2130 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 32, NO.

4, AUGUST 2017

Thermal Analysis of Power Cables Installed in Solid

Bottom Trays Using an Equivalent Circuit
Shahriar Saadat, Student Member, IEEE, Akim Borbuev, Student Member, IEEE, and Francisco de León, Fellow, IEEE

Abstract—Cables in ventilated and ladder-type trays have treated as a homogeneous rectangular mass with uniform heat
been extensively studied and are rated according to ANSI/NEMA generation. It is also assumed that all cables are operated at their
standards. The National Electric Code (NEC) provides guidelines ampacity limit (maximum load).
on ampacity for cables installed in ventilated and ladder-type
trays. However, for solid bottom trays, there is very little published The study carried out in [5] used ventilated trays without
material; there are neither standards nor guidelines. This paper cover. The effect of tray covers on cable operating temperature
proposes a methodological approach for the thermal rating of or ampacity was developed by Engmann [6]. It was concluded
power cables installed in solid bottom trays with and without that the ampacities of cables in ventilated trays with cover are
cover. An analog thermal–electrical circuit is derived from first 70% to 75% of the open-top tray ampacities [6].
thermodynamic principles. The circuit parameters are easy to
compute. The method is completely general and is applicable to Uniform heat generation within cable mass was assumed in
trays with any number of cables (not only to integer fill depths). [5], [6]. Harshe and Black introduced ampacity calculation of
The validity of the method is corroborated with numerous cables in ladder type trays with and without covers with non-
finite-element method simulations and laboratory experiments. A uniform heating within the cable pack, i.e. load diversity is
hand-worked example is given for illustration.
taken into account [7], [8]. Harshe and Black also studied two-
Index Terms—Cable ampacity, cable thermal rating, solid dimensional heat flow within the cable mass in [9].
bottom trays. ANSI/NEMA WC 51 ICEA P-54-440 standard [10] is based
on the models established in [5], [7], [8] and provides ampacity
I. INTRODUCTION values for cables installed in ventilated trays. A transient thermal
model of cables installed in trays and ladders was developed in
ABLES are installed in solid bottom trays in schools,
C hospitals and retail environments [1], [2]. In some cases
power generation plants and transmission and distribution sub-
[11]. It was assumed that the cable bundle is located in free
air. However, the calculations are limited to the cable bundle
only, i.e. the heat transfer problem is solved within cable mass
stations also utilize solid bottom trays to route cables [3]. Solid only. For convection and radiation calculations from the cables
bottom cable trays completely eliminate cable sagging and of- to ambient [11] points to [5], [6], [10].
fer protection against dirt, dust, and rodents. Solid bottom trays The thermal behavior of cables installed in solid bottom trays
(with covers) provide excellent EMI/RFI shielding protection with and without covers is reported in [3]. It follows the same
for very sensitive circuits against external electromagnetic fields procedures as [5] and [8] with a few modifications in the cal-
[4]. However, there are neither standards nor explicit industry culation of the effective thermal emissivity of the cable mass
guidelines on ampacity calculations for cables installed in solid and tray surface to account for the solid bottom. However, the
bottom trays. findings of [3] are not supported by experimental verification.
The pioneering work by Stolpe [5] introduced a simple It will be demonstrated in this paper that the method of [3] can
method for ampacity calculations of cables installed in ven- overestimate the ampacity by up to 40%.
tilated trays, which is the base of the current standards. Stolpe’s The National Electric Code (NEC) [12] introduces guidelines
model assumes that the cables in a tray form a composite cable on ampacities for cables installed in ventilated trays with and
mass (combination of conductive and insulating layers) of uni- without covers. In the case of solid bottom trays NEC points to
form depth (depth of fill) spread across the full width of the tray. Section 310.15(C). This section states that engineering supervi-
The model also assumes that all cables in the tray are power sion is necessary to compute ampacities, but no information is
cables that uniformly generate heat throughout the tray. There- provided on how to evaluate the effective thermal resistance be-
fore, the cable mass consisting of various types of cables can be tween the conductors and surroundings. Among the most pop-
ular cable ampacity programs used in the industry, including
Manuscript received April 7, 2016; revised July 8, 2016; accepted August CYMCAP and USAmp+, only ETAP and AmpCalc advertise
17, 2016. Date of publication August 24, 2016; date of current version May 19, calculations for cables installed in trays. The calculations are
2017. Paper no. TPWRD-00428-2016.
The authors are with the Department of Electrical and Computer Engineering,
based on [10] and therefore only applicable to ventilated trays.
New York University, Brooklyn, NY 11201 USA (e-mail: ss4993@nyu.edu; To close the gap with the standards and industry guidance and
ab5163@nyu.edu; fdeleon@nyu.edu). practice, a methodological thermal-electrical model for ampac-
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
ities of cables installed in solid bottom trays is developed in this
Digital Object Identifier 10.1109/TPWRD.2016.2602184 paper. The model uses fewer assumptions than [5]–[10] and is

0885-8977 © 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications standards/publications/rights/index.html for more information.
SAADAT et al.: THERMAL ANALYSIS OF POWER CABLES INSTALLED IN SOLID BOTTOM TRAYS USING AN EQUIVALENT CIRCUIT 2131

Fig. 1. Equivalent representation of cables installed in trays.

based on analytically derived heat transfer equations. Different

from [3], [5]–[8] and [10], the model of this paper considers
two-dimensional heat flow within the cable pack. Moreover,
all previous studies combine the thermal emissivities of cable Fig. 2. Proposed model for solid bottom trays with cover.
pack and tray into one effective thermal emissivity. The method
proposed here treats them separately producing more accurate
results. Unlike [3], [5], [7]–[10], the model proposed in this
paper, accounts for the fact that the top and bottom surfaces
of the cable bundle are at different temperatures. All studies
based on [5] treat the cable mass as a homogeneous rectangu-
lar (infinite) slab with uniform heat generation resulting in a
one-dimensional heat transfer problem whereas in the model
proposed in this paper the generated heat is concentrated in an
equivalent rectangular conductor (as shown in Fig. 1) yielding
Fig. 3. Proposed model for solid bottom trays without cover.
a more accurate two-dimensional heat transfer problem.
The proposed technique is validated with experimental mea-
surements using a 15 kV aluminum cable and Finite Elements
(FEM) simulations. The calculated ampacities show 4% to 5%
differences (conservative) with respect to laboratory tests. A
comparative study is made between the proposed method and [3]
using the 5 kV copper cable from [10]. The validity range of the
model is established with numerous FEM simulations using ca-
bles from 600 V to 35 kV. The effect of the tray surface emissiv-
ity on the cable rating is studied. A step-by-step example is pre-
sented in the Appendix to illustrate the simplicity of the method.
Fig. 4. Equivalent thermal-electrical circuit for cables installed in solid bottom
II. THERMAL-ELECTRICAL CIRCUIT tray with cover.

Computation of the ampacity of cables installed in trays is

a complex and tedious task because of the large number of demonstrate that (d) and (e) are also reasonable and present
cables, various loading scenarios, and cable sizes. Therefore, good agreement with experiments.
the following assumptions are made in the development of the With the above assumptions one can represent cables installed
circuits and models: in trays with the equivalent model given in Fig. 1. In fact this
a) Cables are installed at a uniform depth. feature of the model allows computing the ampacity of any num-
b) Conductors are assumed to be at the same temperature. ber of cables, i.e. the model is not limited to integer fill depths.
c) The fill depth of the cable pack is computed as in [10]: Detailed analytical models for cables installed in solid bottom
n1 d21 + n2 d22 + · · · + nn d2n trays with and without covers are presented in Figs. 2 and 3,
Calculated Depth = (1) respectively. The corresponding analog thermal-electrical cir-
width of tray
cuits are given in Figs. 4 and 5. Note that in the circuit of Fig. 4
where: there are additional resistors in the upper branch to account for
d1 , d2 , . . . , dn : Diameter of cables (in) the cover and its effects on convection and radiation. The rect-
n1 , n2 , . . . , nn : Cable number with respective diame- angular channel that represents the cable insulation and jacket
ters of d1 , d2 , . . . , dn . has a uniform thickness. As it can be seen from Figs. 2 and 3
d) Conductors are combined to form a rectangular equiva- the presence of a tray cover results in convection and radiation
lent conductor with cross-sectional area equal to the total inside the tray while in the case of open tray the cable pack is
cross-sectional area of all conductors in the tray. directly exposed to the ambient.
e) Temperatures of bottom and side of the tray are equal. One can note that the circuit of Fig. 4 is more difficult to
f) Tray has negligible thermal resistance in steady state. solve than the one of Fig. 5 because the cover introduces more
g) Convection and radiation inside the tray take place be- variables. The variables used in the thermal-electrical circuits
tween the top surface of the cable pack and the inner are:
surface of the tray cover. Q: Total power losses (W/m)
All assumptions except (d) and (e) were initially made in [5] Tcon d : Cable conductor temperature (°K)
and validated in the laboratory. Results obtained in Section VII Tsu r f : Top surface temperature of cable mass (°K)
2132 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 32, NO. 4, AUGUST 2017

Fig. 6. Two-dimensional heat flow in a square channel [13] and partitioned

cable mass for conduction shape factor calculation.

where:
Fig. 5. Equivalent thermal-electrical circuit for cables installed in solid bottom Rcon d Thermal resistance for conductive heat transfer
tray without cover.
(°K/W)
S Conduction shape factor (m)
Tcov er : Temperature of tray cover (°K) k Thermal conductivity (W/m·°K)
Tb&s : Temperature of tray bottom and side surfaces The conduction shape factor depends on the geometry of
(°K) the solid body. Yovanovich [14] presents a general expression
Tam b : Ambient temperature (°K) for computing the conduction shape factor. Conduction shape
Rca − t : Equivalent thermal resistance between the rect- factors for various geometries can be found in [13]–[17].
angular conductor and the top or bottom surface Analytical formulae for the calculation of the conduction
of the cable mass (°K/W) shape factors for a square channel (the channel can be a solid
Rca − s : Equivalent thermal resistance between the rect- body) shown in Fig. 6 are [13]:
angular conductor and the side surface of the
W 2πL
cable mass (°K/W) < 1.4 → S=   (3)
Rcon v : Equivalent thermal resistance for convection in- w 0.785 ln W
w
side the covered tray (°K/W) W 2πL
Rr ad : Equivalent thermal resistance for radiation inside > 1.4 → S= W  (4)
w 0.930 ln w − 0.050
the covered tray (°K/W)
Rtr ay : Conductive thermal resistance of tray (°K/W) For W/w = 1.4, (3) and (4) give almost the same values. As
Rcon v − B A : Equivalent thermal resistance for convection it can be seen from Figs. 2 and 3, the cable pack composed
over the bottom of tray (°K/W) of insulation and jacket forms a rectangular shell. This shell is
Rr ad − B A : Equivalent thermal resistance for radiation over partitioned into four sections as illustrated in Fig. 6. Conduction
the bottom of tray (°K/W) shape factors in (3) and (4) are for a complete shape, therefore,
Rcon v − C A : Equivalent thermal resistance for convection as it is shown in [14]–[17] for a quarter of the rectangular shell,
over the tray cover (°K/W) the conduction shape factors (3) and (4) should be modified as
Rr ad − C A : Equivalent thermal resistance for radiation over follows:
the tray cover (°K/W) π 
Rcon v − S A : Equivalent thermal resistance for convection W 2 L 
< 1.4 → S = (5)
over the tray side (°K/W) w 0.785 ln W w
Rr ad − S A : Equivalent thermal resistance for radiation over π 
W L
the tray side (°K/W) > 1.4 → S = 2W  (6)
Rcon v − T A : Equivalent thermal resistance for convection w 0.930 ln w − 0.050
over the top of cable pack when there is no cover
The conduction shape factor is calculated for each section
(°K/W)
separately using (5) and (6). Finally conductive thermal resis-
Rr ad − T A : Equivalent thermal resistance for radiation over
tances between the conductor mass and top, bottom and sides
the top of the cable pack when there is no cover
of the cable pack utilizing (2) are computed.
(°K/W)

III. THERMODYNAMICS (HEAT TRANSFER) B. Convection

The heat transfer from the conductors to the surroundings Heat transfer by convection is characterized by a convective
takes place through conduction, convection, and radiation. heat coefficient as follows [13]:
kair · N u
A. Conduction hcon v = (7)
L
Heat transfer by conduction takes place within the cable mass. where:
The two-dimensional thermal resistance for conduction is com- hcon v Convective heat transfer coefficient (W/m2 ·°K)
puted using [13]: kair Thermal conductivity of air (W/m·°K)
1 Nu Nusselt number
Rcon d = (2) L: Characteristic length (m)
Sk
SAADAT et al.: THERMAL ANALYSIS OF POWER CABLES INSTALLED IN SOLID BOTTOM TRAYS USING AN EQUIVALENT CIRCUIT 2133

The equivalent thermal resistance for convection is computed 2) Convection at the Bottom and Cover of Tray: For hori-
using the following expression [13]: zontal surfaces the Nusselt number is defined as follows [13]:
1 N u = C · Ram (19)
Rcon v = (8)
hcon v · A For the upper surface: C = 0.54 and m = 0.25; for the lower
where: surface: C = 0.27 and m = 0.25.
Rcon v : Thermal resistance for convective heat transfer 3) Convection on the Sides of Tray: In case of vertical sur-
(°K/W) faces the Nusselt number is given by [13]:
A: Area of convecting surface (m2 ) 0.670Ra1/4
1) Convection Inside the Tray: The Nusselt number depends N u = 0.68 +  4/9 (20)
on the geometry of the convective system. Convection inside the 1 + (0.492/P r)9/16
tray is determined by the following equation [13]:
C. Radiation
N u = 0.069Ra1/3 P r0.074 (9)
Radiation takes place in the tray between the top surface of
Ra is Rayleigh number and P r is Prandtl number which have the cable pack and the tray cover and then from the tray outer
the following expressions [13]: surfaces to ambient.
1) Radiation Within the Tray: Since the tray is covered, ra-
gβ(T1 − T2 )L3 diation inside the tray happens between the top surface of cable
Ra = (10) mass and the inner surface of the cover. This exchange depends
αν
1 strongly on the surface geometry and orientation, as well as on
β= (11) the surface emissivity and temperature. The concept of view
Tf
factor is used to compute radiation exchange between any two
T1 + T2 surfaces [13].
Tf = (12)
2 The following view factor is used for a covered cable tray
k which yields a system of two rectangular surfaces of equal size
α= (13) in opposite location [6]:
ρ · Cp
εcable εtr ay
k = 1.5207 · 10−11 · Tf3 − 4.8574 · 10−8 · Tf2 F =
3.2 − εcable − εtr ay
(21)

+ 1.018 · 10−4 · TF − 3.9333 · 10−4 (14) Radiative heat transfer coefficient and thermal resistance are
p given by [6]:
ρ =
R · Tf
(15)  
hr ad = F σ T12 + T22 (T1 + T2 ) (22)
ν = μ/ρ (16) 1
Rr ad = (23)
C1 · Tf 3/2 hr ad · A
μ= (17) where:
Tf + C2
F Radiation view factor
μCp εcable Surface emissivity of cable pack
Pr = (18)
k εtr ay Surface emissivity of tray
where: hr ad Radiative heat transfer coefficient (W·m−2 ·°K−1 )
g Acceleration of gravity (g = 9.81 m/s2 ) σ Stephen-Boltzmann constant (σ = 5.67 × 10−8 W/m2 ·
◦
β Volumetric thermal expansion coefficient (1/°K) K)
Tf Film temperature (°K) T1 Temperature of emitting surface (°K)
T1 Temperature of hot surface (°K) T2 Temperature of absorbing surface (°K)
T2 Temperature of cool surface (°K) Rr ad Thermal resistance for radiative heat transfer (°K/W)
α Thermal diffusivity at Tf (m2 /s) A Surface area (m2 )
ν Kinematic viscosity at Tf (m2 /s) 2) Radiation on Outer Surfaces of Tray: Radiative processes
μ Dynamic viscosity at Tf (kg/m·s) that occur at the surface are described by [13]:
 
ρ Density of air at Tf (kg/m3 ) hr ad = εtr ay σ T12 + T22 (T1 + T2 ) (24)
p Atmospheric pressure (p = 101 325 Pa)
R Gas constant for dry air (R = 287.058 J/ kg · ◦ K) 1
Rr ad = (25)
Cp Specific heat capacity at constant pressure (Cp = hr ad · A
1006 J/ kg · ◦ K)
IV. SOLVING THE THERMAL–ELECTRICAL CIRCUIT
C1 Sutherland viscosity law coefficient (C1
= 1.458 · 10−6 kg/m · s · ◦ K1/2 ) The proposed model is based on the equivalent thermal-
C2 Sutherland viscosity law coefficient (C2 = 110.4 ◦ K) electrical circuits of Figs. 4 and 5.
2134 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 32, NO. 4, AUGUST 2017

A. Cables Installed in Solid Bottom Trays With Covers The objective of these iterations is to determine the total heat
generated Q in Watts per unit of length.
The following system of equations can be derived from the
circuit for solid bottom trays with covers applying KVL: Q = Qt + Qb + Qs (36)
Tcon d = Tsu r f + Qt · R1 (26) Applying Joule’s law, the ampacity of the cable can be com-
puted from:
Tsu r f = Tcov er + Qt · R2 (27) 
Tcov er = Tam b + Qt · R3 (28) Q
I= (37)
nRac
Qb · R1 Qs · R5
Tcon d = Tb&s + + (29) where:
2 2 I Ampacity of cables (A)
Tb&s = Tam b + Qb · R4 (30) Q Heat generated in the cables (W/m)
Tb&s = Tam b + Qs · R6 (31) n Number of cables
Rac ac resistance of the conductor at the operating tempera-
where: ture (Ω/m)
Qt Heat flow in upper direction (W)
Qb Heat flow in lower direction (W) B. Cables Installed in Solid Bottom Trays Without Covers
Qs Heat flow to the sides (W)
The system of equations in the case of solid bottom trays
R1 Rca − t
without covers is:
R2 Rcon v ||Rr ad
R3 Rcon v − C A ||Rr ad − C A Tcon d = Tsu r f + Qt · R1 (38)
R4 Rcon v − B A ||Rr ad − B A
Tsu r f = Tam b + Qt · R7 (39)
R5 Rca − s /2
R6 (Rcon v − S A ||Rr ad − S A )/2 Qb · R1 Qs · R5
Tcon d = Tb&s + + (40)
The given system of equations is non-linear due to the nature 2 2
of the convection and radiation that depend on temperature. The
Tb&s = Tam b + Qb · R4 (41)
only known parameters of the system of equations are the con-
ductor and ambient temperatures. Note that thermal resistances Tb&s = Tam b + Qs · R6 (42)
for conduction are constant while thermal resistances for con- where:
vection and radiation are temperature-dependent and computed R7 Rcon v − T A ||Rr ad − T A
accordingly. To solve such system of equations iterative meth- The solution of the system of equations (38)–(42) follows the
ods are required. The successive relaxation method is utilized as same procedure described in the previous section except that
in [18]. The relaxation parameter λ = 0.5 is used to obtain the there are fewer unknowns and equations.
best results. The iterative method is described in the following
steps: V. LABORATORY TESTS
1) Start with reasonable initial guesses for Tcov er (0) and
Tb&s (0), which are used to evaluate Qt (0), Qb (0) and A single core 350 kcmil, aluminum conductor, 15 kV EPR
Qs (0) from (28), (30) and (31), respectively insulated, PVC jacket, copper tape shield cable was used in
2) Compute Tsu r f (0) using (26) the laboratory experiments. The cable has a measured outside
3) Determine Qt (1) solving (27) diameter of 36.2 mm (1.425 in). The cable tray used in the
4) Use Qt (1) to calculate new values for Tcov er (1) and experiments is made of galvanized steel, 3.048 m (10 ft) long,
Tb&s (1) from linear combinations of the previous value 0.6096 m (24 in) wide, 0.1016 m (4 in) height and thickness of
and the one computed with (28) and (29) as follows: 1.5 mm.
Tcov er (1) = (1 − λ) Tcov er (0) + λ (Tam b + Qt (1) · R3 ) A. One Inch Fill Depth
(32) The first experiment was conducted using twelve cables cor-
Tb&s (1) = (1 − λ) Tb&s (0) responding to a fill depth of slightly larger than one inch as per
  ICEA/NEMA; see (1). The cables were connected in series and
Qb (0) · R1 Qs (0) · R5 looped back and forth so that they can be fed from one source.
+ λ Tcon d − −
2 2 The total length of the cables is 120 ft.
(33) The thermal behavior of the conductors and jackets were
measured with thermocouples and monitored from a computer.
5. Find the errors:
The locations of the thermocouples (TN ) are the same for both
Ecov er (1) = Tcov er (1) − Tcov er (0) (34) open and covered tray tests (see Fig. 7). The experimental setup
is shown in Fig. 8(a). For illustration purposes the cover is re-
Eb&s (1) = Tb&s (1) − Tb&s (0) (35)
moved. The ambient air temperature was measured as 23 °C and
6. If the errors are greater than 0.5 °C repeat the steps. 24 °C for open and covered trays, respectively. The experiments
SAADAT et al.: THERMAL ANALYSIS OF POWER CABLES INSTALLED IN SOLID BOTTOM TRAYS USING AN EQUIVALENT CIRCUIT 2135

TABLE II
CURRENT & TEMPERATURE COMPARISON FOR TWO INCH FILL DEPTH

Tray Current (A) HST Rise (°C)

Fig. 7. Locations of thermocouples for the one-inch depth experiment. Note Installation
that there are two thermocouples per location, one touching the jacket and one
touching the conductor. Experiment Proposed Difference Experiment & FEM Difference
& FEM Model (%) Proposed Model (%)

Open 230 220 4.3 60 63 5.0

Covered 171 155 9.0 56 59 5.4

Fig. 10. Steady state temperature distribution for one inch fill depth with
cover.

Fig. 8. Experimental setup for one inch (a) and two inch (b) tests.

TABLE I
CURRENT & HST COMPARISON FOR ONE INCH FILL DEPTH

Tray Current (A) HST Rise (°C)

Installation

Experiment Proposed Difference Experiment & FEM Difference

& FEM Model (%) Proposed Model (%)
Fig. 11. Steady state temperature distribution for one inch fill depth without
Open 371 390 5.1 72 76 5.6 cover.
Covered 296 292 1.4 99 97 2.0

Fig. 9. Locations of thermocouples for the two-inch experiment. Note that

there are two thermocouples per location, one touching the jacket and one
touching the conductor.
Fig. 12. Steady state temperature distribution for two inch fill depth with
cover.

ran for 15.5 (open tray) and 17.5 hours (covered tray) until the
temperature reached steady-state. The results are presented in
VI. FEM MODEL
Table I.
COMSOL Multiphysics, Finite Elements Method (FEM)
software is used to evaluate the experimental results in terms
B. Two Inch Fill Depth
of final temperatures. The software solves the complete com-
Twenty four cables were tightly packed with a fill depth of putational fluid dynamics, nonlinear electromagnetic, and heat
2.03 inches as depicted in Fig. 8(b). Again for illustration pur- transfer equations simultaneously by coupling Heat Transfer,
poses the cover is removed. Thermocouple locations are shown Fluid Flow, and Electromagnetic multi-physics packages.
in Fig. 9. The ambient air temperature was recorded as 23 °C Figs. 10–13 provide the steady-state temperature distribution
for both open and covered trays. Eleven hours (open tray) and obtained from FEM simulations with and without covers for one
25 hours (covered tray) were required to make sure that the and two inch-fill depths, respectively. The purpose of the simu-
temperatures reached steady-state; see Table II. lations is to reproduce laboratory tests for one and two inch-fill
2136 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 32, NO. 4, AUGUST 2017

TABLE IV
ALLOWABLE HEAT INTENSITIES OF VENTILATED VS. SOLID BOTTOM TRAYS
WITH COVERS; REPRODUCED FROM [3]

Fill Depth Derating for Solid Additional Derating Total Derating

of Cables Bottom Tray to account for Solid Bottom
(in) without Covers for Cover Trays with Covers

1 0.942 0.72 0.68

2 0.951 0.74 0.70
3 0.957 0.77 0.74
Fig. 13. Steady state temperature distribution for two inch fill depth without
cover.
TABLE V
CALCULATED AMPACITIES FOR A SINGLE CORE 250 KCMIL, COPPER
TABLE III CONDUCTOR, 5 KV SHIELDED CABLE
ALLOWABLE HEAT INTENSITIES OF VENTILATED VS. SOLID BOTTOM TRAYS
WITHOUT COVERS; REPRODUCED FROM [3]
Fill Depth of Calculated Cables (in)
Cables (in) Ampacity (A)
Fill Depth of Allowable Heat Intensity (W/ft·in2 ) Derating
Cables (in) Factor Open Covered

Ventilated [3] Solid Bottom Using Proposed Difference Using Proposed Difference
[3] Model (%) [3] Model (%)
1 5.925 5.257 0.942
1.5 3.557 3.186 0.946 1 314 280 12.3 267 191 39.7
2 2.427 2.194 0.951 1.5 244 215 13.5 N/A 154 N/A
2.5 1.784 1.624 0.954 2 202 177 14.0 174 131 32.6
3 1.377 1.26 0.957 2.5 173 152 13.7 N/A 116 N/A
3 153 133 14.7 134 104 28.8

depths and validate the FEM model. The experiments have run 2) For FEM simulations the current applied in the experiment
until the temperature in the system reaches the steady-state. is used as the input data and the temperature is computed.
It has been observed that due to the large number of highly Therefore, currents of the laboratory tests and FEM sim-
non-linear equations solved by the software, simulation of ulations are the same and the temperatures show a small
11 to 25 hours of actual cable transient response, could take difference.
from several hours up to weeks using a server that has 24 cores All differences between calculated and measured parameters
in its central processing unit (CPU) running at 3.33 GHz each as are within engineering accuracy (2% to 9%). Note that the per-
well as 96 GB of DDR3 random-access memory (RAM). For this cent difference in the temperatures are calculated based on the
reason the advantage of the closed system in covered tray sim- temperature rise above the ambient temperature (24 °C for one
ulations was utilized. COMSOL Multiphysics has a feature that inch fill depth with cover, 23 °C for one inch fill depth with no
allows defining external natural convection and radiation on the cover and two inch fill depth with both cover and no cover). For
outer surfaces of the covered tray, thus dramatically decreasing instance, from Table I for one inch fill depth with no cover, the
the simulation time. This feature uses much simpler algebraic difference is computed between 72 °C (Experiment & Model)
equations rather than partial differential equations. Therefore, and 76 °C (FEM). A detailed example for two inch fill depth
although the region outside the tray with cover is not illustrated with cover is provided in the Appendix.
(see Figs. 10 and 12), it is physically present and represented
by natural convection and radiation on the boundaries of the VIII. CALCULATED AMPACITIES AND EVALUATION OF [3]
system. The Tables I to II show a good match of the hot spot
temperatures (HST’s) between laboratory experiments and FEM As it was mentioned in the Introduction, the study of [3]
simulations. FEM simulations yield a maximum difference of suggests that the ampacities provided in [10] need to be re-
5.6% when compared to experiments. duced when solid bottom trays are utilized. The derating fac-
tors are given in terms of allowable heat generation q (see
Tables III and IV). The ampacity is calculated using the
VII. EVALUATION OF THE PROPOSED MODEL following expression [10]:
In this section the evaluation of the proposed model and val-
I =d× q/(nRac ) (43)
idation of the thermal-electrical circuit are carried out. The re-
sults given in Tables I and II are obtained in the following way: where:
1) The HST measured in the experiment is entered as data I: Ampacity of a cable (A)
in the model and the current is computed. Therefore, the d: Cable diameter (in)
HSTs of experiments and model are the same and the q: Allowable heat generation (W/ft·in2 )
currents are slightly different (by a few percent). n: Number of cables
SAADAT et al.: THERMAL ANALYSIS OF POWER CABLES INSTALLED IN SOLID BOTTOM TRAYS USING AN EQUIVALENT CIRCUIT 2137

TABLE VI TABLE IX
VALIDITY RANGE OF THE MODEL FOR 600 V CABLES VALIDITY RANGE OF THE MODEL FOR 15 KV CABLES

1/0 AWG 500 kcmil 2 AWG 500 kcmil

Fill Depth of HST Current (A) HST Current (A) Fill Depth of HST Current (A) HST Current (A)
Cables (in) (°C) (°C) Cables (in) (°C) (°C)

FEM Model Diff. (%) FEM Model Diff. (%) FEM Model Diff. (%) FEM Model Diff. (%)

1 83 65 60 7.7 103 308 303 1.6 1 82 69 64 7.2 94 332 344 3.6

2 – – – – 102 209 208 0.5 2 – – – – 90 228 228 0.0
3 – – – – 96 150 158 5.3 3 – – – – 90 181 181 0.0

TABLE VII TABLE X

VALIDITY RANGE OF THE MODEL FOR 2 KV CABLES VALIDITY RANGE OF THE MODEL FOR 25 KV AND 35 KV CABLES

1/0 AWG 500 kcmil 1/0 AWG 350 kcmil

Fill Depth of HST Current (A) HST Current (A) Fill Depth of HST Current (A) HST Current (A)
Cables (in) (°C) (°C) Cables (in) (°C) (°C)

FEM Model Diff. (%) FEM Model Diff. (%) FEM Model Diff. (%) FEM Model Diff. (%)

1 82 66 61 7.6 90 277 277 0.0 1 84 128 119 7.0 98 302 322 6.6
2 – – – – 88 192 187 2.6 2 – – – – 82 206 191 7.3
3 – – – – 90 153 153 0.0 3 – – – – 84 162 155 4.3

TABLE VIII
VALIDITY RANGE OF THE MODEL FOR 5 KV CABLES be multiplied by the square root of the corresponding derating
factor from Tables III and IV to account for the solid bottom
2 AWG 500 kcmil of the tray. Conductor operating and ambient temperatures are
Fill Depth of HST Current (A) HST Current (A) 90 °C and 40 °C, respectively.
Cables (in) (°C) (°C) The differences between the two methods in Table V were
computed with respect to the ampacity determined using the
FEM Model Diff. (%) FEM Model Diff. (%)
proposed model. One can note that in the open tray installations
1 81 55 51 7.3 92 308 314 1.9 the differences are from 12% to 15% for this particular cable.
2 – – – – 91 212 214 0.9
3 – – – – 90 169 169 0.0
For covered tray installations the differences grows larger from
28% to 40%. One must conclude that the method of reference
[3] overestimates the ampacity excessively.

Rac : ac resistance of the conductor at the operating temper-

ature (μΩ/ft) IX. VALIDITY RANGE OF THE PROPOSED MODEL
Note that the allowable heat generation q is given in Watts per This section obtains the validity range of the model. Since
unit volume while the heat generated by the cables Q in (30) is solid bottom trays are utilized in low and medium voltage ap-
given in Watts per unit length. This difference is reflected in the plications, cables from 600 V to 35 kV are evaluated. The study
ampacity formulae (30) and (36); the ampacity in (36) is directly is carried out on the cables given in [10] by comparing FEM
proportional to the cable diameter while the ampacity in (30) is simulations and the model in terms of the calculated current, i.e.
related to the cable diameter through the heat generated Q in the final temperatures of the two methods are the same. Because
the system. [10] provides the ampacity of cables up to 15 kV only, cables
To obtain ampacity tables for cables installed in solid bottom from the General Cable catalogue rated 25 kV and 35 kV are
trays based on the proposed model and assess the results given also selected [19].
in [3] a comparison between the two methods is presented in It is worth mentioning that FEM simulations take a long
Table V. Due to space limitations, Table V is based on only one time to reach the steady state due to the complexity of natural
cable type. The single core 250 kcmil, copper conductor, 5 kV, convection of free air. For this reason, for small size cables
shielded power cable from [10], Tables 5–18 is used. The cable (lower limits) FEM simulations are performed for one inch of
has an outside diameter of 25.4 mm (1.0 in), as an example cables while for large size cables (upper limits) the model and
twenty four of these cables are used to fill one inch fill depth. FEM simulations are compared for one, two and three inch
The derating factors in Tables III and IV apply to the allowable fill depths. Additionally, the evaluation is done on covered tray
heat intensity. Therefore, the ampacity values in [10] need to installations only. The ambient temperature is 40 °C as in [10].
2138 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 32, NO. 4, AUGUST 2017

TABLE XI TABLE XIII

EFFECT OF TRAY SURFACE EMISSIVITY ON RATING OF CABLES IN SOLID INPUT & CALCULATED DATA
BOTTOM TRAYS WITH COVERS
Parameter Description Value
Fill Depth of Calculated Ampacity (A)
Cables (in) λ Relaxation parameter 0.5
w Width of tray 0.6096
Surface Emissivity of Tray ht Depth of tray 0.1016
dc o n d Conductor diameter 0.0156
= 0.23 = 0.88 Ampacity Increase (%) di n s Insulation diameter 0.0304
dc a b l e Overall cable diameter 0.0362
1 166 217 30.7 L Length of cable/tray 3.048
2 115 145 26.1 Ah Surface area of the tray cover and bottom 1.8581
3 91 113 24.2 Av Surface are of the tray sides 0.3097
4 78 93 19.2 pd Characteristic length for convection over 0.2540
tray cover and bottom (area/perimeter)
th Thickness of rectangular shell that 0.0217
represents insulation and jacket
TABLE XII w1 Width of equivalent rectangular conductor 0.5662
EFFECT OF TRAY SURFACE EMISSIVITY ON RATING OF CABLES IN SOLID w2 Height of equivalent rectangular conductor 0.0082
BOTTOM TRAYS WITHOUT COVERS R n d c (Ω/m) Nominal dc resistance of the conductor at 1.624·10–4
20 °C
R d c (Ω/m) dc resistance of the conductor with 2% of 1.65648·10–4
Fill Depth of Calculated Ampacity (A) stranding factor at 20 °C
Cables (in) R a c (Ω) ac resistance of the conductor 3.048 meter 6.2814·10–4
long at the operating temperature
Surface Emissivity of Tray ki n s Thermal conductivity of EPR insulation 0.2
kj a c k Thermal conductivity of PVC jacket 0.2
ε = 0.23 ε = 0.88 Ampacity Increase (%) εp v c Thermal emissivity of PVC jacket 0.9
εt r a y Thermal emissivity of galvanized steel 0.2
1 256 271 5.9 tray (new)
2 159 169 6.3 F View factor for radiation inside tray; 0.0857
3 119 128 7.6 see (21)
4 96 103 7.3 Tc o r e Conductor temperature 79.0
Ta m b Ambient temperature 23.0
Tc o v e r Initial value for temperature of tray cover 28.0
Tb & s Initial temperature of tray sides and bottom 28.0
Tables VI–X present the results for the lower and upper limits
of the validity range of the model.
The conclusions of this study are: (1) for 600 V and 2 kV
and without covers. The technique is completely general and
cables the method is valid from 1/0 AWG to 500 kcmil; (2) for
allows for the calculation of ampacities of any number of ca-
5 kV and 15 kV cables the method works well from 2 AWG to
bles, i.e. the model is not limited to integer fill depths. The
500 kcmil; (2) for 25 kV and 35 kV cables the model is valid
accuracy of the model has been verified with experimental re-
from 1/0 AWG to 350 kcmil.
sults and FEM simulations. The validity range of the model has
been established and the effect of tray surface emissivity on the
X. EFFECT OF TRAY SURFACE EMISSIVITY ON RATING OF cable ampacity has been analyzed.
CABLES IN SOLID BOTTOM TRAYS Since there is little published data on the subject the results of
New galvanized steel trays tend to be shiny and have low this research can be used as a basis for establishing standard or
surface emissivity ( new = 0.23). On the other hand, old gal- guidelines for the calculation of ampacities of cables installed
vanized steel trays are matte and have high surface emissivity in solid bottom trays with and without covers.
( old = 0.88) [20]. To analyze the effects of the tray surface
emissivity on the cable ratings, the cable studied in Section VIII APPENDIX
is used to obtain the values given in Tables XI and XII for trays
with and without covers. The ampacity percent increase is com- Illustration of the proposed model is provided through a nu-
puted between the two cases, i.e. new = 0.23 and old = 0.88 merical example based on two inch fill depth with cover. Ta-
for a given fill depth of the cables. ble XIII provides input and calculated data. Table XIV presents
The results clearly show that there are important advantages the solution of the system of equations (26)–(31) with parame-
of using trays with high surface emissivity. One can see that for ters calculated at each iteration. The initial values are ambient
covered installations having the tray painted in a matte color temperature plus 5 °C. All parameters are given in SI units ex-
can increase the ampacity from 6% for trays with no cover and cept the temperature which is presented in °C for convenience.
as much as 30% for trays with cover. Nominal dc resistance and stranding factor of 8000 series alu-
minum compact conductor are in accordance with ASTM Specs
B800 and B801 and ICEA Part 2, Sections 2.1 and 2.5 as indi-
XI. CONCLUSION
cated in the cable specs sheet. The ac resistance of the conductor
A methodological model has been proposed for the calcula- is computed as per IEC 60287-1-1 Sections 2.1.2 and 2.1.4.1
tion of ampacity of cables installed in solid bottom trays with and adjusted to the operating temperature.
SAADAT et al.: THERMAL ANALYSIS OF POWER CABLES INSTALLED IN SOLID BOTTOM TRAYS USING AN EQUIVALENT CIRCUIT 2139

TABLE XIV TABLE XIV: (CONTINUED)

SOLUTION OF TWO INCH FILL DEPTH WITH COVER
Rc o n v (8) 0.1943 0.2700 ... 0.2319 0.2319
Parameter Equation Iteration Number hr a d (22) 0.7289 0.7452 ... 0.7406 0.7406
Rr a d (23) 0.7384 0.7222 ... 0.7267 0.7267
0 1 ... 9 10 R2 (8)||(23) 0.1539 0.1965 ... 0.1758 0.1758
Qt (27) 239.2 154.7 ... 165.2 165.2
Step 1 Step 4
S (5) 2.8740 2.8740 ... 2.8740 2.8740
Tc o v e r (25) 28.0 46.0 ... 39.9 39.9 R5 (2)/2 0.8699 0.8699 ... 0.8699 0.8699
Tb & s (26) 28.0 49.8 ... 41.9 42.5 Tc o v e r (32) 46.0 39.0 ... 39.4 39.4
Tf − C A (12) 25.5 34.5 ... 31.5 31.5 Tb & s (33) 49.8 37.1 ... 42.5 42.1
βC A (11) 0.0033 0.0033 ... 0.0033 0.0033 Step 5
ρC A (15) 1.1819 1.1473 ... 1.1587 1.1587 |E c o v e r | (34) 100 18.0 ... 0.0 0.0
kC A (14) 0.0261 0.0268 ... 0.0266 0.0266 |E b & s | (35) 100 21.8 ... 0.6 0.4
–6
μC A (17) 18.4·10 18.8·10–6 ... 18.7·10–6 18.7·10–6 Total generated heat & current
νC A (16) 15.6·10–6 16.4·10–6 ... 16.1·10–6 16.1·10–6 Q (36) 279.3 445.6 ... 356.7 364.0
αC A (13) 22.0·10–6 23.2·10–6 ... 22.8·10–6 22.8·10–6 I (37) 136 172 ... 154 155
RaC A (10) 0.8·107 3.1·107 ... 2.4·107 2.4·107
N uC A (19) 28.6 40.5 ... 37.9 37.9
hc o n v − C A (7) 2.9391 4.2694 ... 3.9655 3.9655
REFERENCES
Rc o n v − C A (8) 0.1831 0.1261 ... 0.1357 0.1357
hr a d − C A (24) 1.2083 1.3228 ... 1.2832 1.2832 [1] Mono-Systems. 2016. Solid bottom cable tray system. [Online].
Rr a d − C A (25) 0.4454 0.4069 ... 0.4194 0.4194 Available: http://www.monosystems.com/products/item/trough-tray/129/
R3 (8)||(25) 0.1298 0.0962 ... 0.1025 0.1025 10017
Qt (28) 38.5 239.2 ... 165.2 165.2 [2] SpecMate Wireway Systems. SPM series—solid bottom cable
Tf − B A (12) 25.5 36.4 ... 32.5 32.8 tray system, Feb. 2005. [Online]. Available: http://www.cs.unc.edu/
βB A (11) 0.0033 0.0032 ... 0.0033 0.0033 sittsouth/network/product_cut_sheets/ed579
ρB A (15) 1.1819 1.1403 ... 1.1550 1.1539 [3] R. V. Rebbapragada, D. O’Neill, M. Zuzovsky, and R. Weronick, “Am-
kB A (14) 0.0261 0.0269 ... 0.0266 0.0266 pacity of cables in solid bottom trays with and without covers,” in Proc.
μB A (17) 18.4·10–6 18.9·10–6 ... 18.7·10–6 18.7·10–6 Conf. Rec. IEEE Nuclear Sci. Symp., 1999, vol. 3, pp. 1704–1707.
–6
νB A (16) 15.7·10 16.6·10–6 ... 16.2·10–6 16.2·10–6 [4] Cable Tray Institute. 2016. Cable tray type selection. [Online].
αB A (13) 22.0·10–6 23.5·10–6 ... 22.9·10–6 23.0·10–6 Available: http://www.cabletrays.org/pdfs/Cabletrays-Institute-Techni-
RaB A (10) 0.8·107 3.6·107 ... 2.7·107 2.8·107 cal-Bulletin4.pdf
N uB A (19) 14.3 20.9 ... 19.4 19.6 [5] J. Stolpe, “Ampacities for cables in randomly filled cable trays,” IEEE
hc o n v − B A (7) 1.4696 2.2139 ... 2.0361 2.0515 Trans. Power App. Syst., vol. PAS-90, no. 3, pp. 962–974, May 1972.
Rc o n v − B A (8) 0.3662 0.2431 ... 0.2643 0.2623 [6] G. Engmann, “Ampacity of cable in covered tray,” IEEE Trans. Power
hr a d − B A (24) 1.2083 1.3481 ... 1.2959 1.2998 App. Syst., vol. PAS-103, no. 2, pp. 962–974, Feb. 1984.
Rr a d − B A (25) 0.4454 0.3992 ... 0.4153 0.4141 [7] B. L. Harshe and W. Z. Black, “Ampacity of cables in single open-top cable
R4 (8)||(25) 0.2010 0.1511 ... 0.1615 0.1606 trays,” IEEE Trans. Power Del., vol. 9, no. 4, pp. 1733–1740, Oct. 1994.
Qb (30) 24.9 177.5 ... 117.1 121.5 [8] B. L. Harshe and W. Z. Black, “Ampacity of cables in single covered
Tf − S A (12) 25.5 36.4 ... 32.5 32.8 trays,” IEEE Trans. Power Del., vol. 12, no. 1, pp. 3–14, Jan. 1997.
βS A . (11) 0.0033 0.0032 ... 0.0033 0.0033 [9] W. Z. Black and B. L. Harshe, “Ampacity of diversely loaded cables in
ρS A (15) 1819 1.1403 ... 1.1550 1.1539 covered and uncovered trays,” IEEE Trans. Power Del., vol. 15, no. 1,
kS A (14) 0.0261 0.0269 ... 0.0266 0.0266 pp. 3–7, Jan. 2000.
μS A (17) 18.4·10–6 18.9·10–6 ... 18.7·10–6 18.7·10–6 [10] ICEA/NEMA ampacities of cables installed in cable trays, ICEA Publ.
νS A (16) 15.7·10–6 16.6·10–6 ... 16.2·10–6 16.2·10–6 No. P-54-440; NEMA publ. no. WC51, NEMA, Washington, DC, USA,
αS A (13) 22.0·10–6 23.5·10–6 ... 22.9·10–6 23.0·10–6 1986.
RaS A (10) 0.5·106 2.3·107 ... 1.7·107 1.8·107 [11] S.-B. Liu and X. Han, “Analytical method of calculating the transient and
P rS A (18) 0.7092 0.7064 ... 0.7074 0.7073 steady-state temperature rises for cable-bundle in tray and ladder,” IEEE
N uS A (20) 14.4 20.7 ... 19.3 19.4 Trans. Power Del., vol. 13, no. 3, pp. 691–698, Jul. 1998.
hc o n v − S A (7) 3.6937 5.4793 ... 5.0525 5.0895 [12] NFPA-70, National Electric Code, 2014.
Rc o n v − S A (8) 0.8742 0.5893 ... 0.6391 0.6345 [13] F. P. Incropera and D. P. DeWitt, Introduction to Heat Transfer, 3rd ed.
hr a d − S A (24) 1.2083 1.3481 ... 1.2959 1.2998 Hoboken, NJ, USA: Wiley, 1996, p. 698.
Rr a d − S A (25) 2.6724 2.3953 ... 2.4918 2.4844 [14] M. M. Yovanovich, “A general expression for predicting conduction shape
R6 ((8)||(25))/2 0.3294 0.2365 ... 0.2543 0.2527 factors,” in Proc. AIAA 11th Aerosp. Sci. Meeting, Jan. 1973, pp. 266–291.
Qs (31) 15.2 113.4 ... 74.4 77.2 [15] S. S. Kutateladze and V. M. Borishanskii, A Concise Encyclopedia of Heat
Step 2 Transfer. Cambridge, MA, USA: Pergamon Press, 1966.
S (5) 82.5271 82.5271 ... 82.5271 82.5271 [16] W. M. Rohsenow, J. P. Harnett, and Y. I. Cho, Handbook of Heat Transfer,
R1 (2) 0.0606 0.0606 ... 0.0606 0.0606 3rd ed. New York, NY, USA: McGraw-Hill, 1998.
Ts u r f (26) 76.7 64.5 ... 69.0 69.0 [17] J. H. Lienhard IV and J. H. Lienhard V, A Heat Transfer Textbook, 3rd ed.
Step 3 Cambridge, MA, USA: Phlogiston Press, 2008.
Tf (12) 52.3 55.3 ... 54.5 54.5 [18] M. Terracciano, S. Purushotaman, F. de León, and A. R. Farahani, “Ther-
β (11) 0.0031 0.0030 ... 0.0031 0.0031 mal analysis of cables in unfilled troughs: Investigation of the IEC standard
ρ (15) 1.0845 1.0748 ... 1.0774 1.0774 and a methodological approach for cable thermal rating,” IEEE Trans.
k (14) 0.0281 0.0284 ... 0.0283 0.0283 Power Del., vol. 27, no. 3, pp. 1423–1431, Jul. 2012.
μ (17) 19.6·10–6 19.8·10–6 ... 19.7·10–6 19.7·10–6 [19] General Cable. Apr. 2015. Spec. [Online]. Available: http://
ν (16) 18.1·10–6 18.4·10–6 ... 18.3·10–6 18.3·10–6 www.generalcable.com/na/us-can/products-solutions/industrial/medium-
α (13) 25.8·10–6 26.2·10–6 ... 26.1·10–6 26.1·10–6 voltage-cable/copper-tape-shielded-type-mv-cable
Ra (10) 3.9·105 1.4·105 ... 2.3·105 2.3·105 [20] The Engineering ToolBox. 2016. Emissivity coefficients of some com-
Pr (18) 0.7024 0.7017 ... 0.7019 0.7019 mon materials. [Online]. Available: http://www.engineeringtoolbox.com/
Nu (9) 4.9241 3.5168 ... 4.1041 4.1041 emissivi ty-coefficients-d_447.html
hc o n v (7) 2.7693 1.9933 ... 2.3213 2.3213

Authors,’ photographs and biographies not available at the time of publication.