Case Studies in Thermal Engineering 28 (2021) 101338

Contents lists available at ScienceDirect

Case Studies in Thermal Engineering

journal homepage: www.elsevier.com/locate/csite

Experiment and numerical simulation of cable trench

fire detection
Xiaosheng Liu a, b, Dong Hou a, b, Jun Ji a, b, Hui Zhu c, *
a
State Key Laboratory of Advanced Power Transmission Technology, Global Energy Inter-connection Research Institute Co.Ltd., Beijing, 102201,
China
b
Global Energy Interconnection Research Institute Co.Ltd, Future City for Science and Technology, Changping District, Beijing, 102201, China
c
Sichuan Fire Research Institute of MEM, Chengdu, Sichuan, 610036, PR China

A R T I C L E I N F O A B S T R A C T

Keywords: In this paper, the full-scale cable fire test is carried out in the fire compartment of cable trench in
Cable trench fire standard substation in order to explore the influence of fire detector sensitivity. The point-type
Numerical simulation temperature detector, point-type smoke detector, cable-type temperature detector, and suction-
Fire detector
type smoke detector are used to test the fire alarm sequence. The temperature field, combus­
FDS
tible gas concentration, and fire alarm sequence of fire detector with different sensitivity are
simulated by FDS. The results show that: The numerical simulation results are in good agreement
with the experimental results. The lower the alarm threshold, the shorter the response time of fire
detector. The sensitivity of suction-type smoke detector has great influences on the detection
results.

1. Introduction

With the rapid development of urbanization, the pace of distribution network construction is accelerated, and the number of
substations is increasing significantly. As an important part of the normal operation of power grid system, cable is an important carrier
of power transmission and signal transmission in substation, which is almost all over the substation. The rapid development of power
grid system also leads to frequent cable fire accidents. For example, the explosion accident occurred in Xi’an substation on June 18 in
China. It not only caused huge economic losses to the country, but also affected the growth of local economy and the daily life of people
in nearby areas.
Based on the above reasons, some scholars have carried out researches on the combustion characteristics of cable fire and the
spread characteristics of cable trench fire [1,2]. Magalie [3] studied the fire behavior of electrical cables in cone calorimeter. The fire
performances of small telecom halogen-free cables were measured using cone calorimeter by changing several test conditions and
cable properties. An et al. [4] studied temperature distribution and CO diffusion induced by cable fire in L-shaped utility tunnel. Based
on cone calorimeter, Martinka et al. [5] revealed the influence of cable spacing and thermal conductivity of cable support material on
its fire characteristics. Zavaleta and Hanouzet [6] improved the previous cable fire model (flash-cat) to predict the heat release rate of
horizontal multi-layer cable fire under wall sticking condition. Based on the classical fire heat transfer model, Hasegawa et al. [7]
proposed a prediction model for the ignition time and upward fire spread rate of cables under the action of external radiant heat flux,
and then evaluated and graded the fire performance of different types of cables. Tewarson et al. [8,9] obtained an empirical formula for

* Corresponding author.
E-mail address: [email protected] (H. Zhu).

https://doi.org/10.1016/j.csite.2021.101338
Received 25 February 2021; Received in revised form 23 July 2021; Accepted 10 August 2021
Available online 21 August 2021
2214-157X/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
X. Liu et al. Case Studies in Thermal Engineering 28 (2021) 101338

Nomenclature

ρ Density
t Time
u Velocity
g Acceleration of gravity
f Volume force vector
τ Viscous tension per unit area
p Pressure
∇ Hamiltonian operator
q̇ Heat
Y Mass concentration
O Shading degree
l Unit length of light attenuation
h Specific enthalpy
qr Thermal radiation flux
T Temperature
D Diffusion coefficient
p0 Background pressure
R Molar gas constant
M Molecular weight of mixed gas
Exp Experiment
RTI Time response index of the detector
δt Characteristic time
Km Extinction coefficient
Sim Simulation

the vertical upward fire propagation rate of cables and proposed the critical conditions for cables to sustain fire spread. Moreover, the
influences of cable material, size and other parameters on fire spread behavior were revealed. Delichatsios et al. [10] proposed the
equivalent parameters to characterize the ignition and combustion characteristics of PE (polyethylene)/PVC (Polyvinyl chloride) cable
bundles, and established the prediction model of fire spread behavior of PE/PVC cable bundles in open space. Recently, some re­
searches have been carried out in the aspect of temperature field simulation of cable trench fire. Zhang [11] investigated the maximum
temperature and smoke back-layering length in the downhill direction from the fire source in a tilted tunnel under natural ventilation.
Numerical simulations were conducted using FDS to study the smoke flow behaviors for a fire in a tunnel with nine tunnel slopes of 0,
1%, 2%, 3%, 4%,5%, 6%, 7% and 8%.
At the same time, in terms of cable trench fire detection, the provisions of GB 50414-2007 [12] clearly put forward that the key
areas such as cable interlayer, electrical basement and cable tunnel in the plant should be equipped with automatic fire alarm system.
Hausner et al. [13] established the working model between temperature field and temperature sensing fire detector, tunnel specific
flame detector and tunnel specific image fire detector.
However, the existing cable trench fire monitoring system still has the phenomenon of false alarm. In order to verify the accuracy, it
will cost more to carry out the full-scale fire monitoring system alarm experiment. At the same time, there is a lack of standards in the
field of fire monitoring system accuracy. Therefore, on the basis of the existing research, the following work were carried out: (1) The
alarm sequence of the fire monitoring system is recorded through the full-scale substation cable trench fire experiment. (2) FDS
software is used to simulate the temperature field and combustible gas concentration of cable trench fire. (3) The accuracy of fire
detector with different sensitivity based on simulation method is discussed. This study provides a reference for the design of substation
cable trench fire monitoring system based on multi-sensor fusion technology.

Fig. 1. Schematic diagram of experimental scene.

2
X. Liu et al. Case Studies in Thermal Engineering 28 (2021) 101338

2. Experimental setup and models of FDS

2.1. Experimental setup

In order to ensure the authenticity of the experimental results of cable trench fire alarm, the experiment is carried out in the cable
trench of substation. The total length of the test site for cable trench fire is 60 m. The cross section of cable trench is 0.9 m × 0.78 m.
Fire experiments are carried out in the cable trench 10 m away from the left port. The arrangement of various detectors is shown in
Fig. 1. A group of point-type smoke detectors (model jty-gm-d12) and point-type temperature detectors (model jtw-zdm-d13a) are
arranged every 10 m from the ignition position. Two cable-type temperature detectors (model jtw-ld-sm3003c) with threshold values
of 85 ◦ C and 105 ◦ C are arranged in the experimental cable line. Suction-type smoke detector (model asd-200d) is arranged on the
channel wall. Cameras (model ds-ipc-b12v2-i) are arranged at both ends of the cable trench to observe the fire dynamics in the cable
trench. In order to simulate the real environment of the cable trench, the experimental cable is placed on the foot of the cable trench,
0.1 m away from the bottom of the trench.
Industrial alcohol with 99% concentration is selected as ignition source, and oil basin with diameter of 100 mm is used as fuel
carrier. Eight cross-linked polyethylene (XLPE) cables (25 mm diameter) are used in the experiment. After the cable is ignited by the
open fire, the alarm sequences of the fire detectors are recorded through the background control terminal. In order to reduce the
experimental uncertainty, a total of 5 groups of alarm experiments (Exp 1 to 5) are carried out, and finally the average value of the
alarm time of each sensor is counted. During the experiment, in order to ensure that other power cables and control cables are not
affected by fire, fireproof cotton is used to wrap them to ensure safety.

2.2. Theoretical method of FDS

(1) Governing equation

FDS (Fire Dynamics Simulator) is a computational airport simulation software developed by NIST (National Institute of standards
and Technology). It is a computational fluid dynamics (CFD) model with fluid movement in fire as the main simulation object, which is
suitable for the study of smoke and heat transfer caused by fire. In FDS calculation, the model space is divided into many micro element
control bodies, and the micro element control body is taken as the basic unit. It is assumed that the values of various physical quantities
such as temperature, pressure, speed, density, etc., in the control body of each element are the same, and they only change with time.
The accuracy of simulation calculation is mainly related to the number of micro element control bodies, and the number of micro
element control bodies is mainly limited by the ability of computer. Like other simulation software, FDS is also based on the N-S partial
differential equations for numerical solution, and gives the distribution fields of various physical quantities (temperature, pressure,
velocity, density, heat release rate, smoke components, etc.) [14,15]. The governing equations are used by FDS for solution via nu­
merical method.
Mass conservation equation:
∂ρ
+∇·→
uρ=0 (1)
∂t

where ρ is the density, t is the time, →

u is the velocity vector, and ∇ is the Hamiltonian operator.
Momentum conservation equation:
[ → ]
∂u
ρ + (→
u · ∇)→u + ∇ · p = ρg + f + ∇ · τ (2)
∂t

where g is the acceleration of gravity, f is the volume force vector, τ is the viscous tension per unit area, and p is the pressure.
Energy conservation equation:
∂ ∂p ∑
(ρh) + ∇(ρh→
u )= +→u · ∇p − ∇ · qr + ∇ · (k∇T) + ∇ · (hi ρDi ∇Yi ) (3)
∂t ∂t i

where h is the specific enthalpy, qr is the thermal radiation flux, T is the temperature, k is the heat conductivity coefficient, Dis the
diffusion coefficient of the ith ingredient, and Yl is the mass fraction of the ith ingredient.
Ingredient conservation equation:
∂ρYi
+∇·→
u ρYi = ∇ · ρDi ∇Yi + Mi m (4)
∂t

where Mi m is the mass production rate of the ith ingredient.

Ideal gas state equation:
/
∑
P0 (t) = ρTR (Yi / Mi ) = ρTR M (5)
i

3
X. Liu et al. Case Studies in Thermal Engineering 28 (2021) 101338

where p0 is the background pressure, R is the molar gas constant, M is the molecular weight of mixed gas, and i is the ith ingredient.

(2) Response model of temperature detector

The working principle of the temperature detector is that the thermal radiation and convection of the high temperature flame are
detected by the thermal element in the sensor. When the temperature rises to the alarm threshold, the detector is activated. The heat
transfer model of the temperature detector in FDS software can be expressed by the following formula:
q̇total = q̇cond + q̇conv + q̇rad (6)

where q̇total is the total heat received by the temperature detector, q̇cond is the heat of conduction, q̇conv is the heat of convection, and q̇rad
is the heat of thermal radiation. In the early stage of fire, the heat transmitted to the detector by thermal radiation can be ignored.
Compared with convective heat transfer, the proportion of heat conduction is also very small, which can be ignored [16]. Therefore,
the total heat can be expressed as:
( )
q̇total = q̇conv = hA Tg − Td (7)

where Td is the detector temperature, Tg is the flue gas temperature. In the process of simulation, the detector is regarded as the point of
mass concentration. By introducing equation (7) into the energy conservation equation of the particle, the energy conservation
equation of the detector can be obtained as follows:
( )
dTd Tg − Td mc
= ,τ= (8)
dt τ hA

where m is the mass of the detector, c is the specific heat capacity of the detector, h is the convection heat transfer coefficient. For a
given detector, m, c, and A are constant values. When the flue gas velocity is u, h is approximately proportional to u0.5 , Therefore, the
following formula can be deduced:

τu0.5 = RTI (9)

where RTI is the time response index of the detector, which is 96 m0.5s0.5 [17]. When FDS is running, the difference method is used to
solve the partial differential equation of detector heat transfer, and the time and space are discretized.

(3) Response model of smoke detector

FDS can not directly simulate the working process of smoke detector. Therefore, in FDS, the response process of smoke detector is
simplified to calculate the CO shading degree in the smoke detector. When the CO opacity reaches a certain critical value, the smoke
detector will be activated [18]. The mass fraction of CO in the sensing chamber of the smoke detector is calculated by the following
formula:
dYCO Ye (t − δte ) − YCO (t)
= (10)
dt δtc

where YCO is the mass concentration of CO in the cable trench. Ye is the mass concentration of free flue gas outside the detector. YCO and
Ye can been obtained by FDS simulated. δte is the characteristic time of CO filling the whole detector. δtc is the characteristic time when
CO fills the sensing chamber of the detector. Cleary et al. [19] considered that these two characteristic times are functions of the
external free air velocity u:

δte = αe uβe
(11)
δtc = αc uβc

where αe , αc , βe , βc are the empirical coefficient. The default values in FDS are: 0.8, -0.9, 2.5, -0.7 [19]. The theoretical solution of
sensing indoor CO concentration is obtained from (10)and (11). Furthermore, the calculation of CO mass concentration is converted to
the calculation of the unit shading rate per unit length:
( ) /
O = 1 − e− Km ρYCO l × 100% l (12)

where O is shading degree, Km is the extinction coefficient. For most combustible materials, when the light wavelength is 633 nm, the
recommended value of Km is 8700 m2/kg±1100 m2/kg l is the length of light attenuation.

2.3. Cable trench model of FDS

In order to study the influence of FDS simulating different fire detectors on cable fire early warning mechanism, the experimental
and simulation settings are as follows: For the open fire and smoldering experiments, the fire source power and position of each

4
X. Liu et al. Case Studies in Thermal Engineering 28 (2021) 101338

experiment are set unchanged.

According to the actual size of cable tunnel, a simulation model is established by FDS. The model cable trench is 60 m in length with
cross section of 0.9 m × 0.78 m, which is the same as the experimental condition. In the process of FDS simulation, the mesh system
directly affects the simulation results. Therefore, the mesh size in cable combustion area is 0.001 m × 0.001 m × 0.001 m, while the
size in non-cable combustion area is 0.05 m × 0.05 m × 0.05 m. The experimental diagram of cable trench fire and the layout of
temperature measuring points are shown in Fig. 2. “THCP” denotes “thermocouple”.
The combustible cable is located 0.1 m from the bottom of the cable trench and 0.25 m from the left side of the cable trench, and the
simulated cable is placed on the cable tray. The positions of various detectors are consistent with the real experiment. Moreover, in the
process of cable trench fire simulation, the concentration measuring points of toxic and harmful gas CO are added. The specific
measuring points are located in the center of cable trench ceiling within the range of 0 ~ 15 m from the fire source.

3. Results and discussion

The temperature field, combustible gas concentration, and fire alarm sequence of fire detector with different sensitivity are dis­
cussed in this section.

3.1. Experimental results of alarm time

The alarm time of different cable fire detectors is shown in Table 1. Exp1 to 5 represent the five experiments carried out in this
paper. The values in the table correspond to the alarm time of each experiment. At the same time, the average value of five experiments
is counted. According to the different physical quantities, the detection system with the fastest trigger action is the cable-type linear
temperature detector, and the average alarm time is 41.2 s. The second one is the suction-type smoke detector, and the alarm time is
not more than 120 s. Due to the low temperature and high humidity of the environmental thermometer in the cable trench, the alarm
time of the point type smoke fire detector in the cable trench is lagging behind. However, the point-type temperature detector does not
give an alarm.

3.2. Analysis of simulated temperature distribution in cable trench fire

Fig. 3 shows the temperature field distribution curve of cable trench fire. In the figure a ~ e, the space temperature field distribution
curves of the cross sections 0 m, 0.5 m, 1.0 m, 1.5 m and 2.0 m away from the fire source in the cable trench fire simulation experiment
are given respectively. In the figure, 0.25 m, 0.45 m, 0.65 m and 0.85 m represent the distance from the bottom of the cable trench in a
certain cross section. In the figure, - 0.2 m, 0 m, 0.2 m, 0.4 m and 0.6 m respectively indicate the horizontal distance from the center of
the fire source in the cross section direction.
It can be seen from Fig. 3 that with the longitudinal increase of the distance from the fire source, the temperature of the measuring
point shows a gradual attenuation trend. At the same fire source section position, with the gradual approach from the cable trench
ceiling, the corresponding temperature also gradually increases.
As shown in Fig. 3a, this section is the location of the fire source, and its temperature shows great fluctuation. At the height of 0.25
m, the maximum temperature at 0 m away from the fire source is as high as 125 ◦ C, while the temperature at 0.4 m and 0.6 m away
from the fire source decreases greatly. In the process of combustion, the heat radiation of flame is the main reason for this phe­
nomenon. In addition, the temperature at the ceiling of fire source section also fluctuates greatly. The heat radiation is most obvious
above the fire source, which leads to the highest temperature at this position. With the increase of the distance from the fire source, the
temperature gradually decreases. At the same time, it is observed at the height of 0.65 m that the temperature is the highest at 0 m
away from the fire source, the second is at 0.2 m, and the lowest is at - 0.2 m. This phenomenon is caused by the combined action of
flame radiation and flue gas temperature. The phenomenon of the highest temperature on both sides and the low temperature in the

Fig. 2. Cable burning experiment and temperature measuring point distribution.

5
X. Liu et al. Case Studies in Thermal Engineering 28 (2021) 101338

Table 1
Alarm time of cable fire detectors.
Types of detectors Exp 1 (s) Exp 2 (s) Exp 3 (s) Exp 4 (s) Exp 5 (s) average value (s)

cable-type temperature detector 45 42 40 39 40 41.2

suction-type smoke detector 115 120 100 109 113 111.4
point-type smoke detector 350 308 322 343 336 331.8
point-type temperature detector / / / / / /

middle is obvious in the range of 0.5 m from the fire source. When it is beyond this range, the temperature will be greatly affected by
the flue gas temperature. As shown in Fig. 3a ~ e, the height of 0.65 m, this phenomenon shows that the closer to the inner wall of the
right side of the cable trench, the higher the temperature.

3.3. Analysis of simulated smoke concentration in cable trench fire

Smoke and toxic gases, such as carbon monoxide, hydrogen cyanide, and hydrogen chloride, are the most fatal factors in fires. Fig. 4
shows the CO concentration curve of the ceiling at different distances from the fire source in the cable trench fire simulation exper­
iment. The results show that the release concentration of CO increases rapidly at first, then decreases, and finally increases. The reasons
are as follows: (1) The fuel of cable material is surface rubber. When combustion occurs, combustible materials on the outer surface
pyrolysis and burn, resulting in an increasing CO release rate. (2)With the deepening of pyrolysis, the combustion of cable gradually
transfers from surface material to intermediate sheath material. A carbonized layer is produced when the surface material is burning.
The carbonization layer hinders the combustion of the material to a certain extent, and then causes the CO emission rate to decline
[20]. (3)When the cable continued to burn, the carbonized layer of the cable pyrolysis and CO began to release again. (4) At the same
time, the cable fire spread transversely, which accelerated the combustion progress of the cable, and then accelerated the release of CO,
resulting in the continuous increase of CO concentration.
By analyzing the CO time curves in the 1 ~ 5 m and 6 ~ 10 m plots, it can be seen that the CO concentration decreases slowly with
the increase of the distance from the fire source at the same time. The concentration of CO in the cable trench increases slowly with
time. By analyzing the CO concentration curve in the 1 ~ 35 m diagram, it can be seen that CO concentration does not exist at the
beginning, but gradually increases with the combustion of the cable. The most obvious position is the initial change time of CO
concentration. Through the analysis of the simulation data, it is known that the initial change time of CO concentration at 10 m, 20 m,
and 30 m from the fire source is 68 s, 144 s, and 251 s, respectively. According to the above data, it can be deduced that the smoke
propagation velocity of cable combustion in substation cable trench is 0.147 m/s, 0.139 m/s, and 0.120 m/s, respectively, and the
average smoke propagation velocity calculated by 10 m interval is basically the same.

3.4. Sensitivity analysis of cable trench fire detection

In order to better study the influence of fire detectors with different sensitivities on cable fire warning mechanism, sensors with
different sensitivities are set, which is shown in Table 2. Sim 1 to 4 indicates that four simulations have been carried out. According to
GB 4715 [21], the critical shading degree of suction-type smoke detector are 3.28%/m and 4%/m, and the critical shading degree of
point-type smoke detector are 4%/m and 5%/m. According GB 16280 [22], the time response index of cable-type temperature detector
are 85 ◦ C（105 ◦ C）.
The parameters of the above sensors are based on the fire alarm experiment of cable trench. The simulation results of temperature
field and combustible gas concentration. In the process of simulation, the position of sensors is corresponding to the real experiment.
The sensitivity of scene Sim 1 is consistent with that of fire detector. The simulation scenario is set as follows: (1)The temperature
inside and outside the cable trench is set at 20 ◦ C. (2)Both sides of the cable trench are opened. (3)The heat release rate of fire source
remains unchanged in combustion experiment. (4)The simulation time of combustion experiment is 400 s. The results are shown in
Fig. 5.
In the Sim 1 scene, the alarm times of cable-type temperature detector, suction-type smoke detector and point-type smoke detector
are 37 s, 115.7 s and 349 s, respectively. Compared with the detectors with the same alarm sensitivity in the real experiment, the alarm
time is basically the same, which indicates that the fire detection and alarm experimental research can be realized by numerical
simulation. In the Sim 2 scenario, the alarm time of cable type linear thermal fire detector is 37.3 s, that of air breathing smoke fire
detector is 115.7 s, and that of point type smoke fire detector is 293.2 s. Compared with Sim 1 scene, the alarm time of point type smoke
fire detector is shortened. In the Sim 3 scene, the alarm time of cable type linear thermal fire detector is 41.1s, that of air breathing
smoke fire detector is 217.0s, and that of point type smoke fire detector is 336.1s. Compared with Sim 1 scene, the alarm time of air
breathing fire detector is longer. In the Sim 4 scene, the alarm time of cable type linear thermal fire detector is 41.1 s, that of air
breathing smoke fire detector is 217.0 s, and that of point type smoke fire detector is 272.3 s.
Fig. 6 shows the comparison of alarm time of each fire detector in cable combustion experiment and numerical simulation
experiment. The point-type temperature detector does not alarm in the real experiment and numerical simulation. Therefore, the
relevant data are not shown in Fig. 6. In view of the above fire detectors, it is obvious that the cable-type linear temperature detector
has the earliest alarm time after the fire, and the simulation results are basically the same as the experimental results. It can be found
that the sensitivity of the suction-type smoke detector has a great influence on the fire alarm time, and the response time of the fire is

6
 X. Liu et al.
7

Case Studies in Thermal Engineering 28 (2021) 101338

Fig. 3. Spatial temperature distribution curve of cable trench fire.
X. Liu et al. Case Studies in Thermal Engineering 28 (2021) 101338

Fig. 4. Variation curve of CO concentration at different distances.

Table 2
Sensor sensitivity setting.
scene suction-type smoke detector point-type smoke detector cable-type temperature detector

Sim 1 3.28%/m 5%/m 85 ◦ C（105 ◦

C）
Sim 2 3.28%/m 4%/m 85 ◦ C（105 ◦
C）
Sim 3 4%/m 5%/m 85 ◦ C（105 ◦
C）
Sim 4 4%/m 4%/m 85 ◦ C（105 ◦
C）

delayed by about 50%. Finally, for the point-type smoke fire detector, the simulation results of this type of detector with different
sensitivity are slightly different. The alarm time of point type smoke detector with high sensitivity threshold will be delayed about 60 s
under the same fire conditions.

4. Conclusion

In this paper, based on the experimental results of different detector alarm time sequence under the condition of cable trench fire,
the alarm characteristics of different sensitivity fire detectors are simulated by FDS software. The main conclusions are as follows: (1)
Through the simulation of fire detectors with different sensitivities, it is found that the response time of fire detector with lower
sensitivity is shorter, otherwise, the response time is longer. (2) For the suction-type smoke detector, the alarm time will be unstable in
the simulation process. For example, the sensitivity of Sim 3 and Sim 4 is 4%/m. Compared with high sensitivity scenes Sim 1 and Sim 2
(sensitivity set to 3.28%/m), the fire response time delay is about 50%.
This study provides a reference for long-distance large space fire detection simulation, the selection of different sensitivity sensors,
and the design of fire monitoring system. At the same time, it provides data support for the development of multi-sensor fusion fire
detection and early warning system.

Author statement

All authors have made substantial contributions to the conception of design of the work, and acquistion, analysis, interpretation of
data for the work.
All authors approved the final version to be published.

8
X. Liu et al. Case Studies in Thermal Engineering 28 (2021) 101338

Fig. 5. Simulation of alarm time sequence of fire detector.

Fig. 6. Alarm time of each fire detector in cable combustion experiment and simulation experiment.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

Acknowledgements

The work was supported by State GRID Company Science and Technology Projects (5455DW210005). Sichuan Fire Research
Institute of MEM Science and Technology Projects (T2019880801).

9
X. Liu et al. Case Studies in Thermal Engineering 28 (2021) 101338

References

[1] P. Zavaleta, R. Hanouzet, T. Beji, Improved assessment of fire spread over horizontal cable trays supported by video fire analysis, Fire Technol. 55 (2018)
233–255.
[2] R. Passalacqua, P. Cortes, N. Taylor, D. Beltran, P. Zavaleta, S. Charbaut, Experimental characterisation of ITER electric cables in postulated fire scenarios,
Fusion Eng. Des. 88 (2013) 2650–2654.
[3] C. Magalie, C. Anne-Sophie, S. Rodolphe, F. Laurent, G. Emmanuelle, L. Christian, Fire behaviour of electrical cables in cone calorimeter: influence of cables
structure and layout, Fire Saf. J. 99 (2018) 12–21.
[4] W. An, Y. Tang, K. Liang, M. Cai, T. Wang, Z. Wang, Study on Temperature Distribution and CO Diffusion Induced by Cable Fire in L-Shaped Utility Tunnel, vol.
62, Sustainable Cities and Society, 2020.
[5] Martinka Jozef, Rantuch Peter, Sulová Janka, Martinka Filip, Assessing the fire risk of electrical cables using a cone calorimeter, J. Therm. Anal. Calorim. 135
(2018) 3069–3083.
[6] P. Zavaleta, R. Hanouzet, T. Beji, Improved assessment of fire spread over horizontal cable trays supported by video fire analysis, Fire Technol. 55 (2019)
233–255.
[7] H.K. Hasegawa, K. Staggs, A.E. lIPSKA-Quinn, N.J. Alvares, A study of the fire performance of electrical cables, Fire Saf. Sci. 3 (1991) 237–247.
[8] A. Tewarson, M. Khan, Fire propagation behavior of electrical cables, Fire Saf. Sci. 2 (1989) 791–800.
[9] A. Tewarson, M.M. Khan, Flame propagation for polymers in cylindrical configuration and vertical orientation, Sympos. Combust. 22 (1989) 1231–1240.
[10] M. Delichatsios, Upward flame spread and critical conditions for Pe/pvc cables in A tray configuration, Fire Saf. Sci. 4 (1994) 433–444.
[11] A Xz, A Yl, B Cs, C Jz, Numerical Simulation on the Maximum Temperature and Smoke Back-Layering Length in a Tilted Tunnel under Natural Ventilation -
ScienceDirect, Tunnelling & Underground Space Technology;vol. 107.
[12] China SPo GB 50414, Code for design of fire protection and prevention for iron steel metallurgy enterprises, in: GB 50414: Code for Design of Fire Protection and
Prevention for Iron Steel Metallurgy Enterprises, 2007.
[13] M.B. Hausner, F. Suárez, K.E. Glander, N.V.D. Giesen, J.S. Selker, S.W. Tyler, Calibrating single-ended fiber-optic Raman spectra distributed temperature sensing
data, Sensors 11 (2011) 10859–10879.
[14] B. Karlsson, J. Quintiere, Enclosure Fire Dynamics, Henry Dream press, UK, 2008.
[15] J.G. Quintiere, Fundamentals of Fire Phenomena, John Wiley & Sons Ltd, 2006.
[16] US Department of Commerce N, Fire Dynamics Simulator, Technical Reference Guide, Sixth Edition, NIST SP - 1018.
[17] National fire alarm and signaling code, in: National Fire Alarm and Signaling Code, US-NFPA, 2019.
[18] Xiao, Fire simulation test and analysis of laminated bamboo frame building, Constr. Build Mater. (34) (2012) 257–266, 2012.
[19] T.G. Cleary, A. Chernovsky, W.L. Grosshandler, M. Anderson, Particulate entry lag in spot-type smoke detectors, Fire Saf. Sci. 6 (2000) 779–790.
[20] L.H. Hu, N.K. Fong, L.Z. Yang, W.K. Chow, Y.Z. Li, R. Huo, Modeling fire-induced smoke spread and carbon monoxide transportation in a long channel: fire
Dynamics Simulator comparisons with measured data, J. Hazard Mater. 140 (2007) 293–298.
[21] China SPo, GB 4705-2005: point type smoke detector, in: GB 4705-2005: Point Type Smoke Detector, 2005.
[22] China SPo GB 16280, Line type heat detectors, in: GB 16280: Line Type Heat Detectors, 2014.

10