1 Thermal modeling of an Aero Z Linnet overhead conductor considering the

2 effects of high-emissivity coating ink on ampacity

3
4
5 Lucas Lanza Bernardes1, Lucas Carmona1, Gabriel Fonseca1, Fernanda Souza da Silva1, Carlos
6 Alexandre Meireles do Nascimento2, Sidnei Ueda3, Rudolf Huebner1
7
8 1- Group of Analysis and Modeling of Thermal Systems and Equipment, Department of Mechanical
9 Engineering, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil.
10 2- Companhia Energética de Minas Gerais – Distribuição.
11 3- Nexans do Brasil.
12
13
14 Abbreviated title: CFD thermal modeling of an Aero Z Linnet conductor
15
16
17 Corresponding Author: Lucas Lanza Bernardes
18
19 Group of Analysis and Modeling of Thermal Systems and Equipment
20 Department of Mechanical Engineering
21 Universidade Federal de Minas Gerais
22 Av. Presidente Antônio Carlos, 6627
23 31270-010 Belo Horizonte - MG – Brazil
24 Phone: +55 31 32913812
25 E-mail: [email protected]
26

Highlights

 The thermal state of the Aero Z Linnet conductor was studied using CFD modeling.

 Steady-state was studied for a range of operating conditions.

 CFD results were compared to those from industrial standards and experiments.

 High-emissivity coating ink was used to improve ampacity.

27
28
29
30
31 Conflict of interest:

32 We certify that there are no conflicts of interest associated with the work described in the article.
 Abstract
33 This work proposes the development of a numerical model that effectively represents the heat

34 transfer phenomena on transmission line cables. It was validated by comparison with results from

35 tests in a wind tunnel and analytical calculations proposed by technical standards in the electrical

36 engineering area for three different wind speeds (0, 1 and 3 m/s) and two different electric currents

37 (200 A and 400 A). It considered the effect of the use of high-emissivity coating ink as a tool for

38 ampacity increase. The study made possible a decrease in temperature of up to 13 % based on the

39 use of the coating ink. CFD numerical results were satisfactory, presenting good agreement with

40 experimental and analytical data.

41

42 Keywords: Ampacity, transmission line cables, coating, numerical simulation.

43

44 1. Introduction

45 The power plants are typically built away from consumer centers (cities and

46 industries). For this reason, the electricity produced by the generators runs long distances in

47 a complex energy transmission system, increasing the system power losses .

48 Associated with this, there is a high difficulty of expansion into new spaces for new

49 transmission lines. Society currently imposes restrictions on the expansion of airlines due to

50 the need for environmental preservation and visual pollution. In addition, the cost of

51 deploying new lines is very high [1].

52 In this context, this study aims to present ways to improve the thermal performance

53 of conductor cables used in overhead transmission lines. It is expected to contribute to the

54 security of the operation and have the real possibility of raising the transmission capacity of

55 the overhead lines (increase the ampacity). It made use of auxiliary materials, such as high-

56 emissivity coating. With this, it aims to get around, where possible, the difficulty of
57 installing new transmission lines due to the mentioned reasons. A numerical approach is

58 used to predict the aerial conductor steady state temperature for twelve different

59 combinations of wind speed, electric current and cable superficial emissivity.

60

61 1.1 Ampacity

62 The ampacity is the technical term representing the electrical power transmission

63 capacity of a transmission line, considering the heat limitation permitted for conductor

64 cables, either for technical specifications or security factors [2]. The weather conditions in

65 the vicinity of transmission lines, in combination with the electric current, determine the

66 temperature distribution in the conductors. Wind speed and direction, ambient temperature,

67 solar radiation, the presence of rain or snow are the prevailing climatic factors in

68 determining the line ampacity, among others of less influence [3].

69 The main technical standards related to the ampacity are Morgan [4], IEEE [5] and

70 Cigrè [6]. The most used standards are the last two, which are actually based on several

71 studies on thermal ratings, including mainly the Morgan study and also ECAR [7]. What

72 differentiates the approach taken by the technical standards of the area, companies and

73 industry experts are the calculations used in determining the gain rates and loss of thermal

74 energy as well as the simplifications, e.g., neglecting the contribution of corona effects,

75 magnetic heating and evaporation. Moreover, different calculation methods are presented for

76 conductor properties, such as electrical resistance. Each standard uses specific correlations

77 of convection (natural, forced and mixed) and solar gain. The other terms of gain and heat

78 loss are calculated in the same way as in the standards.

79

80 1.2 State-of-the-art
81 Makhkamova [8] applied the finite volume technique using commercial software in

82 the study of wind flows over aerial cables and underground power lines. The numerical

83 results were consistent with the analytical and experimental data, being the differences not

84 more than 5 K. But there are limitations on the proposed model as some important

85 simplifications were made. For aerial case, the "Lynx" cable was modeled in a simplified

86 way as massive concentric aluminum and steel cylinders. It is known that normal aerial

87 cables are composed by circular wires stranded normally in opposite directions and so they

88 are not massive. Atmospheric air lies within the wires in the cable and because of that the

89 effective thermal conductivity of the cable is not equivalent to the aluminum thermal

90 conductivity. According to Morgan [9], normal effective thermal conductivity values are just

91 around 2 % of the normal aluminum thermal conductivity. Moreover, the turbulence was

92 modeled using the k-ε model which is not the most appropriate to predict flow separation

93 when analyzing flows around circular objects [10]. The flow separation can have an

94 influence in the heat dissipation. The losses by radiation are included as a constant flow

95 from the surface of the outer cylinder and not determined by the finite volume model. In this

96 way the directional attribute of radiation is neglected. The solar gain was patterned similarly.

97 Joule effect was transformed into volumetric heat generation and was not considered during

98 the simulation the electrical current distribution inside the aluminum layers.

99 Portela [11] did a similar study to the one proposed in this paper. The analyzed cable

100 was the “Linnet” type in a not compact pattern. The work development was also done from

101 numerical and analytical technique, and the results compared with data obtained from wind

102 tunnel tests. It evaluated the influence of a high emissivity coating ink on this type of cable

103 as well as the inclusion of an internal air gap between aluminum and steel layers. In this

104 case, the external diameter of the conductor was increased (1 mm) to include extra air space.

105 The cross section area of steel and aluminum was kept constant. But with this there was an

106 increase in drag force (up to 16%) as its frontal area also increased. So the proposed
107 geometries were no longer comparable to the standard Linnet cable. The use of coating ink

108 allowed an electric current gain of up to 13% in 400 A electric current and without cross-

109 flow wind case. The combined use of coating ink and air gap showed a gain of 14 % for this

110 same studied case. The used equivalent thermal conductivity was the one that got minor

111 errors between simulation and experimental results, and not based on real material properties

112 or on previous works. The influence of air gaps and thermal contact resistance between

113 wires in the equivalent thermal conductivity of aluminum was not mathematically evaluated.

114 This study aims to propose a new numerical model that is more accurate with reality

115 and that does not adopt the described simplifications made by the mentioned recent studies

116 in the field.

117

118 2. Materials and methods

119 2.1 Experimental setup

120 For the experimental analysis of this work the new wind tunnel of the Federal University of

121 Minas Gerais (UFMG) was used, which is the second largest Brazil subsonic wind tunnel. It can get

122 in tests section wind speeds of up to 400 km/h. The tunnel has an engine of 485 hp and overall

123 dimensions of 25 meters long, 15 meters wide, 6 meters high and weighs 15 tons. In Fig. 1 it is

124 possible to see the wind tunnel modeled in 3D software.

125
126
127 Fig. 1. UFMG Wind tunnel modeled in 3D.

128

129 Measurements were made according to the 12 cases described below, being the

130 conditions which were addressed in this study:

131

132 Table 1

133 Studied cases.

CASE Electrical Current Wind speed Emissivity

(A)
(m/s)

1 400 3 0.23

2 400 3 0.91

3 200 3 0.23

4 200 3 0.91

5 400 1 0.23

6 400 1 0.91
 7 200 1 0.23

8 200 1 0.91

9 400 0 0.23

10 400 0 0.91

11 200 0 0.23

12 200 0 0.91

134

135 Temperature, emissivity, electrical current and wind speed were monitored using:

136 thermocouples at four separate temperature measuring points; a thermographic camera for

137 the emissivity definition; a voltmeter to capture the potential difference at the end of the

138 cable; an anemometer to control speed inside the wind tunnel; a current source for

139 controlling the intensity of DC current. Half of the cable was covered with a coating ink,

140 which has a high emissivity value, and the other half remained without coating.

141 The Aero Z Linnet cable surface emissivity (with and without coating) was obtained

142 with the use of thermal imaging during tests in the wind tunnel. The naked cable has

143 emissivity of 0.23, and the cable with high emissivity paint has an emissivity of 0.91, close

144 to the value of 0.95, indicated by the manufacturer of the paint.

145

146 2.2 Numerical Model

147 As described by Portela [11], the creation of a 3D model of the actual cable, considering at

148 least one stranding step, implies problems for the subsequent steps. As the actual cable has small

149 interstices of air, as well as a number of complex contact regions between the wires, a very large

150 mesh refinement is needed to properly capture the physical effects present in these regions. Thus,

151 the number of elements and nodes generated is very large, which prevents such simulations with the
152 current resource used. Therefore, a simplified analysis was proposed, considering the cable as a

153 solid cylinder. The thermal conductivity was adjusted to account for the presence of internal air.

154 The difference in results between the two models was no more than 2 %, and thus only the

155 simplified model was used.

156 In this study, the object of interest is a compact cable that practically has no great presence

157 of air between the aluminum wires. Thus, the Aero Z Linnet is much closer to a massive cylinder

158 than the standard Linnet cable. This indicates that the simplified model is more appropriate for this

159 study.

160 Geometric models of the Aero Z Linnet conductor (Fig. 2) have been developed in

161 Solidworks® software based on available cable samples. This conductor consists of two layers of

162 aluminum wires stranded concentrically with alternate left-hand and right-hand directions, an air

163 layer inside and a stranded steel core, represented by regions 1, 2, 3 and 4, respectively. For this

164 study, a reduced section of the conductor was considered to analyze mainly two-dimensional

165 characteristics. Table 2 shows the geometric parameters used in this study.

166

167

168 Fig. 2. Cable geometry.

169

170 Table 2

171 Geometry Parameters.

 Dimensions Values
D1 18.9 (mm)
D2 13.2 (mm)
D3 8.5 (mm)
D4 2.7 (mm)
L 15.0 (mm)
172

173 The thermal contact resistance between the two aluminum thermal layers (0.0005 W/m 2K)

174 was calculated using the method of contact pressure determination [12] and plastic correlation given

175 by [13]:

( )
k sms P 0.95
(1)
h c=1.25
σs H v

176 Where:

ks effective thermal conductivity of the contact materials (W/mºC)

ms Inclination of asperities (m)

σs Roughness RMS (m)

P Contact pressure (MPa)

Hv Surface microhardness (MPa)

177

178 For the simplified model, an interface between the two layers of aluminum was included, so

179 that it allows inclusion of thermal contact resistance. The contact thermal resistance between wires

180 in the same layer is neglected since the conduction of heat occurs predominantly in the radial

181 direction, rather than circumferentially.

182 To determine the thermal contact resistance were considered Aero Z Linnet cable installation

183 conditions in line LD Itabirito - Ouro Preto 2, 138kV, provided by CEMIG Company:
184  Span length: 450 m.

185  Stretching condition of the cable conductor for typical span: 1240 kgf and 18,5 % of the

186 breaking load at 20 °C windless - final condition.

187 The data from Table 3 were used in the calculation, these being provided by the

188 manufacturer Nexans:

189

190 Table 3

191 Conductor properties used to calculate the thermal contact resistance.

Thermal Mean surface Poisson Superficial

conductivity of roughness coefficient hardness (MPa)
conductor (microns)
(W/mK)

203 4 0,35 490

192

193 From the simplified cable model, a region of influence and interest was created with

194 approximately ten times the cable diameter after this section and three diameters before the section.

195 Ansys Meshing from Ansys Workbench® was used for mesh generation on the model. The

196 length of mesh element criteria [14] was used to refine the initial mesh. A coarse mesh was initially

197 generated, and each simulation and refinement was performed with the intention of obtaining the

198 most suitable mesh possible. That is, a mesh that results in relation to the coarser mesh previously

199 simulated does not vary significantly. Thus, it is guaranteed that the numerical results found no

200 longer depend on the mesh used without requiring a greatly exaggerated refinement and an

201 unnecessary computational effort. The criterion mesh element length can be determined by:
 (2)
[ ]
N 1
1
∑∆Vi
3
h=
N i=1

202 Where:

∆ V i Volume of each element i in the mesh (m³)

N Total number of mesh elements (dimensionless)

h Characteristic length of mesh element (m)

203

204 In determining h, it is possible to set the mesh refinement factor r determined by Eq. (3),

205 which is the ratio between h of a coarser mesh and h of the refined mesh.

.
h cm
r= (3)
hrm

206

207 According to Celik et al. [14], refinement factor should be greater than 1.3. Moreover,

208 refinement factor should be uniform in all the studied meshes.

209 Three meshes were simulated, and a grid independence test is presented in Fig. 3. Mesh 2

210 was chosen for its low computational cost and high accuracy.

211

212

213 Fig. 3. Mesh used in the simulation and mesh test.

214

215 Table 4

216 Tested meshes.

Mesh h (mm) Elements / Knots r factor y+

1 0.085 9450967 / 2986757 1.33 0.32

2 0.064 12569787 / 3977592 1.33 0.19

3 0.048 16739719 / 5297128 1.33 0.18

217

218 For the physical modeling of the problem, CFX-Pre was used, which is also part of the

219 Ansys Workbench® program, where all the relevant materials and information of the boundary

220 conditions were inserted.

221 In the aluminum volume, a zero-volt tension was inserted on one side of the cable cross-

222 section tension, and a voltage value that induced the software to determine a consistent current

223 value of the values studied in this paper (200 and 400 A) was inserted in the other side. Only small

224 differences were allowed (1 %).

225 The Monte Carlo method is used in this study for radiation modeling. Aluminum and steel

226 are considered opaque media, i.e., radiative phenomena in these areas are surface phenomena. The

227 aluminum surface emissivity is determined in experiments for cases with (0.91) and without coating

228 ink (0.23).

229 Regarding turbulence modeling, the SST model (Shear-Stress-Transport) is recommended

230 for simulations that require higher accuracy for boundary layer behavior [15]. Another advantage of

231 this formulation is the treatment of regions near the walls in case of a low Reynolds number, which

232 is more accurate and more robust [16]. The results in terms of heat transfer are also better [17]. It is

233 therefore an appropriate model for the work treated in this case, which simulates the airflow at low

234 Reynolds numbers on cylindrical objects (the cable).

235

236 3. Results and Discussion

237 3.1 Steady-state analytical solutions

238 Using a program developed for analytical calculations according to the technical standards,

239 it was possible to obtain analytical temperature results at steady-state for the twelve cases studied.

240 IEEE, CIGRE and MORGAN, the three reference standards in the area, were considered.

241 Importantly, the solar gain calculation functions have been disabled because all study was

242 done considering the cable inside the wind tunnel where there is no direct sunlight.

243 Table 4 shows the analytical results obtained for the studied cases:

244

245 Table 5

246 Studied cases and analytical temperature values under steady-state conditions.

Case Emissivity Current Wind speed Temperature Temperatur Temperature

(A) (m/s) IEEE e CIGRÈ MORGAN
(oC) (oC) (oC)

1 0.23 400 3 33 33 33

2 0.91 400 3 33 32 32

3 0.23 200 3 27 27 27

4 0.91 200 3 27 27 27

5 0.23 400 1 40 40 40

6 0.91 400 1 38 38 38

7 0.23 200 1 29 29 29

8 0.91 200 1 28 28 28
 9 0.23 400 0 68 69 69

10 0.91 400 0 55 55 55

11 0.23 200 0 37 37 38

12 0.91 200 0 33 33 33

247

248 As noted in Table 4, all standards corroborated each other with respect to the values found. This

249 indicates that, for the same conditions studied, the main source of difference in results between the

250 standards lies in the calculation of the solar heat gain. The standards’ correlations of natural, forced

251 convection and heat loss by radiation result in very similar values for the evaluated conditions (up

252 to 3 m/s and 400 A).

253

254 3.2 Steady-state numerical solutions

255 Table 6 shows a summary of the results (analytical, experimental and numerical) and

256 deviations obtained.

257

258 Table 6

259 Comparison of numerical, analytical and experimental results.

Case Numerical Analytical Experimental Difference Difference Difference

results (N) results (A) results (E)
A and N A and E N and E
(oC) (oC) (oC)
(%) (%) (%)

1 34 33 36 2.9 8.0 5.6

2 33 33 35 0.0 7.3 5.7

3 27 27 29 0.0 7.0 6.9

 4 27 27 28 0.0 4.3 3.6

5 41 40 45 2.4 10.9 8.3

6 39 38 43 2.6 12.1 9.1

7 30 29 31 3.3 7.8 3.6

8 29 28 30 3.4 6.4 4.8

9 63 68 69 -7.9 0.4 8.7

10 54 55 60 -1.9 8.5 10.0

11 36 37 39 -2.8 3.8 7.7

12 33 33 35 0.0 4.9 5.7

260

261 In general, the analytical and numerical results were below the experimental results,

262 showing differences smaller than 12 %. The numerical and analytical results were very close, with

263 an average difference of only 2.3 %. The mean percent difference of all analytical results was 7 %.

264 The maximum deviation value was 12.1 %, and the minimal deviation was 0.5 %. This shows that

265 the experimental data are consistent with the values found by the standards, which strengthens the

266 validity of the analytical method and the experimental data for the development of this work. The

267 largest error was found in the case of 1 m/s, and worst was in the cases with the 400-ampere

268 current. A larger error was expected for the no-wind cases (0 m/s) because it is known that the

269 calculations involving natural convection are inherently more complex and difficult to be modeled.

270 However, differences of up to 5 °C were found in cases of 1 m/s, corroborating the findings of

271 Portela [11]. Thus, the results indicate the direction of worse analytical results for low wind speed

272 and Reynolds number.

273 The average deviation of all numerical simulations was 6.6 %. The maximum deviation

274 amount was 10 %, and minimum deviation was 3.6 %. This shows that the experimental data are

275 consistent with the values obtained from the numerical method, which strengthens the validity of
276 this and the experimental data for the development of this work. It notes that the errors found go

277 beyond the measurement uncertainty of the thermocouple, and thus the analytical results are not

278 within the range of possible values for the temperature values measured by thermocouples. This is

279 clear, considering that the thermocouples were inserted properly and correctly set during

280 measurements, eliminating the intrinsic measurement errors. According to Portela [11], the

281 expanded uncertainty of the temperature measurement with the thermographic camera, for this

282 experiment, is approximately 1 °C. According to Teixeira [18], the uncertainty of measurements

283 with a thermocouple in electrical connectors (aluminum) of aerial conductors is 1 % for the entire

284 measurement range (up to approximately 100 °C). Thus, due to the similarity of the surface and

285 measured material and the temperature range, it is possible to assume the same uncertainty in the

286 measurements made by thermocouples during the wind tunnel test.

287 As expected, the calculated drag coefficient by Ansys® resulted in 0.92 compared to the

288 approximation of 0.90 specified by the manufacturer for this level of Reynolds number (maximum

289 of 3.330 for the 3 m/s case).

290

291 3.3 Validation of numerical simulation

292 To validate the numerical calculation, the convective and radiative coefficients should also

293 be compared. Fig. 4 and 5 show the heat transfer coefficient by convection and the radiative heat

294 transfer coefficients. In this case, the numerical results were compared with the analytical results set

295 out in the standards (the average values of the standards were used).

296
 48.0
44.0
40.0
36.0
32.0
28.0
hc (W/m²K)

24.0
20.0 Numerical
16.0 Analytical
12.0
8.0
4.0
0.0
1 2 3 4 5 6 7 8 9 10 11 12
CASE
297

298 Fig. 4. Comparison of the convective heat transfer coefficient.

299

7.0

6.0
hr (W/m²K)

5.0

4.0

3.0 Numerical
Analytical
2.0

1.0

0.0
1 2 3 4 5 6 7 8 9 10 11 12
CASE
300

301 Fig. 5. Comparison of the radiative heat transfer coefficient.

302

303 As can be observed, the results were consistent, and the maximum deviation found for

304 convection was 2.7 W/m²K for the case of 0 m/s with an emissivity of 0.91 and a current of 200 A.

305 Generally, cases where there is no crosswind displayed large differences, as expected because

306 determining the heat transfer coefficient by convection presents greater uncertainty in those cases.
307 However, the values are still close. Furthermore, in the case of radiation, the convergence of the

308 results is evident, and the measured values are numerically slightly smaller than those determined

309 analytically. Case 8 showed the highest deviation of 1.1 W/m²K. However, the results generally

310 show a very close proximity of the results obtained with the two techniques, which also helps to

311 validate the numerical methodology.

312 Consolidating the values found, it is possible to obtain the correlation of the Nusselt number

313 with the Reynolds number of the flow for the studied cases, as shown in Fig. 6. The curve behavior

314 is similar to that found in the work of Makhkamova et al. [8] for the Lynx cable:

315

30

25

20
Nusselt [-]

15

10

0
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Reynolds [-]
316

317 Fig. 6. Dimensionless evaluation of cross wind flow around the Aero Z Linnet cable.

318

319 3.4 High emissivity coating ink effect

320 Tables 7 and 8 summarize the results obtained in the experimental tests of each wind

321 condition, emissivity and electric current, as well as the corresponding percentage reduction in the

322 conductor temperature at steady-state. As indicated in the tables, the cables that were painted

323 showed a decrease in temperature compared to the unpainted cables for both natural convection and
324 forced convection. In natural convection, there is an average reduction of 10.5 % in the standard

325 cable temperature. For forced convection with the incidence of 1 m/s wind, the painted cables

326 showed an average reduction of 3.8 % in temperature. Of course, when using the high-emissivity

327 coating, the emissivity of the surface becomes approximately four times the normal value,

328 increasing the effect of radiation energy dissipation. This effect is more significant when the

329 radiation is more important in the total heat dissipation, as in the case of natural convection (37 %

330 for the emissivity of 0.91). For high wind speeds, such as greater than 1 m/s, the coating ink usage

331 showed no significant average gain for practical applications (1.6 %).

332

333 Table 7

334 High-emissivity coating ink effect, 200 A.

Wind speed Emissivity Average temperature Temperature reduction

(m/s) (-) (ºC) (%)

0 0.23 38
7.9
0 0.91 35

1 0.23 31
3.2
1 0.91 30

3 0.23 29
0.5
3 0.91 28

335

336 Table 8

337 High-emissivity coating ink effect, 400 A.

Wind speed Emissivity Average temperature Temperature reduction

(m/s) (-) (ºC) (%)

 0 0.23 69
13.0
0 0.91 60

1 0.23 45
4.4
1 0.91 43

3 0.23 36
2.8
3 0.91 35

338

339

340 From the presented results, is possible to see that the use of coating paint is interesting for

341 application in transmission lines installed in areas with low wind speeds (near 0 m/s), being better

342 in colder regions where the radiative heat loss is even greater. It is also possible to show that the

343 greater the current value, that is, the greater the heat dissipated by the Joule effect, the more

344 interesting the usage of coating ink becomes. When comparing the above tables, it is possible to see

345 that the largest reduction is given for the higher current of 400 A. In the case of higher gain (13 %

346 reduction), it is possible to increase the electric current level by 22 % (490 A) and maintain the

347 same temperature (69 °C).

348

349 4. Conclusion

350 With the work done, it was possible to simulate and evaluate the thermal behavior of the

351 commercial cable Aero Z Linnet with and without coating ink.

352 All simulations have shown a good approximation of the experimental data (with an error

353 below 10 %). This therefore confirms that the model adequately reproduces the heat transfer

354 phenomena and fluid flow around conductor cables in aerial power transmission lines. In general,

355 the temperatures obtained numerically were underestimated with respect to practical data.
356 The analytical results were also lower than the results measured in wind tunnel tests. Thus, a

357 close proximity between the numerical and analytical results was obtained, with an average

358 difference of 2.3 %.

359 The three evaluated standards showed very similar results for temperature prediction in

360 cases where there is no direct incidence of sunlight. When the sun is included in the analysis, it is

361 possible to obtain differences of up to 8 %, with the MORGAN methodology as the one that is most

362 different (lower temperatures) over the others (CIGRE and IEEE).

363 It has been proven that the use of high-emissivity coating ink enables the reduction of the

364 mean cable temperature. However, it is clear the most significant effect on the conditions under

365 which the influence of convection in the total heat transfer is lower. Therefore, the largest gain was

366 obtained for the case without crossflow wind (natural convection). Furthermore, the use of the ink is

367 also more effective the higher the current level is, in which case the conductor temperature is higher

368 and thus the loss of radiation energy is significant. Reductions of up to 13 % in temperature value

369 have been obtained. For this same case, the use of ink coverage allows raising the electric current

370 by 22 % while maintaining the same average temperature value of the previous cable.

371

372 Acknowledgements
373 The authors would like to thank the ANEEL-CEMIG P&D 309 project and NEXANS companies
374 for supporting the project developed in this work.
375

376

377 References
378 [1] ALMEIDA, V. F. de. Metodologia para avaliar o impacto de técnicas de repotenciação no
379 desempenho de linhas de transmissão em regime permanente. 2013. 173 f. Dissertação
380 (Mestrado em Engenharia Elétrica)- Universidade de São Paulo, São Carlos, 2013.
381 [2] CIGRE-Task Force B2.12.3. Sag-Tension Calculation Methods for Overhead Lines. June,
382 2007.
383 [3] CIGRE-Task Force B2.12.6. Guide for selection of weather parameters for bare overhead
384 conductor ratings. December 28, 2005.
385 [4] MORGAN, V. T., “The Thermal Rating of Overhead-Line Conductors, Part I The Steady-
386 State Thermal Model”, CSIRO Division of Applied Phisics, Sydney, Australia, 1981.
387 [5] IEEE Std 738™-2006. IEEE Standard for Calculating the Current-Temperature of Bare
388 Overhead Conductors, January 30, 2007.
389 [6] CIGRE- WG B2.43. Guide for thermal rating calculations of overhead lines. 2014.
390 [7] EAST CENTRAL AREA RELIABILITY COORDINATION AGREEMENT- Paper 68-
391 TAP-28. Transmission Conductors Thermal Ratings.
392 [8] MAKHKAMOVA, I.; MAHKAMOV, K.; TAYLOR, P. CFD thermal modelling of Lynx
393 overhead conductors in distribution networks with integrated Renewable Energy Driven
394 Generators. Applied Thermal Engineering, v. 58, p. 522-535, 2013.
395 [9] MORGAN, V. T. The radial temperature distribution and effective radial thermal
396 conductivity in bare solid and stranded conductors. IEEE Transactions on Power Delivery,
397 v. 5, n. 3, p. 1443-1452, jul. 1990.
398 [10] MENTER, F. R. Turbulence Modeling for Engineering Flows, 2013. Disponível em:
399 <http://cfd.spbstu.ru/agarbaruk/c/document_library/DLFE-41517.pdf>. Acesso em: 12 mar.
400 2015.
401 [11] PORTELA, T. M. Avaliação teórica e experimental da transferência de calor em
402 cabos de linhas aéreas com foco na análise da ampacidade. 2014. 175 f. Dissertação
403 (Mestrado em Engenharia Mecânica)- Universidade Federal de Minas Gerais, Minas Gerais,
404 2014.
405 [12] CARDOU, A. Stick-slip Mechanical Models for Overhead Electrical Conductors in
406 Bending. GREMCA, 2013; ISBN 978-2-9812337-2-1.
407 [13] SRIDHAR, M.; YOVANOVICH, M. Elastoplastic Contact Conductance Model for
408 Isotropic Conforming Rough Surfaces and Comparison with Experiments. Journal of Heat
409 Transfer, v. 118, n. 1, p. 3-9, 1996.
410 [14] CELIK, I. B., GHIA, U., ROACHE, P. J., FREITAS, C. J., COLEMAN, H., RAAD,
411 P. E., Procedure for Estimation and Reporting of Uncertainty Due to Discretization in CFD
412 Applications. Journal of Fluids Engineering, v. 130, n. 7, p. 078001-1 - 078001-4, 2008.
413 [15] MENTER, F.R. Review of the Shear-Stress Transport Turbulence Model Experience
414 from an Industrial Perspective. Int. J. of Comp. Fluid Dynamics, v. 23, n. 4, p. 305-316,
415 2009.
416 [16] WHITE, F. M. Viscous Fluid Flow. 3. ed. Singapore: MC GRAW HILL, 2006.
417 [17] ALEKHIN, V.; ANTIPIN, A.; GORODILOV, S.; KHRAMTSOV, S. Numerical
418 simulation of wind loads on high rise buildings. In: CONVR 2013, London. Proceedings of
419 the 13th International Conference on Construction Applications of Virtual Reality. London:
420 Teesside University, oct. 2013.
421 [18] TEIXEIRA, G. G. D. Confiabilidade metrológica em termografia aplicada em
422 sistemas elétricos. 2012. 146 f. Dissertação (Mestrado em Engenharia Mecânica)-
423 Universidade Federal de Minas Gerais, Belo Horizonte, 2012.