RATING OF CABLES IN A NONUNIFORM THERMAL

ENVIRONMENT

G. Anders, J.M. Braun, M. Vainberg, S . Rizzetto Heinrich Brakelmann

Ontario Hydro Technologies Gerhard-Mercator-Universitat
Toronto, Canada Duisburg, Germany

Abstract. When cables cross external heat sources or cross development of a systematic method for determining the
the areas of high thermal resistivity, the conductor feasibility of extending cable life and/or increasing current
temperature will be higher than the values attained outside of ratings is of paramount importance. This is particularly so
the unfavourable area. For perpendicular and oblique when cables cross unfavorable thermal environment, e.g.,
crossings as well as street crossings, these efsects are usually crossing other cables, steam pipes, parking lots, streets, and so
ignored for distribution circuits, whereas for transmission on.
cables, corrective actions in physical installation condition When power cables cross regions with unfavourable
are sometimes taken. Almost never analytical solutions are thermal conditions, temperatures higher than the design value
used to determine the efSect of external heat source and high can occur. If the region is wide enough, the rating of the cable
thermal resistivity regions on the ampacity of the rated cable. will usually be based on the assumption that the entire route is
The main reason no computations are performed is an characterized by the same conditions. In a majority of cases,
absence of either derating formulas or derating tables the unfavourable thermal environment will be very short,
(curves) and not the lack of a need. Tof i l l this gap, analytical usually a few meters (e.g., street crossing or perpendicular
solutions for the computation of the derating factors have cable crossing). In these cases, the effect of the crossing is
been developed and sample computational results are usually ignored. However, the conductor temperature in such
presented in this paper. Alternative approach to address the cases may be much higher than in the remainder of the route
issue of higher operating temperatures is to design and install and cable derating is required.
dynamic feeder rating systems. This topic is also discussed in The most accurate method of assessing the short- and
the paper. long-term cable ratings at any given point in time is to perform
continuous measurements of conductor temperature. Since, in
1. Introduction the great majority of cases, this is not technically feasible, the
next most effective option is to measure cablelpipe surface
Underground cables are far more expensive to install temperature and the current flowing in the conductor and use
and maintain than overhead lines. The greater cost of this information as an input to the mathematical models.
underground installations reflects the high cost of the These models compute real-time conductor temperature and
equipment, labor and time necessary to manufacture the cable, cable ratings.
to excavate and backfill the trench, and to install the cable.
Because of the extra expense, most underground installations In this paper, the issues involved in rating of cables in a
nonuniform thermal environment are discussed. Analytical as
are constructed in congested urban areas and as leads from
generating plants and in substations. The large capital cost well as experimental methods are reviewed and numerical
associated with cable installations also makes it necessary that examples presented.
they are used to their full potential. 2. Unfavourable thermal environment
Information about the maximum current-carrying
capacity, which a cable can tolerate throughout its life without 2.1 Cable crossings
risking deterioration or damage, is extremely important in
In congested urban areas, we often find power cables
power cable engineering and operation. Ampacity values are
crossing steam pipes or other cables. When a rated cable is
required for every new cable installation, as well as for cable
crossed by another heat source, its temperature will be higher
systems in operation. With some underground transmission
than if it was isolated. As shown in [ I ] , this additional
cable circuits approaching the end of their design life, the
conductor temperature rise may exceed 20°C. Therefore, the

0-7803-5515-6/99/$10.00 Q 1999 IEEE 83

ampacity of the rated cable should be reduced accordingly. 0 usage of larger cable
Ten and often 20% reduction may be required. This is much installation of ventilating manholes at crossing points
larger than up to 5% reduction considered by some utilities. 0 at steam heat crossings, utilities sometimes build a
On the other hand, many utilities do not derate cables crossed Styrofoam (closed cell) box around the steam heat pipe,
by other heat sources. This practice may lead to premature then fill the box with Gilsulate, an underground
cable failures, as reported for example in [2]. An informal insulating material used by steam heating companies
survey conducted by the authors has clearly indicated a need to in one instance, for 90-degree crossings, a utility
establish derating factors for cable crossings. sometimes would follow a very rough approximation by
The temperature increase in the rated cable resulting doing ratings assuming parallel, but at twice the actual
from the presence of another heat source will depend on vertical separation distance.
several parameters. The most important are: Finally, several engineers indicated that the main reason
the amount of heat dissipated by the crossing heat source no computations are performed is an absence of either derating
formulas or derating tables (curves) and not the lack of a need.
the distance between rated cable and the heat source at the
All indicated that it would be very beneficial if a simple
point of intersection
derating method was available. Such a method is presented in
the angle between the heat source and the cable (the more [I] and briefly summarized below.
parallel the two are, the larger the influence of the heat
source on the rated cable) Suppose that cable 2 is to be derated because of the
0 the size of the rated cable. heat Wl (W/m) generated by cable 1. Let A8- be the
In order to find out how the cable crossings are dealt maximum permissible conductor temperature rise of the
with in electric utilities, an informal survey was conducted conductor of cable 2 for current dependent losses. The
among several engineers active in the field of cable ratings [ 11. reduced rating I,, of cable 2 can be obtained by combining as
In general, all the contacted utilities deal with the subject of
cable crossings in a similar way. For distribution installations [I1
(low and medium voltage), in a great majority of cases, no
derating is considered at all. The most often quoted reason
was that the cables are rated very conservatively (e.g., high
soil thermal resistivity or high ground ambient temperature are
assumed). In some cases, for example in Sweden, special where
precautions are taken in the case when a distribution cable (0.4 =is the rating of the isolated cable 2.
1,
- 24 kV) crosses a district heating pipe. In particular, the
Swedish standards state that the cable should always be above W,, = losses in conductor of cable 1 (W/m).
the pipe, at the distance not smaller than 20 cm as =the mutual thermal resistance between cables 1 and 2
T,*
perpendicular as possible. A heat isolating plate 50 mm thick
is required between cable and the pipe system. In one utility, (K."
for cable ratings at the steam pipe crossings it is assumed that Thus, the derating factor can be obtained from equation
the ambient temperature is higher by 5°C than the values used ( 1 ) when the mutual thermal resistance T,, is known. The
for isolated cables, which results typically in 3% decrease in equation for the evaluation of this resistance is developed in
the cable rating. In several countries, a minimal distance is [I].
specified between rated cable and the external heat source. To illustrate the effect of cable crossings, we will
This distance varies between 20 and 50 cm. consider a crossing of a IO kV XLPE cable circuit and a 138
Situation is only slightly different when a high voltage kV high-pressure, oil-filled, pipe-type cable illustrated in
cable is involved. With the exception of France, no Figure 1. The cables and the their characteristics are described
computations are performed to determine the actual effect of in [3] (cable model 1 and 3).
cable crossing another heat source (EDF may conduct a 3
dimensional finite element analysis in some important cases). Cable No. 1 is a I O kV single conductor XLPE cable.
However, most utilities would take special actions to reduce The cable has copper screen wires and a PVC jacket. The
the mutual heating effect. Several approaches have been laying conditions are assumed as follows: cables are located
applied: 1.2 m below the ground in a flat configuration. Uniform soil
isolation of the heat source from the cable properties are assumed throughout. Spacing between cables is
usage of heat transfer pipes equal to one cable diameter (spacing between centers equal to
usage of thermal backfills in an increased trench cross- two cable diameters). Ambient soil temperature is 25°C. The
section thermal resistivity of the soil is equal to 0.8 K m " . The
usage of insulating liquid circulation cables are solidly bonded and are regularly transposed.

84
y = 0.91 m c
v
'I /
pipe-type cable I / cable 2
Y
A
n &
0.95

0.9
-

- Parallel Cable 1 -- Cable1

--
Z1

0.072 m

y = 1.2m

"
0.5 " 30 i
60

Crossing Angle (deg.)

o
90

Figure 1 Cable configurationf o r the numerical example Figure 2. Derating factors as a function of the crossing angle

Cable No. 2 is a 138 kV high-pressure liquid-filled The temperature influence of the external heat source
(HPLF) cable. All parameters are the same as in the on the rated cable will depend on both the crossing angle and
NeherMcGrath (1957) paper. The cables lie in a cradle the distance from the crossing. Figure 3 illustrates this
configuration and operate at 85% load-loss factor. The laying dependence.
conditions are assumed as follows. The cables are located in a
From Figure 3, it is apparent that the influence of the
steel pipe 8.625 inches outside diameter. The pipe is covered distribution circuit (cable 1) on the pipe-type cable (cable 2)
with an asphalt mastic covering 0.5 inch thick. The center of
disappears at the distance of exceeding 6 m for crossings with
the pipe is located three feet (0.91 m) below the ground.
angles smaller than 60". On the other hand, for the 10 kV
The derating factors will, of course be affected by the cable, this influence extends to about 5 m.
value of the crossing angle. Figure 2 shows this relation for 25
both cables. We recall that p = 90" represents the parallel t
route of both circuits.

From Fig. 2, we can observe that the decrease of the G

15
derating factor is more pronounced for the pipe-type cable as e
1

the circuits become more parallel. This higher sensitivity is E

partially caused by much smaller permissible temperature rise E" 10
due to joule losses for the pipe-type cable than for the
distribution circuit and can also be explained by the fact that, 5
in this example, the distribution circuit has about 25% higher
losses than the pipe-type cable. Thus, for the 138 kV cable,
n
the derating factor decreases from 0.76 for perpendicular -0 1 2 3 4 5 6
crossing to 0.62 when the circuits are parallel. A conservative Distance from the Intersection (m)
practice, used by many utilities in cases like this, would be to
Figure3. Temperature influence on the rated cable as a
assume that the cables are parallel. For the parallel laying, the function of the crossing angle and the distance from the crossing.
ampacity of the pipe type cable is 562 A. Thus, when this
conservative assumption is used, we would unnecessarily Referring to Figure 3, we note that, as expected, the
decreases the ampacity of the pipe-type cable by 120 A or more parallel the cables are, the higher will the influence
17.6%. temperature be. This influence is much greater for the pipe-
type cable, and hence lower derating factor is required. The
influence temperature can be quite high at the point of
intersection even for a perpendicular crossing and reaches
17°C in our example (this can be compared with the 23°C
influence when the circuits are parallel).

85
2.2 Street crossing
Power cables are usually laid in congested urban areas
and very seldom the thermal conditions along their route are
constant. Cables are usually laid along the street length.
Occasionally, they will be located along a green belt and they
would cross a street at a perpendicular angle. Even if no street
crossing occurs, the presence of the trees will considerably
change the local thermal environment. When a rated cable
crosses the area of unfavourable thermal characteristics, its
temperature will be higher than if it was in a uniform
Cable
environment for which the rating is normally calculated. A
common occurrence of cables crossing a paved parking lot
Region 2
may raise the conductor temperature by as much as 10°C if the
soil composition is the same as outside the crossing, [4]. As is
shown in [SI, the additional conductor temperature rise may
exceed 20°C in some particularly unfavourable conditions.
Therefore, the ampacity of the rated cable should be reduced
accordingly. Up to 40% reduction may be required. On the
other hand, many utilities do not derate cables crossing
unfavourable thermal environment. This practice may lead to
premature cable failures as the circuits become more heavily
loaded.
The magnitude of the required cable derating will
depend on the following factors:

1. The length of the unfavourable portion of the route;

2. The difference in the thermal resistivities of the native soil
and the crossed area;
3. The difference in the ambient temperature at the two
sections; and
Figure 4. Cable crossing thermally unfavourable region
4. The laying condition of the cable in the unfavourable
environment. The derating factor and the derated current for these
In the presentation of the results of a cable crossing a conditions are equal to: d = 0.574, I = 5 18 A
street, we will use cable laying example shown in Figure 4. In The reduction of ampacity is very significant. Figure 5
this figure, the cable is rated for the conditions characterizing shows the variation of the derating factor with changes of soil
soil 2, and we assume that it crosses a short section of thermal resistivity and the ambient temperature in region 1 .
unfavourable thermal conditions, represented by soil 1. The Let us define two parameters:
length of this section is equal to b m. In addition, the ambient
(1)
temperature may be different in both sections, as illustrated in x=- Ti'" 'amb
Figure 1. When crossing streets, cables are often placed in (2)
short sections of pipes. This is also shown in Figure 4. Ti2' and =-' a m b
We will consider a street crossing of a 138 kV high- where T4 and Oamh represent the external thermal
pressure, oil-filled, pipe-type cable described above. The resistance and ambient temperature in the two regions,
ampacity of this cable for the uniform soil conditions is equal respectively.
to 902 A. In practical cable installations, the value of x can reach
3 and the ambient temperature in the street crossings can easily
The cable crosses a street 10 m wide with the soil be 5°C higher than outside the crossing. Figure 5 shows the
thermal resistivity under the asphalt equal to 2.5 K.m/W. The dependence of the derating factor on parameters x and y and as
ambient temperature in this section is equal to 25°C. a function of the width of the crossing with zo = b / 2 .
We can observe from Figure 5 that the derating factor
becomes almost constant at the width of the road equal 3 m
( zo = 1.5 m). The required reduction in cable rating becomes

86
very significant and reaches 50% when the soil thermal
resistivity is three times greater than in the rest of the route and
the ambient temperature is increased by 5°C.
- - - y = 1.2

1 .o
M
c 0.8
.
Ye

e 0.6 Figure 6. Ontario Hydro's Dynamic Feeder Rating system

&
".7 ~~~

The key performance characteristics of the RTUs are to

0 2 4 6 8 1Om provide reliable acquisition and storage of cable and ambient
20 temperatures, and load current; these data are transferred to the
CPU, typically via modem, at regular intervals under typical
Figure 5. Derating factor as a function of the width of the
crossing, thermal resistance and ambient temperature of the crossing
field conditions, including transient potential rises.
area The distributed RTU units are commercial data
acquisition units requiring a power supply and a link for
Thus, if the pipe-type cable crosses a street 10 m wide, communication to the CPU. The units are capable of long
and the thermal resistivity of the soil is three times higher than term (1-2 week) operation solely on battery power for
the rest of the route, the rating can be based on the uninterrupted collection of data. The units are provided with
unfavourable route conditions alone. However, if the sufficient memory to retain upwards of 500k points of data.
unfavourable region is only 3 m wide (e.g., cable passes in a To overcome possible malfunction of the remote system and
vicinity of a tree), then the cable rating should be reduced by maintain remote operation, the RTUs can be restarted directly
about 30% assuming that the thermal resistivity of the soil is 3 by the CPU. The RTUs have a broad operating range,
time higher in this region than outside. typically from -4OCto +70°C.
3. Dynamic Feeder Rating The technology has proven highly reliable for over 10
years outdoors in Northern Ontario, monitoring transmission
Dynamic feeder rating systems are an important equipment mechanical performance.
alternative to analytical approaches for determining a real-time RTU power can be provided from a conventional ac
conductor operating temperature and corresponding real-time
source or, where appropriate, from a solar battery pack to
cable rating. One such system developed by Ontario Hydro
minimize installation costs. Similarly, to minimize installation
Technologies is briefly described below.
costs, the communications link can be via cellular phone or
3.1 Data Acquisition System dedicated wireless transmitter. Depending on requirements,
RTUs can accommodate up to 25 thermocouple inputs.
Ontario Hydro's Dynamic Feeder Rating (DFR) system
is configured to link several distributed terminal units (RTU), 3.2 DFR Software
positioned at strategic locations along the cable route, to a
DFR program can perform calculations in real-time of
main computing unit (CPU) which is generally installed at a
steady state, and emergency ratings. User specifies the
substation. The system is based on an open architecture,
relevant maximum operating temperatures and the duration of
relying on proven third party sub components. An overview of
the emergency and the software computes the maximum
the OH'S DFR is shown in Figure 6.
current that the cable can carry for the specified conditions.
Operation of the system is under control of the CPU, Calculation of the steady-state rating can be performed in two
typically an industrial grade 233 MHz PC; to maintain modes. In one case, standard NeherMcGrath or IEC 287
continuous service, the CPU is powered by an uninterruptible calculations are performed with the user-specified load-loss
power supply which also has provision for safe shutdown of factor. In the second case, program considers recorded load
the system in case of extended power outages. To overcome variations over the last 24 hours and assumes the same
possible malfunction of the system and maintain remote variations during the following period. Duration of this period
operation, the CPU can be restarted at will by a simple remote is such that the steady-state conditions are achieved.
telephone command. Time-dependent ratings are based on the method
described in the IEC Standards 853-1 and 2. Similarly as in
the steady-state case, two modes of load representation can be

87
considered during the emergency period: (1) a step function As an alternative, real time cable rating systems could
can be applied, or (2) the load curve from the last 24 hours can be applied for the installations where the cables are loaded
be scaled. close to their thermal limits or where emergency conditions
During the steady-state and time-dependent may force short-term high loading.
calculations, the thermal and electrical parameters of the
model are continuously adjusted so that the computed and
measured cable surface temperatures match. REFERENCES
3.3 DFR System
The following are the main features of the OH’S DFR [ l ] Anders, G.J., Brakelmann, H., “Cable Crossings -
system. Derating Considerations”, submitted for publication in the
IEEE Trans. On Power Delivery.
General characteristics:
PC anywhere - main access screen [2] H. Orton, A. MacPhail, V. Buchholtz, “Elevated
Temperature Operation-A Precautionary Note”, IEEE
graphing & storage of historical data PES, Insulated Conductor Committee, Minutes of the 99’
download to local WS Meeting, Houston Texas, April 14-17, 1996, Appendix 5-
alarming capabilities K-I.
dedicated single circuit [3] Anders, G.J., “Rating of Electric Power Cables.
multiple circuits from single CPU Ampacity Computations for Transmission, Distribution
and Industrial Applications”, IEEE Press, 1997, New
Cable types York.
5 through 500 kV
[4] Williams, J.A., “Distributed Fiber Optic Temperature
Paper (HPFF, SCFF), extruded, gas, in air, directly Monitoring Results on 69 kV XLPE Cable in Duct”, IEEE
buried, submarine, ducts, tunnels PES, Insulated Conductor Committee, Minutes of the 101
Single, multiple cables, forced-cooled Meeting, Scottsdale, Arizona, April 20-23, 1997,
n
IJl Appendix 7-D-1.
Loadcurrent [5] Brakelmann, H., Anders, G.J., “Ampacity Reduction
Ambient & cable surface temperatures Factors for Cables Crossing Thermally Unfavourable
Conventional, fiber optic, wireless Regions”, submitted for publication in the IEEE Trans,
On Power Delivev.
Ratings
User definable (e.g., 20 min, 4 h, 24 h, etc.)
Steady-state, time to specified temperature
DFR systems can be used to enhance current-carrying
capacity of power cables as well as to eliminate risk of
overheating. They allow utilization of cable systems to their
maximum capabilities and are particularly useful when a
deferment of costly capital programs is desirable.
4. Conclusions
From the information presented above, it is quite clear
that the presence of external heat sources or unfavourable
thermal environment may have a significant effect on the
rating of power cables. In particular, the derating of 3 to 5%
used by some utilities may be insufficient. Hence, the need for
either derating curves or computational procedures, or both,
seems to be quite evident.

The methodology for developing derating factors has

been developed and preliminary results are presented in this
paper. The derating factors can be applied to either existing or
new cable installations.

88