are used to specify the elementary processes in GFSA's gas. Avalanche coefficients that are most commonly used are ɑ, the number of electronic ionization collisions per cm of a distance in the direction of the electric field; ƞ, the number of electrons per cm of a distance in the direction of the field attached to the electrically negative atoms or molecules; and γ, the number of electrons generated

**217**

*Influence of Gamma Radiation on Gas-Filled Surge Arresters*

from secondary processes per primary avalanches. The avalanche coefficients vary with pressure and electric field and are not constant for any particular gas or a gas mixture. This dependence has been experimentally determined for most insulating gases and is described as a range of values of the *pd* product, where p is the insulator

The electrical breakdown in gases can be dynamic or static, not depending on the precise mechanism of the electrical breakdown. In gases, static breakdown develops when the rate of voltage variation is significantly lower than the rate of elementary processes occurring during the electrical breakdown of the gas. The breakdown becomes dynamic in the moment these rates become comparable. Depending on the localization of the dominant secondary processes in the system, the static breakdown occurs through the Townsend mechanism or the streamer mechanism. Static breakdown voltage is a deterministic feature of the system, unlike dynamic break-

Townsend mechanism is the main pathway of direct current (DC) breakdown, if the secondary processes on the electrodes (e.g., the ionic discharge, photoemission, metastable discharge, etc.) are more prevalent than the secondary processes in the gas (the ionization by positive ions, photo ionization, metastable ionization, etc.) [3]. The value of DC breakdown voltage by Townsend mechanism is described by

where ɑ is the number of ionization collisions per cm along the electric field, ƞ is the number of electrons per cm along the field associated to electrically negative atoms or molecules, and γ is the number of electrons generated from secondary

When the secondary processes in the gas are more prevalent than the secondary processes on the electrodes, DC breakdown will occur as streamers, and the breakdown voltage will be dictated by the streamer mechanism [18, 19]. The value of a DC voltage for streamer mechanism breakdown is described by the following

To describe, determine, and compare the pulse shape characteristics of GFSA, an analytical algorithm has been developed. This approach allows a fast and accurate determination of the GFSA characteristics, without including the irreversible processes resulting from the repeated measurements of the "pulse breakdown voltage" random variable, using different shapes of rapid voltage changes. A pulse shape characteristic (volt-second) represents the breakdown voltage of the GFSA electrode configuration as a function of the applied voltage pulse duration. A decrease in the duration of the pulse results in the increase of the breakdown voltage. Determining the exact pulse characteristic experimentally would require a large number of tests with differently shaped voltage pulses. The application of the area law allows the determination of the pulse shape characteristic based on a single set of measurements using a single pulse voltage shape. The basic assumption in the area law is that plasma-spreading rate in the inter-electrode gap increases linearly

*V*(*x*,*t*) = *k* (3)

*(α−η)*dx*α*dx = 1 (1)

[*α*(*x*) − *η*(*x*)] ⋅ dx ≥ *k* (2)

gas pressure and d is the distance between the electrodes [1, 11, 16].

down that is a stochastic process occurring in a range of voltages [17].

0 d *e*∫ 0 *d*

o d

*DOI: http://dx.doi.org/10.5772/intechopen.83371*

the following equation:

equation:

γ ∫

processes per primary avalanche.

∫

due to the rise of the electric field:

**Figure 1.** *Schematic of a gas-filled surge arrester [9].*

*Influence of Gamma Radiation on Gas-Filled Surge Arresters DOI: http://dx.doi.org/10.5772/intechopen.83371*

*Use of Gamma Radiation Techniques in Peaceful Applications*

radioactive filling) [6–8].

**2. Gas-filled surge arresters (GFSA)**

voltage reaching a certain threshold. Commonly used are transient suppression diodes (TSD), metal-oxide varistors (MOV), and gas-filled surge arresters (GFSA) [4]. The advantages of GFSA compared to the other overvoltage components protection are (1) the ability to conduct high currents (up to 5000 A), (2) low intrinsic capacity (~1 pF), and (3) low costs [5]. The disadvantages of GFSA are (1) practical irreversibility of characteristics after the electric arc effect, (2) delayed response, and (3) unsuitability with respect to environmental protection (if GFSA have a

The basic design of gas-filled surge arresters is a simple configuration of two or three electrodes encased in a glass or ceramic enclosure filled with insulating gas (**Figure 1**) [9]. The common insulation medium of choice is noble gases [10]. The form of the electrodes is such that it provides a pseudo-homogeneous macro component of the electric field [11]. The strength of the electric field impacts the ratio of free electron generation and loss in the insulating gas. If the field is strong enough, a self-sustained avalanche process will lead to the electrical breakdown of the gas and discharge of the overvoltage [12, 13]. The performance of GFSA is best described by its pulse shape characteristic. The narrower the pulse characteristic is, and the smaller the slope is, the better protective characteristics of GFSA are [6, 7]. Solutions to improve the performance of GFSA in response time include the use of hollow cathodes, thus avoiding radioactive materials or implementing internal radioactive sources [14, 15]. The differences in regulations and standards and the lack of clear strategy for their clear storage and disposal make radioactive sources a challenging solution in the context of radiation protection and potential environmental contamination. Still, there are some natural (cosmic radiation, solar flares, coronal mass ejection) and man-made phenomena (nuclear explosions, especially at high altitude) where the power grid, electrical machinery, and electronic compo-

nents can experience voltage transients under ionizing radiation.

In order to describe the electrical breakdown of the insulating gas, avalanche coefficients are used to specify the elementary processes in GFSA's gas. Avalanche coefficients that are most commonly used are ɑ, the number of electronic ionization collisions per cm of a distance in the direction of the electric field; ƞ, the number of electrons per cm of a distance in the direction of the field attached to the electrically negative atoms or molecules; and γ, the number of electrons generated

**216**

**Figure 1.**

*Schematic of a gas-filled surge arrester [9].*

from secondary processes per primary avalanches. The avalanche coefficients vary with pressure and electric field and are not constant for any particular gas or a gas mixture. This dependence has been experimentally determined for most insulating gases and is described as a range of values of the *pd* product, where p is the insulator gas pressure and d is the distance between the electrodes [1, 11, 16].

The electrical breakdown in gases can be dynamic or static, not depending on the precise mechanism of the electrical breakdown. In gases, static breakdown develops when the rate of voltage variation is significantly lower than the rate of elementary processes occurring during the electrical breakdown of the gas. The breakdown becomes dynamic in the moment these rates become comparable. Depending on the localization of the dominant secondary processes in the system, the static breakdown occurs through the Townsend mechanism or the streamer mechanism. Static breakdown voltage is a deterministic feature of the system, unlike dynamic breakdown that is a stochastic process occurring in a range of voltages [17].

Townsend mechanism is the main pathway of direct current (DC) breakdown, if the secondary processes on the electrodes (e.g., the ionic discharge, photoemission, metastable discharge, etc.) are more prevalent than the secondary processes in the gas (the ionization by positive ions, photo ionization, metastable ionization, etc.) [3]. The value of DC breakdown voltage by Townsend mechanism is described by the following equation:

$$\gamma \int\_0^d e^{\int\_{(\alpha-\eta)d\mathbf{x}}^d d\mathbf{x}} d\mathbf{x} = \mathbf{1} \tag{1}$$

where ɑ is the number of ionization collisions per cm along the electric field, ƞ is the number of electrons per cm along the field associated to electrically negative atoms or molecules, and γ is the number of electrons generated from secondary processes per primary avalanche.

When the secondary processes in the gas are more prevalent than the secondary processes on the electrodes, DC breakdown will occur as streamers, and the breakdown voltage will be dictated by the streamer mechanism [18, 19]. The value of a DC voltage for streamer mechanism breakdown is described by the following equation:

$$\int\_{a}^{d} \left[ a(\mathbf{x}) - \eta(\mathbf{x}) \right] \cdot d\mathbf{x} \ge k \tag{2}$$

To describe, determine, and compare the pulse shape characteristics of GFSA, an analytical algorithm has been developed. This approach allows a fast and accurate determination of the GFSA characteristics, without including the irreversible processes resulting from the repeated measurements of the "pulse breakdown voltage" random variable, using different shapes of rapid voltage changes. A pulse shape characteristic (volt-second) represents the breakdown voltage of the GFSA electrode configuration as a function of the applied voltage pulse duration. A decrease in the duration of the pulse results in the increase of the breakdown voltage. Determining the exact pulse characteristic experimentally would require a large number of tests with differently shaped voltage pulses. The application of the area law allows the determination of the pulse shape characteristic based on a single set of measurements using a single pulse voltage shape. The basic assumption in the area law is that plasma-spreading rate in the inter-electrode gap increases linearly due to the rise of the electric field:

$$V(\mathbf{x}, t) = k \tag{3}$$

where k is the parameter that depends on electrical discharge mechanism and electrode polarization and Es is the electric field value corresponding to the DC breakdown voltage Us. The voltage pulse breakdown condition is given by u(t) > Us, given that the DC breakdown voltage Us represents the lowest possible value of the breakdown voltage for a specific electrode configuration. Without taking the charge spreading in the inter-electrode gap into account, the following stands:

$$E(\mathbf{x}, t) = \boldsymbol{\mu}(t) \cdot \mathbf{g}(\mathbf{x}) \tag{4}$$

where g(x) is a function that is to be determined from the specific electrode configuration.

If the complete breakdown occurs through the Townsend mechanism (i.e., k is assumed to be constant along the inter-electrode gap), then according to expressions (3) and (4)

$$\frac{1}{k\_{t\_1}} \stackrel{t\_1 \leftrightarrow t\_s}{\text{g}(\text{x})} = \stackrel{t\_1 \leftrightarrow t\_s}{\text{l}} \begin{bmatrix} \mu(t) \ -U\_s \end{bmatrix} \text{dt} = P = \text{const} \tag{5}$$

where x = xk indicates the zone where breakdown through Townsend mechanism changes into breakdown by a streamer, in the moment t = t1 + ta. In order for a breakdown to occur, a constant surface has to be formed in the voltage-time plane between u(t) and Us. Thus, measuring the surface P and the DC breakdown voltage Us is sufficient to determine the pulse shape and pulse breakdown voltage, given that these characteristics of the system do not depend on the applied voltage [7].

A semiempirical method to determine the pulse shape characteristic is performed in the following way: a sequence of DC breakdown voltages is measured (with at least 20 measurements), followed by a sequence of pulse breakdown voltage measurements, by applying the pulse voltage of a stable and defined shape (at least 50 measurements). The corresponding distribution function is obtained by statistical analysis of the measured values. Distribution function allows the determination of the quantities of pulse shape characteristics Ux and Uy desired boundaries (most frequently one takes x = 0.1% and y = 99.9%). When the mean value of DC breakdown voltage Us is known, and the quantities are determined, following system of equations can be solved:

$$\begin{aligned} \boldsymbol{u}\,(t) &= U\_{\text{s}}, \boldsymbol{t} = \boldsymbol{t}\_{1} \\ \boldsymbol{u}\,(t) &= U\_{\text{x}}, \boldsymbol{t} = \boldsymbol{t}\_{\text{ax}} \\ \boldsymbol{u}\,(t) &= U\_{\text{y}}, \boldsymbol{t} = \boldsymbol{t}\_{\text{ay}} \end{aligned} \tag{6}$$

Values t1, tax, and tay allow the determination of surfaces Px and Py by applying the area law:

$$\begin{aligned} P\_x &= \int\_{t\_1}^{t\_1 \ast t\_{xx}} [\mu(t) - U\_s] \mathbf{d}t = \text{const} \\ P\_y &= \int\_{t\_1}^{t\_1 \ast t\_{xy}} [\mu(t) - U\_s] \mathbf{d}t = \text{const} \end{aligned} \tag{7}$$

**219**

*Influence of Gamma Radiation on Gas-Filled Surge Arresters*

**3. Radiation resistance of gas-filled surge arresters**

**4. Induced radiation effects on GFSA characteristics**

The radiation resistance of gas-filled surge arresters (GFSA) is of great importance, especially if the devices are applied in operating regimes with constant, occasional, or potential exposure to ionizing radiation. Natural or artificial atmospheric electromagnetic pulses can cause varying levels of damage to electronic components. Thus, GFSA are very important overvoltage protection components for low-voltage applications in both the military industry and space exploration technologies. Testing of GFSA in comparison to other nonlinear surge arresters has indicated the feasibility of replacing commonly used semiconductor overvoltage components (transient suppressor diodes, metal-oxide varistor), whose protective

characteristics significantly degrade when subjected to the radiation [4, 6].

In order to test the performance of GFSA under the influence of n + γ radiation, the following variables were determined in a n + γ field: (1) the random variable "pulse breakdown voltage," (2) the random variable "DC breakdown voltage," and (3) the volt-second characteristic. The n + γ source was californium isotope 252Cf. This source was selected due to its neutron spectrum resemblance of a nuclear blast's neutron spectrum [20]. Since the nuclear cross section for capturing a neutron is large enough only for thermal and slow neutron capture and due to the structure of californium source fission spectrum, a relatively small part of neutrons takes part in the neutron activation of GFSA materials. GFSA was subjected to two

and 16.24 × 1011 n/cm2

ponent, emitted radiation also has a γ component. The latter influences the electric characteristics of GFSA only for the duration of the exposure to the radiation field. Also, the inelastic interaction cross section of the neutron component is larger than the corresponding γ component cross section [21]. This allows the observation of the effects of radiation resulting from the neutron fluency only. In the experiment, the type and pressure of gas varied in order to get a detailed insight into how

By measuring 1000 values, the influence on the "DC breakdown voltage" and "pulse breakdown voltage" random variables was tested. During the measurement series, discharge energy (current) was maintained constant. Results of the breakdown voltage obtained in the measurement series were divided into 10 groups of 50 successive values. Statistical tests were performed on each group of results by graphical visualization and chi-squared and Kolmogorov-Smirnov tests. Within each group of measurements, the measured values of breakdown voltage were tested with respect to the type of theoretical distributions (normal, exponential, double exponential, and Weibull's). U test was used to determine the groups of measurement series having the same random variable (with significance level α = 5%) [17, 22]. The area law was used to explore the effects on the pulse shape

The experiments show that the standard deviation of the static breakdown voltage significantly decreased after the irradiation of the GFSA. The pulse voltage tests show that an irradiated GFSA reacts more readily and has somewhat narrower volt-second characteristic than unirradiated GFSA. Effectively, irradiation has improved GFSA's protective traits. GFSA DC breakdown voltage versus neutron fluency is presented in **Figure 2**. The GFSA volt-second characteristics before and after exposure to the radioactive source, respectively, is presented in **Figure 3A** and **B**.

. Along with the neutron com-

*DOI: http://dx.doi.org/10.5772/intechopen.83371*

neutron fluencies: 5.41 × 109

(volt-second) characteristic.

the radiation influences the GFSA characteristics.

Upon the determination of surfaces Px and Py, it is possible to calculate the xth and yth values of the "pulse breakdown voltage" random variable (by applying the area law) for any form of pulse voltage u(t). If the form of pulse voltage is taken as a parameter (in the time interval considered), it is possible to determine xth and yth pulse shape characteristics [7].

*Use of Gamma Radiation Techniques in Peaceful Applications*

configuration.

sions (3) and (4)

\_\_1

system of equations can be solved:

pulse shape characteristics [7].

*k* ∫ *t*1 *t*1+*ta* \_\_\_\_ dx g(x) = ∫ *t*1 *t*1+*ta*

where k is the parameter that depends on electrical discharge mechanism and electrode polarization and Es is the electric field value corresponding to the DC breakdown voltage Us. The voltage pulse breakdown condition is given by u(t) > Us, given that the DC breakdown voltage Us represents the lowest possible value of the breakdown voltage for a specific electrode configuration. Without taking the charge

*E*(*x*,*t*) = *u*(*t*) ⋅ *g*(*x*) (4)

where g(x) is a function that is to be determined from the specific electrode

If the complete breakdown occurs through the Townsend mechanism (i.e., k is assumed to be constant along the inter-electrode gap), then according to expres-

where x = xk indicates the zone where breakdown through Townsend mechanism changes into breakdown by a streamer, in the moment t = t1 + ta. In order for a breakdown to occur, a constant surface has to be formed in the voltage-time plane between u(t) and Us. Thus, measuring the surface P and the DC breakdown voltage Us is sufficient to determine the pulse shape and pulse breakdown voltage, given that these characteristics of the system do not depend on the applied voltage [7]. A semiempirical method to determine the pulse shape characteristic is performed in the following way: a sequence of DC breakdown voltages is measured (with at least 20 measurements), followed by a sequence of pulse breakdown voltage measurements, by applying the pulse voltage of a stable and defined shape (at least 50 measurements). The corresponding distribution function is obtained by statistical analysis of the measured values. Distribution function allows the determination of the quantities of pulse shape characteristics Ux and Uy desired boundaries (most frequently one takes x = 0.1% and y = 99.9%). When the mean value of DC breakdown voltage Us is known, and the quantities are determined, following

> *u*(*t*) = *Us*,*t* = *t*<sup>1</sup> *<sup>u</sup>*(*t*) <sup>=</sup> *Ux*,*<sup>t</sup>* <sup>=</sup> *<sup>t</sup>*ax

*u*(*t*) = *Uy*,*t* = *t*ay

Values t1, tax, and tay allow the determination of surfaces Px and Py by applying

[*u*(*t*) − *Us*]dt = const

[*u*(*t*) − *Us*]dt = const

Upon the determination of surfaces Px and Py, it is possible to calculate the xth and yth values of the "pulse breakdown voltage" random variable (by applying the area law) for any form of pulse voltage u(t). If the form of pulse voltage is taken as a parameter (in the time interval considered), it is possible to determine xth and yth

*Px* = ∫ *t*1 *t*1+*tax*

*Py* = ∫ *t*1 *t*1+*tay* [*u*(*t*) − *Us*]dt = *P* = const (5)

(6)

(7)

spreading in the inter-electrode gap into account, the following stands:

**218**

the area law:
