Lightning Protection Improvement and Economic Evaluation of Thailand’s 24 kV Distribution Line Based on Difference in Grounding Distance of Overhead Ground Wire

This study determines the voltage across insulators after a direct lightning strike to an overhead ground wire on a 24kV pole structure for diﬀerent grounding distances of overhead ground wire, to calculate the maximum ground resistance required to avoid disruption of the distribution line system using ATP-EMTP software. The results show that when a 40 kA lightning current, the average lightning current in Thailand, strikes a 24kV pole structure, the maximum ground resistance should not exceed 4 Ω for a 40 m grounding distance of overhead ground wire, based on an existing critical insulator ﬂashover of 205kV. However, because the average ground resistance in Thailand is approximately 10 Ω , this study proposes increasing the insulation level from 205 kV to 300kV to reduce the likelihood of power outage. The cost-eﬀectiveness of such an investment is assessed in terms of net present value (NPV), internal rate of return (IRR), proﬁtability index (PI), and discounted payback period (DPP) using existing economic tools. Results show that when the critical insulator ﬂashover is increased from 205kV to 300kV for a 40 m grounding distance of overhead ground wire, the project is likely to have a DPP of 15.12 years, NPV of 143,321.87USD, IRR of 12%, and PI of 1.15. On the other hand, grounding distances greater than 40m for overhead ground wire result in negative NPV, although the back ﬂashover rate can be reduced by 1.51–5.71% with grounding distances of 80–200m compared to the situation in the absence of grounding.


Introduction
kV power distribution line networks in ailand are the most important infrastructures for sending electricity to customers. erefore, disruptions on these distribution lines will cause significant economic losses. Lightning strikes on these distribution lines are the main reason for permanent and temporary power disruptions. Results of studies performed in Asian countries such as Malaysia, Japan, and China show that lightning is the main reason for power disruption. In ailand, many disruptions and power outages occur because of lightning [1]. erefore, 24 kV distribution lines should be equipped with some form of lightning protection. Many methods to decrease the disruption of distribution lines from lightning strikes to the system have been developed, and many variations in lightning protection have been presented in literature [2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21]. For example, Yokoyama [2] indicated that there are many power failures due to lightning overvoltages that occur on the coast and in the mountainous areas of Japan. e installation of lightning arresters every 200 meters with insulation levels of 100 kV was found to greatly reduce the rate of power outages due to indirect lightning overvoltages. Additionally, it was found that the lightning protection design of a distribution system depends on the type of lightning, type of materials used, and location of the distribution system. All variables are considered to determine the appropriate lightning protection system. Zhang et al. [3] determined that ground conductivity and lightning channel position have effects on lightning-induced voltage on overhead ground wires (OHGWs). Omidiora and Lehtonen suggested that structures of medium voltages could be protected by OHGW lines from direct lightning strikes or flashovers across trees, with the OHGW lines providing effective protection for lowground-resistance systems [4]. Han et al. [5] demonstrated that the OHGW is an important equipment for lightning protection that functions at protection angles not exceeding 45°. Meanwhile, in another study, the voltage induced by a lightning current was reduced using two OHGWs below and above the phase conductor [6]. Paulino et al. [7] showed improvements in lightning protection by increasing the dielectric strength from 95 kV to 170 kV based on the variation of the ground resistivity in the simulations. Additionally, the lightning protection of 95 kV and 170 kV insulation systems was evaluated in a real environment. It was found that using an insulation rating of 170 kV could reduce the power outage rate by more than 50%. Michishita and Hongo [8] reported that the flashover rate is lower than 0.0001 times/year when indirect lightning strikes a distribution line in which lightning arresters are installed every 200 m, and they also found that the subsequent lightning strikes are key to the system design because their flashover rates in distribution systems are higher than those of the first lightning strokes; further, the use of vertical rods is recommended to improve the grounding system. Meanwhile, arc-extinguishing lightning protection gaps can be applied onto both ends of a 10 kV insulator to protect the insulation from burning [9]. Piantini [10] indicated that the flashovers of an insulator in a medium voltage system due to indirect lightning current are less frequent if the critical flashover (CFO) of the insulator is more than 300 kV and also found that the lightning protection level will decrease when the grounding distance of OHGW is wider and the soil resistance is higher. If the system is not equipped with OHGW, the lightning arrester may be damaged due to excessive energy power. erefore, the best protection is to install OHGW with lightning arresters at every pole and every phase; however, the cost-effectiveness must be considered. Lowvoltage systems are more sensitive than distribution systems because of their lower insulation level. e research shows that more transformers are damaged if the lightning arrester is attached to the high side; therefore, the lightning arrester on the low side can greatly reduce the damage of the transformer.
To protect against lightning overvoltages in combined underground cables and overhead lines, installation of surge arresters is required on both ends of the underground cables [11]. In places with low ground conductivities and high insulation levels of 300 kV or high ground conductivities and low insulation levels of 150 kV, the distance between arresters should be less than 400 m [12]. According to Chen et al. [13], lightning arresters are the best equipment for protecting a system from lightning. e spacing of lightning arresters is a major consideration in system protection. However, surge arresters have been shown to be unable to suppress overvoltages in cases of direct lightning strikes to the system, and OHGWs were demonstrated to be better equipment for system protection in severe-lightning areas [14].
Many factors must be considered in the provision of appropriate lightning protection, such as ground resistance, lightning current waveform, lightning current magnitude, type of lightning current, type of pole structure, and lightning current position. Meanwhile, Paulino et al. [15] showed that lightning current waveforms are affected by ground resistance. ey also determined that the back flashover rate can be decreased via maintenance of the ground resistance to a minimum [16]. Meanwhile, Tossani et al. [17] reported that buildings near overhead lines in a city are expected to decrease electromagnetic lightning pulses radiated by indirect lightning strikes. Mikropoulos and Tsovilis [18] presented a statistical estimation of the surface flash rate in a distribution system in accordance with IEEE Std. 1410:2011. e conditions for a lower flashover in a distribution line system are as follows: a lower height of the distribution line structure, lower ground resistance and lightning current magnitude, and nearby objects close to the distribution system. is paper presents a study of the effects of direct lightning to an OHGW on a 24 kV distribution system for various grounding distances of OHGW.
is study considers only direct lightning strikes to the OHGW. e influence of indirect lightning strikes (lightning induced voltages) on the number of lightning outages per 100 km is not considered. e installation of an OHGW is expected to reduce the number of these outages [2]. is would affect the economic analysis presented. In this study, the flashover of insulators was obtained by simply comparing the overvoltages with the CFO of the insulators. More complex and accurate methods are available in the literature [19][20][21]. e investigation was performed using ATP-EMTP software. In this study, five different values of grounding distance of OHGW are tested with an existing insulator of 205 kV, to determine the permissible ground resistance and insulation that prevent power outages from occurring in the system. In addition, this study is based on the 40 kA average lightning current in ailand striking a 24 kV distribution system with an average ground resistance of 10 Ω. For this setup, the insulation level should be increased from 205 kV to 300 kV for the system to be able to withstand a 40 kA lightning current. Economic evaluations of the different grounding distances of OHGW are shown in Section 4. e paper is arranged as follows: in Section 1, the introduction is presented; in Section 2, the configuration of the system and relevant data on lightning are illustrated; in Section 3, a case study is demonstrated; and, in Section 4, the cost-effectiveness of the investment is explained. Finally, in Section 5, the conclusion is presented.  [22]. is number of thunderstorm days is used to calculate the ground flash density (GFD) using the following equation [23]:

System Configuration and Relevant Data
where N g is ground flash density (flashes•km 2 /year) and T d is the number of thunderstorms (days/year). e probability of the peak lightning current exceeding the critical lightning-current magnitude can be calculated using the following equation [24]: where P(I) is probability of peak lightning current exceeding critical lightning-current magnitude (%). I is critical lightning-current magnitude (kA) from the simulation in ATP-EMTP program. M is median of stroke peak current magnitude (kA), 40 kA in ailand [23]. B is constant (3.09 for ailand power system).
A plot of equation (2) is shown in Figure 1, where the median is 40 kA as of year 1997 [23]. e back flashover rate (BFR) [25] is calculated using equation (3), where N L is calculated using equation (4).
where N g is ground flash density (flashes/km 2/year), [26], h is average conductor height (m), and b is separation distance of OHGW (m). OHGW in this paper is used for single overhead ground wire; therefore, b � 0.

Model of Lightning Current.
Negative lightning current is supposed to strike the OHGW. e lightning-current waveform follows a CIGRE concave shape, as shown in Figure 2 [27,28]. e impedance of the lightning strike channel was 400 Ω parallel to the current source. e 400 Ω value of the lightning strike channel is a typical value used in literature; however, it is more appropriate for higher lightning currents as analyzed in [29]. Using a higher value, such as the 1000 Ω value proposed in [30], would considerably influence the obtained results.
Normally, the time to halve the value, or the tail time, is a stable 77.5 μs, whereas the front time, t f (μs), and maximum steepness, S m (kA/μs), were considered as functions of the lightning current crest, If (kA), where t f and S m are calculated using the two following equations [27,28]: where I is lightning current crest, If (kA). e parameters of the lightning-current source examined in this study are outlined in Table 1.

Economic Evaluation.
e economics study is performed based on calculations of the time value of the project in cash using discounting cash inflow, in terms of the following: net present value (NPV), discounted payback period (DPP), profitability index (PI), and internal rate of return (IRR) [31].

Net Present Value (NPV)
. NPV is the difference between the present values of the cash inflows and cash outflows over a period of time. NPV is applied to analyze the profitability of a projected investment. If the NPV is positive, the project is determined to be a potentially profitable investment; otherwise, the project should not proceed. NPV can be estimated using where A n represents cost savings per nth year, n is lifetime of project: 25 years, I 0 is investment cost at start point, and i is discount rate: 10% (mean of inflation rate of business loans in ailand).

Discounted Payback Period (DPP).
e DPP is used to decide the profitability of project based on the number of years required to recoup expenses from the initial investment (I 0 ). DPP can be calculated as follows: where A represents total n year cost savings, I 0 is investment cost at start point, and i is discount rate: 10% (mean of inflation rate of business loans in ailand).

Profitability Index (PI).
e PI is a calculation of the potential profit of the proposed project. A PI or ratio greater than 1 indicates that the project is a potentially profitable investment, whereas a PI lower than 1 indicates that the project should be avoided.
e PI can be estimated as follows:

Internal Rate of Return (IRR).
e IRR is used to estimate the profitability of a potential investment. e IRR is equivalent to the discount rate at which the NPV is equal to zero. e e IRR value can be calculated from NPV equation in which the NPV value is equal to zero as follows:

Modeling of 24 kV Distribution Line System.
ATP-EMTP was used to analyze a direct lightning strike to an OHGW on 24 kV distribution lines, as shown in Figure 3(a). e configuration and grounding system is shown in Figure 3(b) [32]. e structures are each composed of a 185 mm 2 three-phase all-aluminum space aerial conductor, pin post insulator used to support the three-phase conductor, 7.93-mm-diameter uniform zinc-coated steel wire as the OHGW, external ground wire connected to a 2.4 m ground rod at ground level, and 10.25 m concrete pole above ground as fixture for all equipment. e distances between conductors, phases, and the OHGW are shown in Figure 3(c). Distribution lines are represented by JMARTI frequency-dependent models in the ATP-EMTP program with a frequency dependency of 10 Hz-3 MHz. e interval time step is less than 1/10f max . Normally, the time step value is 1 ns.
In this project, the CIGRE type 15 model was applied, but an insulator was used instead of a capacitor. e parameters for the design are listed in Table 2.

Effects of Lightning Current on Parts of 24 kV Pole
Structure.
e effects of lightning current on the voltage across insulators for different grounding distances of OHGW (40, 80, 120, 160, and 200 m) are investigated using ATP-EMTP software, to determine the permissible ground   Mathematical Problems in Engineering 5 resistance and maximum voltage across the insulators that will not cause flashovers, as shown in Figures 4-14.
For the setup shown in Figure 4, wherein the structure is grounded at every pole or every 40 m, a direct lightning strike to the structure is inferred to cause a voltage drop across the insulator up to a certain value; if this voltage is greater than the CFO of insulation, a power outage will occur. e voltage drop across the insulator is proportional to the current size and ground resistance. For example, if the ground resistance is 10 Ω, the voltage drops across the insulator will be 222 kV and 310 kV at lightning currents of 30 kA and 50 kA, respectively. On the other hand, if the lightning current is 40 kA, the voltages across the insulators will be 268 kV and 350 kV at ground resistances of 10 Ω and 20 Ω, respectively. For a lightning current of 40 kA (i.e., average lightning current in ailand) and a ground resistance of 10 Ω (metropolis area), the voltage drop across the insulation is 268 kV, which is greater than the CFO of insulation of 205 kV (existing structure), resulting in a subsequent power outage. However, if we can improve the ground resistance such that it does not exceed 4 Ω, the likelihood of power outage will reduce because the voltage drop across the insulator will not exceed 205 kV.
Based on the possibility of direct lightning strikes to every other pole or every 80 m, the scenario is divided into two cases. In the first case, lightning strikes the grounded pole as shown in Figure 5. e result shows that voltage across the insulator will increase consistently with increases in either lightning current or ground resistance, which is similar to the results shown in Figure 4. Moreover, for the average lightning current of 40 kA and a ground resistance of 10 Ω, lightning strikes to the grounded pole will result in a 324 kV voltage across the insulation, higher than the critical insulator flashover of 205 kV, resulting in a system power outage. However, if the ground resistance is reduced to less than 2 Ω, the distribution line system would be able to withstand the lightning strikes. In the second case, lightning directly strikes an ungrounded pole, as shown in Figure 6. For the average lightning current of 40 kA and a ground resistance of 10 Ω, the resulting voltage across the insulation will be 793 kV, which is much higher than that in the first case because of the higher surge impedance and longer lightning-current path than those in the first case. On the other hand, reducing the ground resistance to its minimum will not improve the power system because of the lack of reflected waves from the base of the pole for reducing overvoltage at the top of the pole.
According to a simulation wherein lightning strikes directly at a structure grounded every other two poles or every 120 m, the voltage across the insulator of the grounded pole is lower than that of the ungrounded pole, which are similar to the results shown in Figures 5 and 6. For the average lightning current of 40 kA and a ground resistance of 10 Ω, the voltages across the insulation will be 358 kV and 986 kV for the grounded and ungrounded poles, respectively, shown in Figures 7 and 8. Both voltages are higher than the critical insulator flashover of 205 kV, which will also result in power outages. e permissible ground resistance does not exceed 1 Ω for when lightning strikes the grounded pole, whereas there is no permissible ground resistance for when lightning strikes an ungrounded pole.
When lightning strikes a pole structure grounded every other three poles or every 160 m, for the average lightning current of 40 kA and a ground resistance of 10 Ω, the voltage           whereas there is no permissible ground resistance for when lightning strikes an ungrounded pole. In a scenario wherein the structure is grounded every other four poles or every 200 m, as shown in Figures 12-14, the results are similar to those shown in Figures 9-11, that is, low voltage across the insulator when lightning strikes a grounded pole and permissible ground resistance not exceeding 1 Ω for an average lightning current of 40 kA. e observations visualized in Figures 4-14 reveal that the voltage across the insulator changes for different grounding distances of OHGW, different ground resistances, and different lightning magnitudes. e reason for the variations is that the pole top voltage can be attenuated by reflected waves generated by nearby poles and the ground resistance of the ground rod, indicating that system reliabilities with respect to the grounding distance of OHGW and in terms of ground resistance are quite different. e main results are outlined in Tables 3 and 4.  Table 3 shows the permissible ground resistance (Ω) in the event of a lightning strike on a 30-50 kA distribution line system that would allow the insulator to withstand the strike without flashover. e permissible ground resistance decreases either when the lightning current increases or when the grounding distance of OHGW is widened. For example, when the lightning current increases from 30 kA to 50 kA, the permissible resistance is reduced from 8 Ω to 2 Ω at a 40 m grounding distance of OHGW. On the other hand, when the grounding distance of OHGW is increased from 40 m to 120 m, the permissible resistance is reduced from 4 Ω to 1 Ω for a lightning current of 40 kA. e results show that when lightning strikes a structure characterized by a longer grounding distance of OHGW, the acceptable ground resistance is reduced. For example, for a lightning current of 40 kA, a structure grounded at distances greater than 120 m would have a permissible ground resistance not exceeding 1 Ω. However, in practice, it will be difficult for the ground resistance to attain a value as low as 1 Ω. erefore, other improvements to lightning protection, such as increasing the dielectric level of the insulator, should be considered.       In Table 4, the minimum critical impulse flashover (kV) at 10 Ω ground resistance allows the insulator to withstand lightning strikes without flashover. e minimum critical impulse flashover increases either when the lightning current increases or when the grounding distance of OHGW is widened. For example, when the lightning current increases from 30 kA to 50 kA, the minimum critical impulse flashover is increased from 222 kV to 310 kV. On the other hand, when the grounding distance of OHGW is increased from 40 m to 200 m, the minimum critical impulse flashover is increased from 268 kV to 360 kV when the lightning current is 40 kA.

Increase in CFO of Existing Insulator
According to Table 4, the CFO of the insulator should be 268-360 kV to reduce the likelihood of power outage for an average lightning current of 40 kA (as in ailand) and an average ground resistance of 10 Ω (as in a metropolis setting) in case direct lightning strikes the OHGW. e influence of indirect lightning strikes (lightning induced voltages) on the number of lightning outages is not considered. Moreover, the grounding distance of OHGW has an effect on suitable CFO; wider grounding distances of OHGW require corresponding increases in CFO. erefore, the objective of this section is to increase the CFO of an existing insulator from 205 kV to 300 kV for a variety of grounding distances of OHGW. Afterward, the cost-effectiveness is evaluated according to economic principles to determine which projects are probable good investments.
First, the critical lightning current, which causes the voltage across the insulator to be equal to a CFO voltage of 300 kV, is determined. is step is performed manually via increases in current in the ATP-EMTP software. Second, the critical current is used to calculate the probability of peak lightning current exceeding the critical lightning-current magnitude (%) from equation (2). ird, the back flashover rate is calculated using equation (3) in Microsoft Excel software.
Fourth, the BFR is multiplied with the outage cost in column (e) to determine the damage cost in column (f ) and reduction or savings cost in column (g) of each grounding structure. e project periods and discount rates used for economic evaluation are 25 years and 10%, respectively.
In this study, increasing the CFO from 205 kV to 300 kV and varying the grounding distance of OHGW from 40 m to 200 m are both analyzed in order to find cost reductions. e results are outlined in Table 5.
Finally, the cost-effectiveness of the investment is determined using economic tools, in terms of measures such as DPP, NPV, IRR, and PI. e project periods and discount rates used for economic evaluation are 25 years and 10%, respectively. e total breakdown of investment cost is equal value of 302.27 USD/pole, comprising materials and services used on the pole, which include the new insulator, zinc-coated steel wire for OHGW and external ground wire, and labor cost and work control, as shown in Table 6. e total investment cost for 100 km distribution lines can be calculated using the data in Table 7.   Lightning current (kA) Grounding distance of OHGW (m)  40  80  120  160  200  30  222  267  292  292  300  40  268  324  358  359  360  50  310  377  416 418 420    e annual costs of operation and maintenance (O&M) for both increase in CFO from 205 kV to 300 kV and installation of grounding distance of OHGW, which is estimated to be at 3% of the total investment cost, are shown in Table 8, column (i).
With regard to the economic results, summation of the total investment cost and annual cost of O&M according to Table 8 in columns (h) and (i), compared with the damage reduction or cost savings in Table 5, column (g), demonstrates that purchases of new or replacement insulators for 40 m grounding distance of OHGW should be considered important investments that require high installation costs but result in greater reductions with better economic results, that is, DPP of 15.12 years, NPV of 143,321.87 USD, greater IRR of 12%, and superior PI of 1.15.

Conclusions
is study investigated the effects of direct lightning strikes to an OHGW on a 24 kV distribution system in ailand. e statistics of power outages have identified large numbers of unknown power outages occurring in urban areas of ailand. erefore, a method for improving lightning protection, via installation of OHGW at various grounding distances, was proposed, to distribute lightning current to the ground level and to account for increases in the critical insulator flashover from 205 kV to 300 kV. e purpose of the study was to reduce power outages among distribution systems. ATP-EMTP software was used for analysis. e study determined that increasing the insulation level via installation of OHGW at every pole or every 40 m can greatly reduce the back flashover rate, thereby causing reductions in power outages by 27.58% compared to when the system is unprotected by OHGW. Although calculations showed that increasing insulation levels to 300 kV with installation of OHGW at every pole has the highest installation costs, the payback period will be fast, at 15.12 years. On the other hand, other grounding distances or wider grounding of OHGW would at least be able to help reduce back flashover rate by 1.51%-5.71% for grounding distances of 80-200 m. If the ground resistance is significantly distinct from 10 Ω, the results of cost analysis of the investment may change, that is, lower ground resistance causes a lower back flashover rate. erefore, a grounding system is imperative for the design of lightning protection in distribution systems, which should be a first priority. Furthermore, another method should be considered if the grounding system is not enough for lightning protection, such as increasing the insulation level of insulator or using a lightning arrester to reduce overvoltage, to enhance the lightning protection performance. All results of this study could serve as a guideline for the design of distribution lines and improvement of power systems.

Data Availability
No data were used to support this study.

Conflicts of Interest
e authors declare that they have no conflicts of interest.