PF-03

RADIAL TEMPERATURE DISTRIBUTION OF AAAC OVERHEAD LINE

IN STATIONARY AND TRANSIENT CONDITIONS
*
M. Kang , M. Strobach and C. M. Franck
Power Systems and High Voltage Laboratories, ETH Zurich,
Physikstrasse 3, 8092 Zurich, Switzerland
*Email: [email protected]

Abstract: Commonly, dynamic line rating (DLR) calculates conductor temperature with a
basic energy balance equation, which assumes homogeneous conductor temperature.
With this 0D temperature assumption, however, there is a risk that inner layers of the
conductor might exceed the limit temperature, if the rating current approaches its
ampacity. Since conductor overheating should be avoided for operational safety, it is
valuable to investigate radial temperature distributions of conductors. Throughout this
study, therefore, homogeneous 1D conductor model with a radial temperature distribution
is simulated in steady and transient states. For steady state simulations, radial
temperature distributions under various ambient conditions are studied. Higher wind
velocity, lower solar radiation, and lower ambient temperature make the temperature
gradients more pronounced. Also, transient temperature distribution is modeled with a
various stepwise current changes. When the core temperature reaches 80°C,
momentous radial temperature distributions are analyzed. Compared to the steady state,
transient phases show smaller temperature differences between core and surface. Both
lower initial and higher final currents result in a flatter radial temperature distribution,
whereas higher final currents cause faster transient heating. Compared to 0D calculations,
the 1D model with a radial temperature distribution reduces the ampacity and prevents
conductor core overheating.

1 INTRODUCTION enough margins exist between allowable and

actual line temperatures. Therefore, the core
Aluminum is the most frequently used material [1, temperatures remain lower than the allowable
pp. 200] for overhead lines (OHLs) because of its temperature in general.
advantages. By mixing magnesium and silicon into
aluminum, all aluminum alloy conductor (AAAC) The 0D models might not be suitable for dynamic
achieves more than 1.5 times higher strength, line rating (DLR). DLR offers an improved
while sacrificing only 5% of ampacity (current temperature management of OHL conductors. It is
carrying capacity, Iccc) [1, pp. 206]. enabled by accessing and real-time processing of
on-site conductor temperature and weather
Despite its advantages, the usability of AAAC is information. DLR then not only optimizes current
limited by its low operation temperature. AAACs loading of OHLs but also, minimizes operational
start recrystallization when they are heated up to risk by improving security and safety. Therefore,
over 100°C, and lose the mechanical strength [1, the conductors can be operated near Tlim. For this
pp. 202]. Also, its high thermal expansion case, the 0D models may result in conductor core
coefficient and low specific weight make AAACs overheating, because they ignore the T(r)
more vulnerable to sagging and vibration [1, pp. distribution.
206]. For these reasons, the maximum allowable
temperature Tlim of AAAC is limited to 80°C [3]. Tlim To analyze the limits of 0D models on DLR, a
determines the ampacity under the given ambient homogeneous cylindrical conductor is modeled
conditions [4], since higher temperatures result in and its radial temperature distribution is studied.
larger sags and worse mechanical integrity of The radial temperature distribution was researched
conductors. extensively by Morgan [7], and its significance on
DLR was already suggested by Douglass [8].
Therefore, Cigre [5], IEC [6], and IEEE [4] Nonetheless, parametric researches on stationary
developed standard OHL models analyzing the and transient behavior of AAACs are still required.
conductor temperature. To simplify the calculation, Throughout this paper, therefore, the conductor
the models commonly disregard the radial temperature T(r) is simulated with a 1D model for
temperature distribution, T(r) of the conductor. various stationary and transient states. Several
groups of relevant weather conditions are varied
These non-dimensional (0D) models are beneficial for steady state simulation, and several current
for the static line rating (SLR). In SLR, the rating steps are applied to transient phase simulation.
currents are kept low for security reasons [10] and

1970
 PF-03

2 1D CONDUCTOR MODELING

2.1 System and environment The T(r) distribution is described by the heat
Conductor: For the modeling, the conductor is equation in the form of [7]:
simplified as a homogeneous cylinder with infinite
length. To study T(r) only, a uniform surface [ ]
temperature Ts without angular and axial
temperature distribution is applied. An AAAC with a
2
conductor cross section A=550mm is used as a where r is the conductor radius (0<r<D/2) and PJ is
reference material. It consists of 61 strands with Joule heating power per unit length. The boundary
d=3.4mm, and has total conductor diameter conditions for (2) are:
D=30.6mm. As the conductor geometry is
comparable to 61/3.5mm AAC in [7], the same T has a finite value at r=0,
effective radial thermal conductivity kr = 1.23 W/mK
is assumed. The other typical material properties of | , and |
AAAC are adopted from [5], and presented in
Table 1.
For AAACs, [5] assumed PJ as
Table 1: Reference AAAC properties at 20°C

Properties Value
Diameter, D 30.51 [mm] where kj=1.023 is a factor for accounting skin
DC Resistivity,  32.7 nm] effect, I (A) is an effective current, Rdc () is the DC
Density,  2703 kg/m3 resistance at 20°C , (1/K) is the thermal resistive
Heat capacity, c  909 J/kgK
coefficient, and Tav is an average line temperature
Absorptivity,  0.5
which is given by:
Emissivity,  0.5
Effective radial thermal conductivity, kr 1.23 [W/mK]

This is used for conductor temperature in the Cigre

Environment: Ambient temperature Ta, 0D model. In addition, the temperature difference
perpendicular wind velocity Vw and solar radiation between Tc and Ts of AAAC is [5]:
S are applied around the whole conductor
circumference. Utilities in Switzerland assume
conservative weather condition for the line rating
[9]. For the operational safety, Vw and S are fixed
to a worst-case value while Ta varies to apply and it is used for analyzing the steady state
seasonal weather differences [10]. The conditions temperature gradient.
are shown in Table 2.
For the transient phase, any difference between
Table 2: Seasonal weather conditions heat gain and loss will change the conductor
temperature [5]. This can be expressed as:
Parameters
2
Ta [°C] Vw [m/s] S [W/m ]
Winter 10 { } { }
Intermediate 20 0.5 900
Summer 40 where, (kg/m).

2.3 Simulation methods

2.2 Heat equations The heat transfer module of COMSOL Multiphysics
The conductor system and environment satisfy the is used for the 1D model simulations. For the
thermal equilibrium at the steady state [5]. The steady state and transient phase evaluation,
power gain of the system is the sum of Joule, ‘stationary’ and ‘time-dependent’ studies of the
ferromagnetic, solar, and corona heating per unit program are used, respectively.
time (PJ, PM, PS, and Pi, respectively). It balances
the power loss, which is the sum of convective, Steady state: The steady state T(r) distribution for
radiative and evaporative cooling per unit time (Pc, Tlim(=80°C) is simulated with different weather
Pr, and Pw, respectively). According to [7], the conditions. Each weather parameter is varied to
reduced equation for AAAC conductors is: analyze the influence on conductor temperature. In
addition to the weather conditions in Table 2,

1971
 PF-03

different values of Vw (0.5, 2.5, and 4.5m/s) and S parameter change. In addition to seasonal weather
2
(500, 900, and 1200W/m ) are used in the conditions in Table 2, each weather parameter is
simulation. For each condition, Iccc is calculated. manipulated to investigate steady state T(r)
These values are then used as references for distribution.
transient phase simulation.

Assuming similar weather conditions, steady state

ampacity from 1D and 0D models are calculated
and compared. All of them use the same maximum
allowable temperature, 1D conductor
model has the maximum temperature at the
conductor core, and the ampacity is the current at
.

Both IEEE and IEC models ignore the T(r)

distribution and the ampacity is the current at
. Among the 0D models, only Cigre model Figure 1: T(r), comparison with different Ta (left),
uses simplified assumptions for the radial Vw (center) and S (right), when .
temperature distribution, and its ampacity is
defined as the current at Although Cigre Among the weather parameters, Vw has the largest
model denotes temperature differences between influence on T(r), and S has the smallest. The
and [5], the radial temperature distribution is not equilibrium ampacity increases significantly with Vw,
considered for its ampacity analysis. as the heat loss Pc increases with the wind velocity.
Analysis for each parameter is done below.
Transient phase: Transient simulations are done
with the ‘winter’ weather condition of Table 2. Ambient temperature: To control the influence of
Therefore, the steady state ampacity of this other parameters, fixed value of Vw=0.5m/s and
2
condition is used as a reference. For the transient S=900W/m are used for Ta simulation. Lower Ta
phase, the allowable time interval tlim is defined as results in larger T(r) gradients. Both Pc and Pr of (1)
the elapsed time to raise the conductor increase with decreasing Ta [5]. Therefore, higher
temperature to Tlim. By raising the current stepwise, currents are needed to increase PJ for equilibrium.
transient T(r) distributions at t = tlim are Since Ta is the only variable of the static weather
investigated. conditions, this part of simulation result can be
used for SLR as well. To reach a core temperature
The stepwise change is done from initial current Ii of 80°C under the three different values of Ta=10,
(30% and 60% Iccc) to final current If (120%, 150%, 20, and 40°C, current values of I=962.3A, 886.0A,
and 200% Iccc). These situations should be and 703.8A are needed, respectively.
investigated extreme situations for a line operating
first in (N)- and then in (N-1)-condition. The plans Wind velocity: To study the influence of Vw only,
2
for the transient current change are presented in fixed values of Ta=10°C and S=900W/m are used
Table 3. for this set of simulation. High wind velocities
decrease Ts significantly and cause larger T(r)
Table 3: Transient current change plans gradients at the ampacity limit. Pc strongly depends
on Vw [5]. For the equilibrium of (1), ampacity
(% Iccc) increase with Vw. To reach a core temperature of
80°C under the three different values of Vw =0.5,
120% 150% 200%
2.5 and 4.5 m/s, the current values of I=962.3A,
Plan A Plan B Plan C 1395.7A, and 1670.8A are needed, respectively.
30%
(30→120) (30→150) (30→200)
(% Iccc)

As mentioned above, the wind velocity is the most

Plan D Plan E Plan F influential parameter.
60%
(60→120) (60→150) (60→200)
Solar radiation: Similarly, to cancel out the
influence of the other parameters, Ta=10°C and
3 RESULTS AND DISCUSSION Vw=0.5m/s are used. Decreasing the solar
radiation causes a reduction of the total heat gain
3.1 Influences of weather parameters on PS. When the other conditions are kept fixed, PJ
steady state temperature distribution has to be increased to satisfy (1) with a higher
current [5]. To reach a core temperature of 80°C
By limiting Tc=Tlim=80°C, the steady state T(r)
under the three different values of S = 500, 900,
distributions for different weather conditions are 2
and 1200 W/m , current values of I=990.3A,
obtained (See Figure 1). Since each weather
962.3A, and 940.7A are needed, respectively.
parameter influences the elements of heat balance
Compared to the other parameters, solar radiation
equation (1), T(r) and ampacity are varied with
has relatively small influence on T(r) distribution.

1972
 PF-03

3.2 Steady state ampacity comparison remain for a long time. According to [1, pp. 203],
the mechanical strength of aluminum decreases
The ampacity values for the standard weather
after long-term exposures to high temperature.
conditions are presented in Table 4 for the 0D
After 100 hours, the residual strength is around 80%
models according to IEEE and Cigre, as well as for
of the initial strength, and after 1000 hours, it
the 1D model. Respecting the Tlim=80°C value for
decreases to 60%.
the core temperature results in a reduction of line
ampacity as well.

Table 4: Seasonal ampacity comparison of

different models

0D Iccc (A) 1D Iccc (A)

IEEE Cigre COMSOL
(Ts=80°C) (Tav=80°C) (Tc=80°C)
Winter 1069.9 1024.4 962.3
Intermediate 996.9 953.0 886.0
Summer 825.3 786.5 703.8

Figure 3: T(r) comparison with different Ta (left),

As shown in Figure 2, if the ampacity value Vw (center) and S (right), when .
calculated from the IEEE (or IEC) 0D model would
be used in the 1D model, the whole conductor 3.3 Transient phase analysis
temperature (except its surface) exceeds Tlim and
reaches a core temperature of 86.7°C. When the The influence of initial and final current is analyzed.
Cigre 0D model is used, half of the conductor area The ampacity of the 1D model at ‘Winter’ condition
exceeds Tlim, and Tc=83.2°C Therefore, the T(r) (Iccc=962.3A, see Table 4) is used as a reference
distribution should be properly considered to ampacity of the transient phase simulations. From
respect the line security. the initial currents 30% and 60% Iccc, Initial steady
state Tc=23.4°C and 38.4°C are obtained,
respectively (see Figure 4). Currents for the initial
and final states are presented in Table 5.

Table 5: Initial and final states for transient phase

Initial currents Final currents

% Iccc 30% 60% 120% 150% 200%
Tlim=80°C I (A) 289 577 1155 1443 1925

The temporal evolutions of under the 6 different

current steps are depicted in Figure 4. Qualitatively,
higher initial current decreases the heating
duration, and results in smaller Also, the
temperature rises faster when final current
becomes higher.
Figure 2: T(r) comparison of COMSOL 1D (solid,
Iccc=1069.9A), Cigre 0D (dashed, Iccc =1024.4A),
and IEEE 0D (dot, Iccc =962.3A) models at ‘Winter’
ambient condition.

The core temperature increase becomes worse

when the weather conditions allow a higher
ampacity, since then PJ and T increase in
equations (4) and (6).

To analyze the influence of overrated ampacity, T(r)

distributions with IEEE 0D model ampacity are
presented (see Figure 3). Among the simulation
results, the maximum Tc=105°C is obtained under
the weather condition of Ta=10°C, Vw=4.5m/s, and
2
S=900W/m . [9] shows that this weather condition
occurs frequently in Switzerland, and it may even

1973
 PF-03

surface temperature increases, and difference

between each transient case is around 0.1°C.

Tlim=80°C

Figure 4: t vsTc with various current plans

Influence of Ii on T(r): With a lower initial current,

it is obvious that increases. Also, the
temperature distribution becomes slightly flatter
because the initial steady state T(r) distribution is
Figure 6: T(r) comparison of transient case D, E,
flatter with a smaller initial current. However, the
and F
influence of Ii has only minor effect on conductor
temperature distribution (cf. Fig. 5). This is
because the Joule heating power caused by the
overcurrent is not large enough compared to the
3.4 Time variables of transient phases
radial heat flux to justify complete adiabatic heating.

Table 6: allowable time interval for different

models

Plans 0D (s) 1D (s)

(change in % Iccc) ,Ts= Tlim ,Tc= Tlim (s)
A (30→120) 1779 1479 300
B (30→150) 784 684 100
C (30→200) 366 328 38
D (60→120) 1606 1263 343
E (60→150) 647 548 99
F (60→200) 288 248 40

with and without T(r) distribution is compared

Figure 5: T(r) comparison of transient case C and and their differences are presented in Table 6. tlim
F decreases by increasing If. For all plans, tlim is
smaller for the 1D model. If a 0D model would be
Influence of If on T(r): As higher If increases the used for transient phase analysis, there is a risk of
temperature rise rate, decreases. In other overheating the wire for Δtlim, as Tc exceeds Tlim for
words, the time to lose heat to environment is the duration.
reduced and the process becomes more adiabatic.
Therefore, the momentous temperature distribution 4 CONCLUSION
is not changed much from the shape of its initial
steady state T(r) distribution, if the applied final The possible benefits of considering the radial
temperature is higher. However, If of 120% to 200% temperature distribution for DLR are elucidated by
ampacity is not enough to observe a dramatic a simple 1D conductor model. If 0D models are
difference in T(r) distribution. For those final used for steady state ampacity calculation, Tc
current values, the momentous T(r) distribution at might exceed Tlim by up to 25°C even in
Tc=80°C is not very different form the steady state reasonable weather conditions. Wind speed is the
T(r) curve (see Figure 6). With higher If, the most influential parameter for the radial
temperature gradient, as it has the largest effect for

1974
 PF-03

increasing the ampacity. On the other hand, 6 REFERENCES

changes in solar radiation do not affect the T(r)
distribution significantly. [1] F. Kiessling et al., Overhead Power Lines,
Springer, 2003.
0D models neglect the T(r) distribution and assume
[2] M. Farzaneh et al., Electrical Design of
Ts or Tc as conductor temperature. This could be
Overhead Power Transmission Lines,
dangerous, since the long-term exposure to high
McGraw-Hill, 2013.
temperatures permanently deteriorates the
mechanical integrity of the conductor. This can be [3] EN 50331-3-4: Overhead electrical lines
avoided by considering a radial temperature exceeding AC 45 kV. Part 3-4: National
distribution. Even with the simple 1D model, the Normative Aspects for Germany. Brussels,
overheating of conductor core can be prevented. CENELEC, 2001.
By applying the model to calculation, ampacity at
different weather conditions is reduced by 10 to [4] Standard for Calculating the Current-
18%. This modified steady state ampacity with Temperature of Bare Overhead Conductors,
IEEE Std 738, 2006.
each weather condition might offer accurate
information for the overhead line operators. [5] Thermal Behavior of Overhead Conductors,
Cigre Working Group 22.12, August 2002.
Also, with a stepwise current change, transient
[6] Overhead electrical conductors-Calculation
phases are simulated. Intensity of initial current
and final current is the governing factors for the methods for stranded bare conductors, IEC TR
transient characteristics. With a lower initial current, 61597, 1995.
the transient temperature profile is flatter. Higher [7] V. T. Morgan, The radial Temperature
final current results in faster temperature increase. Distribution and Effective Radial Thermal
Also, the T(r) becomes flatter with higher If, by Conductivity in Bare Solid and Stranded
keeping the conductor away from the heat transfer. Conductors, IEEE Transactions on Power
Delivery, pp. 1443-1452, 1990.
can be adjusted with the 1D model. By
[8] D. A. Douglass, Radial and Axial Temperature
regulating current with Tc, the 1D model is able to
Gradients in Bare Stranded Conductors, IEEE
control the overhead line temperature. Compared
Transactions on Power Delivery, Vol. PWRD-1,
to 0D models, it prohibits the extra time for heating
No. 2, April 1986.
the inner layers of a conductor. This is important
for real time controlling of current with DLR. [9] Leitungsverordnung LeV, SR 734.31, 30.
March 1994.
5 ACKNOWLEDGMENTS
[10] Guide for Selection of Weather Parameters for
Bare Overhdad Conductor Ratings, Cigre
This study is financially supported by Swiss electric
Working Group B2.12, 2006.
research, Pfisterer, and Swissgrid.

1975