sensors

Article
Mathematical Method of Rational Placement of Gas
Fire Sensors †
Andrey Petrov

Department of Computer-Aided Engineering and Design, National University of Science and Technology
MISIS (NUST MISIS), 4, Leninsky Prospect, Moscow 119049, Russia; [email protected]
† This paper is an extended version of our paper published by Petrov, A.; Fedorov, A.; Mintsaev, M.; Ilyukhin, A.;

Marsov, V. Theoretical Foundations and Methods for the Rational Location of Gas Fire Sensors Based on Gas
Control Technology. Mathematical Modelling of Gas Fire Sensors Location for Early Fire Detection. In XV
International Scientific Conference “INTERAGROMASH 2022”, Rostov-on-Don, Russia, 25–27 May 2022;
Beskopylny, A., Shamtsyan, M., Artiukh, V., Eds.; INTERAGROMASH 2022. Lecture Notes in Networks and
Systems; Springer: Cham, Switzerland, 2023; Volume 574. https://doi.org/10.1007/978-3-031-21432-5_178.

Abstract: Ensuring fire safety is an important condition for the successful operation of industrial
enterprises. It is necessary to detect the threat of fire before the ignition. An indicator of danger
is the appearance of gas, which occurs as a result of unauthorized heating of the equipment. Gas
sensors make it possible to detect the appearance of gases characteristic for the decomposition of
materials at the stage of smouldering and pyrolysis, which cause a change in the composition of the
atmosphere. This paper presents a mathematical method for the rational placement of gas analyzers
in the premises of potentially hazardous industries. The rational placement of gas sensors ensures
safety, taking into account economic feasibility. For the first time, an analysis of the accumulation of
gas flow from the source to the ceiling, and then to the location of the sensors, was used to select the
location of gas sensors. Depending on the permissible values of the gas volume, the placement and
number of gas sensors are selected. The calculation of the conditions for the rational placement of
gas sensors is carried out according to the most common indicators, such as carbon monoxide and
hydrogen at different sizes and heights of premises.

Keywords: gas sensors; mathematical model; rational location; automated system; early fire detection;
gas control systems

Citation: Petrov, A. Mathematical

Method of Rational Placement of Gas
1. Introduction
Fire Sensors. Sensors 2023, 23, 8349.
https://doi.org/10.3390/s23208349 Timely detection of fire hazards is a critical condition for ensuring safety in agriculture,
industry and the environment. In systems for early detection of fire hazards in industry,
Academic Editor: Michele Penza
highly sensitive gas analyzers are used, in which sensors ensure the fixation of trace
Received: 25 August 2023 concentrations of hazardous gases. They are used to detect the possibility of ignition by
Revised: 29 September 2023 gas emission at the stage of smouldering (decomposition) and pyrolysis (overheating), i.e.,
Accepted: 6 October 2023 before the appearance of an open flame.
Published: 10 October 2023 Most of the combustible substances are organic compounds, the smouldering or
combustion of which emits carbon dioxide CO2 and carbon monoxide CO. The sensor is
a sensitive element of the gas detector, which registers an increase in the concentration
of these gases and gives a signal that the specified limit of the controlled parameter has
Copyright: © 2023 by the author.
exceeded. When smouldering, the concentration of CO increases from 20 to 100 mg/m3 ,
Licensee MDPI, Basel, Switzerland.
but when a flame appears, it decreases. Also, during the smouldering of organic materials,
This article is an open access article
hydrogen, H2 , is released, the concentration of which, up to 10 mg/m3 , can be detected
distributed under the terms and
conditions of the Creative Commons
in the presence of semiconductor sensors. When a small amount of material smoulders,
Attribution (CC BY) license (https://
the concentration of hazardous gases is small and they are distributed in the volume of the
creativecommons.org/licenses/by/
room due to diffusion. For gas detectors reacting at this stage of the fire, the requirement for
4.0/). the sensitivity threshold of the sensors is from 0.0001% for CO and 0.00001% for H2 [1,2].

Sensors 2023, 23, 8349. https://doi.org/10.3390/s23208349 https://www.mdpi.com/journal/sensors

Sensors 2023, 23, 8349 2 of 17

The installation of gas control systems in the protected room makes it possible to
detect the appearance of gases characteristic for the decomposition of organic materials,
including insulation of cables, packaging, wood, fabrics and much more [3]. Thermal
degradation in different materials begins at different temperatures, but it is ahead of the
appearance of smoke. The heating process is slow. At the same time, the physical principles
underlying the operation of the fire sensors used (temperature increase, light scattering
by smoke particles, ion current due to radioactive ionization of smoke) reliably register a
fire already at the combustion stage when an open flame appears, which is accompanied
by a strong change in the physical properties of the surrounding air [4]. A fire at the open
flame stage cannot be eliminated by technical measures by changing the operating mode,
for example, turning off the equipment.
A reliable way to prevent a fire at an early stage is to control the chemical composition
of the air, which changes dramatically due to thermal decomposition, pyrolysis, overheating
and the smouldering of combustible materials [5]. It is at this stage of the development of
the fire that the necessary measures can be taken to extinguish it, and the in case of the
overheating of electrical equipment and cables, they can be automatically turned off in time
by a gas sensor signal.
Gas sensors are used all over the world in various fields as fire sensors to improve
the reliability of fire safety systems [6–12]. However, the developed norms and guidelines
do not consider methods for the rational arrangement of sensors. A set of rules for the
design of fire protection systems was developed by the Ministry of Emergency Situations
of Russia for timely and reliable fire detection, providing information to the personnel on
duty [6]. The placement of gas analyzers for plant protection systems and life safety is
presented in [7]; guidelines for the use of stationary or portable gas detectors are proposed
in [8]. The use of gas sensors in industry, as well as in other areas, is considered in [9,10].
The application of gas detectors for real-time detection of chemicals in the workplace
is discussed in [11]. The use of gas sensors for stationary and portable gas detection
in industrial processes associated with the use and production of hazardous substances,
especially toxic and combustible gases, is discussed in [12].
Guidance on the placement of individual sensors of combustible gases in rooms is
given in [13]. The main recommendations for the selection of fire and gas detectors are
presented in [14]. The placement of sensors in industrial premises is not regulated in these
rules and recommendations. In [15], modeling is used to optimize the location of sensors
for early detection of a gas cloud. In operation [16], the system with the microcontroller
should detect a gas leak, issue a warning via SMS, and turn on the fire extinguishing
system. A formulation of nonlinear programming with a mixed integer was proposed for
the optimal placement of flame detectors, taking into account the non-uniform probabilities
of detection failure [17]. An approach to a fire detection system in valuable forests is
proposed, which can detect fires from the earliest stage using a mathematical model of the
detector geometry [18].
Convolutional neural networks are used to detect fire based on aerial photography [19].
The results of experimental investigations of the main characteristics of a typical indoor
seat of fire at the stages of initiating the ignition of combustible materials and in the process
of their burning are presented in [20]. Wood material, wood fiberboard, linoleum, and
plastic panels have been used as combustible materials.
A holistic approach to fire safety design, which describes fire models, is discussed
in [21]. An attempt to create an integrated automated system for the early detection of
fires and environmental monitoring was made in [22]. Gas sensors and methods of their
placement are necessary to ensure the fire safety of underground structures [23].
The author has developed network models for analyzing the flows of hazardous prod-
ucts to ensure the safety of oil refineries with practical application at the Moscow Refinery
and the Ryazan Refinery [24]. A network of branches is a tensor; different connections of
branches and structures are considered as projections of the tensor into different coordinate
systems. Coordinates are paths, lines passing through branches. Closed and open paths
Sensors 2023, 23, 8349 3 of 17

form orthogonal subspaces in the tensor network, the dimension of which changes as the
connections change. In a dual network, connections correspond to breaks, and vice versa.
The sum of the dimensions of the subspaces of closed and open paths in the two networks
remains constant. The duality invariant connects the transformations of the structure [25].
The duality of the networks made it possible to represent mass flows in the distillation
column as a contour network with a basis of closed paths, and heat flows as a nodal
network with a basis for open paths. This made it possible to propose the placement of
pressure sensors in the branches, which determine the basic closed paths in the installation.
Temperature sensors will place in the elements that determine the basic open paths. The
analysis shows where the maximum permissible values are exceeded when the branches
are broken, which simulates accidents in the installation.
In all cases, with the actual size of the production, it is necessary to place the sensors,
taking into account the requirements and conditions of ensuring safety.
The article presents a method for rational placement of sensors in production condi-
tions for early warning of a possible fire. The method makes it possible to create sensor
placement schemes, depending on changes in the size, volume and shape of protected
premises, types of production, creation of new sensors, and changes in the composition
of hazardous gases in different technological processes. The method allows you to adjust
sensor technologies by changing the layout, depending on the increase in the sensitivity
of the sensors, the shape and size of the room, and the requirements for the maximum
permissible volumes of hazardous gases at the stages of smoldering, before the occurrence
of a fire.
It is assumed that in case of malfunctions and heating, dangerous light gases appear;
for example, CO and H2 rise to the ceiling of the room where the sensors are located. The
physical foundations of the first stage of the process of gas propagation from the heating
source to the ceiling of the room were considered in [26,27]. The volume of gas accumulated
during the ascent spreads along the ceiling reaches the sensors, which give a signal about
the danger and the need to take measures to prevent fire. According to the coordinates of
the sensors that gave a signal about the appearance of hazardous gases, it is possible to
localize the heating source and automate the warning of a possible fire.

2. Location Methods for Gas Fire Sensors

2.1. Problem Statement
The volume of the production room is filled with air. The boundaries of the room
are the floor, the ceiling and the walls. In the room there is equipment that performs a
technological process with a potential fire hazard. A harbinger of fire is the appearance of
gas, which occurs as a result of unauthorized heating of the equipment or premises.
It is necessary to place gas fire sensors (GFS), which will record the appearance of
indicator gases such as H2 and CO. The sensor signal will allow you to determine the
possible localization of combustion products in the room to eliminate heat sources and
prevent fire, or take measures to extinguish the fire. The H2 and CO gases that appear as
a result of heating are lighter than air, and therefore, will rise upwards. In addition, the
density of the heated gas is less than the density of the air, so it is affected by a lifting force
equal to the weight of the volume of air displaced by the gas.
It is assumed that the most likely sources of gas are in the area of the floor where the
equipment is located, and there is a risk of fire or explosion. The controlled gases propagate
upwards and to the sides, so the gas sensors must be positioned at the top in such a way as
to ensure that low concentrations of H2 and CO are recorded before the concentration of
the gas emitted reaches a dangerous value.
There are various configuration options for gas sources (emissions):
• At the point;
• At several points;
• In the direction of the line (crack);
• On a surface area (surface rupture).
Sensors 2023, 23, 8349 4 of 17

The release of gas through a line or surface can be thought of as a generalization of

the release through a point. It can be assumed that the air in the room is stationary (or its
movement can be neglected), and the air is “transparent” to the spread of gas.
The placement of sensors should ensure contact with gas molecules when the accumu-
lated volume does not exceed the maximum permissible values set by safety standards.

2.2. Methods of Placement of Gas Sensors

There are various methods to solve this problem. For example, you can consider
the physics of the process from the standpoint of the location of gas sensors. The power
density methods and the location of the lamps solve the problem from the perspective
of the location of the gas sensors. It is assumed that the gas analyzer has a “range”. The
point method uses analogies with the luminous flux reaching the room areas that need to
be illuminated. The disadvantage of these methods is the assumption that the GFS itself
reaches the place of possible occurrence of danger.
In reality, the gas analyzer is in place, waiting for the molecules of the substance to
enter the inlet in order to then send a signal to the security system. For this reason, this
paper uses the method of considering the movement of hazardous gas molecules from the
place of danger in the direction of gas sensors.
The challenge is to determine how many gas sensors are needed and how to place
them in order to detect and locate possible gas emission sites in a timely manner and
monitor the permissible hazardous concentration. This method of placing the GFS solves
the problem from the position of gas movement from sources near the floor of the room
where the equipment operates, in the direction of the location of gas sensors on the ceiling
of the room.

2.3. Physical Model of the Gas Movement Process

Gas can appear as a result of heating the equipment located below, so the gas will
rise up. The density of the heated gas is less than the density of the air, so it is affected
by a lifting force equal to the weight of the volume of air displaced by the gas. Practice
shows that, in the source, the concentration (dynamic ratio) can be 300–350 ppm (parts per
million), and on the ceiling, at the height of H = 5–6 m, the concentration will be 80–90 ppm.
In this case, an approximate concentration gradient is set, which can be considered as a
potential difference.
Let the heat source heat the solid material, converting it into a gaseous state. The
pressure acts the same in all directions, but under the influence of heating, the gas mainly
moves upwards, creating convection currents. There is a stream of gas molecules that
occupy a certain volume at the heating site at the floor level. Let us assume that the source
has a constant intensity, i.e., the same number of gas molecules appear every second. At
constant pressure, the molecules occupy the same volume at any given moment, creating a
cloud of gas step by step.
As the gas cloud propagates, it moves upward. Random factors (convection, turbu-
lence, etc.) give the gas cloud a configuration that is difficult to predict. With a high degree
of certainty, which is sufficient to solve the problem, we will assume that the gas cloud has
the shape of a cone. In the process of movement, the cone occupies all new layers with a
vertex at the ejection point. Rising, the gas cloud will touch the ceiling of the room, and
then spread along the ceiling in all directions, falling into the GFS location areas.
The gas cloud will spread across the ceiling and reach one or more gas sensors, which
will signal the presence of gas in the room. According to the coordinates of the SFS (gas
sensors) that gave the signal, it is possible to solve the inverse problem of determining the
location of the gas source on the floor.
Thus, the three-dimensional problem is reduced to a two-dimensional problem. The
circle of gas on the ceiling that reaches the sensors must have the square that corresponds
to the volume of gas less than the maximum permissible value.
Sensors 2023, 23, 8349 5 of 17

The leading edge of the gas cloud on the ceiling will reach one or more gas sensors,
which will signal the gas presence in the room. According to the GFS (gas sensors) coor-
dinates that signaled, it is possible to solve the inverse problem in order to determine the
location of the gas source on the floor. Thus, a three-dimensional problem is reduced to a
two-dimensional problem—the need to rationally place the GFS on the room ceiling so that
they are touched by the boundary of the gas propagation circle of a certain diameter.
The analysis shows that at an equal density in the gas cloud, the propagation rate
changes in proportion to the cubic root of the time elapsed since the appearance of the
source. With this dependence, the rate of gas propagation decreases rapidly with distance
from the source.

2.3.1. Movement of Gas from the Source of Danger to the Ceiling of the Room
Heating or a chemical reaction converts a substance that can cause a fire from a
liquid or solid state to a gaseous state. The gas occupies a larger volume than the source
substance, and this excess, which creates pressure, begins to push the gas molecules into
the environment. They penetrate between air molecules, spreading throughout the room.
The mass of the gas mg ejected in the first time interval (per second) is equal to the density
of the gas q multiplied by the volume of the cone V (the radius of the base r and the height
h) that the gas will occupy.
mg = V q = 1/3 π r2 h q (1)
where:
mg —gas mass, mg;
V—cone volume, m3 ;
r—cone base radius, m;
h—cone height, m;
q—gas density, mg/m3 .
The concentration unit is ppm (parts per million). With a constant source power, the
same gas mass mg will be emitted in each unit of time. The main parameters reflecting the
state of the gas are the dynamic pressure P, temperature and density q. In this case, the task
is to determine the release of gas and the appearance of hazardous substance molecules.
Thus, the physical effect of a fire hazardous process is the dynamic pressure created as a
result of the emergence and propagation of gas molecules. In addition, the temperature
decreases from the source to the GFS location.
The gas is formed as a result of pyrolysis (thermal decomposition). Heating can lead
to fire, explosion, etc. The heated exhaust gas (H2 , CO) tends to rise. The pressure on
the gas acts in all directions. In addition, the gas molecules are affected by heating and
buoyant force directed upwards, so the gas cloud takes on an upward shape. With a good
approximation, we can assume that this is the shape of a cone. The top of the cone at each
moment of time rests on a source of dangerous gas, where pressure is created.
As time passes, the gas occupies new layers of the truncated cone; let the source have
a constant intensity, i.e., emits the same number of mg molecules in each time interval, then,
the layers of the cone have the same volumes for each time interval. The density of the
gas during propagation remains constant q = const. This is a stronger condition than the
assumption that the density gradually decreases as it moves towards the ceiling. Let us
consider a change in the gas velocity under the condition of a constant flow of molecules
emanating from the source and a constant gas density in a gas cloud.
Suppose that in the lower part of the room, near the floor, the pyrolysis process has
begun due to heating, or for other reasons. The source has a local character, which can
be represented as a point (see Figure 1). Under the influence of pressure, gas molecules
spread evenly, creating a cone-shaped cloud, penetrating between air molecules, displacing
them, as well as each other. It is assumed that the intensity of the source is constant. For
this reason, the number of molecules in all layers of the gas cone per unit time must be
the same.
 where h1—cone height, r1—cone base radius, S—cone base area.
As you can see in Figure 1, the cone radius is equal to the height multiplied by the
tangent of the angle а (equal to half the angle of the apex of the cone). Then, the volume of
the cone V can be expressed in terms of the cone height h and the angle α [27].
Sensors 2023, 23, 8349 6 of 17
V = 1/3 π h3 tg2 α.

Figure1.
Figure Gasdistribution
1. Gas distributionfrom
fromaasource
sourcelocated
locatedon
onthe
theroom
roomfloor.
floor. The
Theupper
upperbase
baseof
ofthe
thegas
gascloud
cloud
cone
coneisisflat.
flat.

Thecone
The coneheight
volume is occupied by the
the distance gas atby
traveled thethefirst moment
leading of of
front time
thewill
gas be:
per unit time,
which is numerically equal to the velocity of gas propagation in the first period. Strictly
speaking, for this reason, if theVcone 1 = 1/3 r1 2 = 1/3
toph1isπlocated h1 S,source at the bottom, then the
in the
base at the top should have a convex spherical shape.
where h1 —cone height, r1 —cone base radius, S—cone base area.
During the second period, per unit of time, the gas will spread further, occupying
As you can see in Figure 1, the cone radius is equal to the height multiplied by the
the next of
tangent layer
the of the αcone.
angle The
(equal to volume
half the of thisoflayer
angle is equal
the apex to cone).
of the the volume
Then,created in the
the volume of
first unit of time. From here you can find out the estimated speed
the cone V can be expressed in terms of the cone height h and the angle α [27]. of gas propagation.
Each layer of the gas volume increment has the shape of a truncated cone.
Let the angle of the cone α = 45°. V =During h3 tgsecond
1/3 π the 2
α. period, the gas will occupy the
volume V2. Since the power of the source is constant, then the volume increment, the
The cone height is the distance traveled by the leading front of the gas per unit time,
which is numerically equal to the velocity of gas propagation in the first period. Strictly
speaking, for this reason, if the cone top is located in the source at the bottom, then the base
at the top should have a convex spherical shape.
During the second period, per unit of time, the gas will spread further, occupying the
next layer of the cone. The volume of this layer is equal to the volume created in the first
unit of time. From here you can find out the estimated speed of gas propagation. Each
layer of the gas volume increment has the shape of a truncated cone.
Let the angle of the cone α = 45◦ . During the second period, the gas will occupy the
volume V 2 . Since the power of the source is constant, then the volume increment, the layer
dV 2 , which the gas occupies in the second unit of time is equal to the volume that the gas
occupied in the first unit of time V 1 . The volume increment layer dV 2 has the shape of a
truncated cone. Its volume is equal to the difference between the volume V 2 = 1/3 π r2 3 =
1/3 π h2 3 and the volume of the cone V 1 = 1/3π r1 3 = 1/3 π h1 3 . Then, one can write that

dV 2 = V 2 − V 1 = 1/3 π r2 3 − 1/3 π r1 3 = V 1 , (2)

where dV 2 is the increase in the volume of gas propagation in the second unit of time; hence
we obtain V 2 = 2 V 1 = 2/3 π r1 3 .
Sensors 2023, 23, 8349 7 of 17

Hence, it follows that the radius of the truncated cone r2 is equal to

√ √
r2 = r1 3 2 or as r = h, then h2 = h1 3 2 (3)

Similarly, we obtain that the radius of the third cone r3 is equal to

√ √
r3 = r1 3 3 or, as r = h, then h3 = h1 3 3 (4)

Continuing step by step to consider the sequential movement of truncated cone layers
from the gas source, we will thus obtain that in time t = n the gas will propagate to a
distance
√ √
rn = r1 3 n or hn = h1 3 n (5)
There will come a moment when the next layer (the base of the truncated cone) reaches
the location of the gas sensors at the level of the room ceiling. Further, the gas area will
begin to spread along the ceiling in concentric circles. The cloud boundary will reach one
or more gas sensors and the security system will signal the presence of gas. Let the angle of
the cone apex be α◦ , (expressed in angles or radians). The volume of the cone V through
the height of the cone h and the angle α, at the first unit time.

V 1 = 1/3 π h1 3 tg2 α,

Comparing the expressions for volumes at the first and second unit’s time, we obtain
that the second height is expressed in terms of the first height as
√ √
h2 = h1 3 2, but in general hn = h1 3 n

Thus, the increase in height, and hence the rate of gas rise, does not depend on the
angle at the cone base at a constant power of the gas source. The expected vertical gas
velocity is equal to the difference in the heights of the gas cloud cones per unit time:

v n = h n +1 − h n (6)

Thus, the room heights and the intensity of the gas source (the number of molecules
that flow out per unit of time and occupy a certain volume) are given. It is necessary to
calculate the time for the gas cloud to reach the room ceiling at the level of the gas sensors,
and the time it takes for it to reach one of the sensors.
Since each height increment is created per unit time, the height increment is numeri-
cally equal to the upward gas velocity, which is shown in Formula (6).
The time to reach the room ceiling is equal to the number of unit layers of the cone t
= n, at which the total cone height of the gas cloud hn becomes equal to the height of the
room H. The upward velocity of the gas, v, will decrease in proportion to the increase in
the area of the cone layer’s base.
Height H, the level of gas sensors location will be reached in time t, equal to the
number of layers of gas n when it propagates to the ceiling t = n
√
3 √
Hn = r1 t = r1 3 n (7)

The volume of the initial gas emission depends on the magnitude of the angle of the
base of the cone. If we assume that the angle at the base of the cone is 60 degrees, then the
initial volume will be three times less than for an angle of 90 degrees. The radius of the
gas cloud on the ceiling will be smaller. The speed of movement along the ceiling to the
locations of gas sensors will also change.
Consider an example of calculating the time and speed of gas propagation in a room.
Height values are selected taking into account the technological features of the protected
premises.
Sensors 2023, 23, 8349 8 of 17

Let the room height be 4 m. Let us take a second as a unit time. The gas ejection cone
in the first second has a height of 1 m, the cone apex angle is equal to the right angle, 90◦ ,
i.e., the cone radius is equal to the height, r = h, the gas source has a constant intensity.
Table 1 gives the characteristics of the cloud: the cone height over time, the velocity of the
upward gas propagation, and the volume dynamics of the gas cloud. It shows the time
during which, under given conditions, the gas cloud reaches the height of the ceiling, and
the accumulated volume of gas in the room.

Table 1. The geometrical characteristics of the gas cloud when rising to the ceiling.

The Difference
Time Since the The Height of the between the Accumulated
Gas Cloud Rise
Beginning of the Gas Cone at Each Room Height Heights of the Volume of Gas
Rate
Ejection, s Moment Room and the Cone
Cone
t, s r = h, m v, m/sec H, m H − ht V, m3
1 1.00 1.0000 4.0 3.00 1.05
10 2.15 0.4445 4.0 1.85 10.47
20 2.71 0.2482 4.0 1.29 20.94
50 3.68 0.1271 4.0 0.32 52.36
60 3.91 0.1119 4.0 0.09 62.83
64 4.00 0.0209 4.0 0.00 67.02
65 4.02 0.0207 4.0 −0.02 68.03

At 64 s, the gas cloud reaches the ceiling. The accumulated volume of gas in the room
is the circle radius and the cone base is equal to the cone height, therefore, it is 4 m, and
then the gas begins to spread horizontally under the ceiling. A similar calculation shows
that at a room height of 6 m, the gas reaches the ceiling at 220 s from the start of the ejection.

2.3.2. Gas Movement from the Place of Reaching the Ceiling to the Sensors
After raising the gas cloud to the ceiling, we obtain a circle at the top, the dimensions
of which may be less than the distance between the sensors. The gas will spread across the
ceiling until it reaches the sensor, which will generate an alarm, i.e., report the danger.
Suppose that when moving along the ceiling, the thickness of the gas layer is constant,
and is equal to the height of the last layer of the truncated cone that reached the ceiling.
In Table 1, the thickness of the gas layer at the ceiling was equal to 0.0207 m. For each
subsequent period of time (second), a ring of gas diverges along the ceiling, the volume of
which is equal to the previous volumes (assuming a constant intensity of the gas source).
When the ceiling is reached, the area of the circle Sn is equal (if the angle of the cone is 90◦ ):

Sn = π rn 2 = π hn 2 = π H2 (8)

This is a two-dimensional case of the pattern that is considered in the analysis of the
rise of a cloud of gas to the ceiling. That is, each ring that spreads across the ceiling has the
same area as the circle, the base of the cone, and a cloud of gas that has risen to the ceiling.
In the second period, the gas will spread further, occupying the next ring. The area
of the second ring, as well as subsequent rings, is equal to the area of the base of the cone.
From here we find the rate of gas propagation. For the second period of time, gas will
occupy an area of Sn+1 .
Sn+1 = π rn+1 2
Since the power of the source is constant, the increment of the area of the circle, the
ring, dSn+1 , occupied by the gas for the second unit of time, is equal to the area S1 , occupied
Sensors 2023, 23, 8349 9 of 17

by the gas for the first unit of time. The area of the ring is equal to the difference between
the area Sn+1 = π rn+1 2 and the area Sn = π rn 2 . Then, we obtain

d Sn+1 = Sn+1 − Sn = π rn+1 2 − π rn 2 = Sn , (9)

where dSn+1 is the increment of the gas propagation area for the second period of time of
movement along the ceiling, from which, we obtain Sn+1 = 2 Sn = 2 π rn 2 . It follows that
the radius of the ring rn+1 is √
r n +1 = r n 2 (10)
For similar reasoning, we obtain, in the third period of time, the gas that will occupy a
ring of radius rn+2 . √
r n +2 = r n 3 (11)
Continuing to consider the sequential propagation of the rings, we obtain, during the
time t = n + m, the gas that will propagate over a distance
√
rn+m = rn m (12)

Two points in time are of interest. The first is when the radius of the ring exceeds the
distance between the sensors. From this point on, gas detection is guaranteed. The second
point is when the radius of the ring reaches the size of the room. From this point on, we
can assume that the room is filled with gas.
Table 2 shows the calculation of the gas propagation on the ceiling before reaching
the sensor. Further, the movement of gas to the boundaries of a square room, with a side
of 15 m.

Table 2. The geometrical characteristics when the gas cloud moves along the ceiling to the sensors.

Time of Distance The Length Distance

The Radius The Speed Gas Cloud
Movement between of the Walls from the Gas
of the Ring of the Gas A − ri Volume V +
on the Sensors, A, of the Room, Ring to the
ri , m Ring Vup
Ceiling, s m L, m Walls L − ri
5 2.24 2.2361 6.0 3.76 72.26 15.00 12.76
10 3.16 0.9262 6.0 2.84 77.49 15.00 11.84
20 4.47 0.5992 6.0 1.53 87.96 15.00 10.53
30 5.48 0.4772 6.0 0.52 98.44 15.00 9.52
40 6.32 0.4085 6.0 −0.32 108.91 15.00 8.68
50 7.07 0.3629 6.0 −1.07 119.38 15.00 7.93
60 7.75 0.3298 6.0 −1.75 129.85 15.00 7.25
70 8.37 0.3043 6.0 −2.37 136.14 15.00 6.63
80 8.94 0.2840 6.0 −2.94 138.23 15.00 6.06

The first column shows the time elapsed since the beginning of the movement of the
gas cloud on the ceiling. The radius of the ring is calculated by Formula (12). The distance
between the sensors is A = 6 m. In this case, when the radius of the ring is equal to A at
40 s, the volume of gas will be 109 m3 . The gas reaches a distance of 6 m to the walls of the
room in 80 s. At the distance between the sensors A = 5, the radius of the ring is equal to A
at 25 s, the volume of gas will then be 93 m3 .
The intensity of the source determines the rate of propagation of the gas. At the same
time, the higher the velocity of the gas, the faster it will reach the sensor layer and will
be registered. In this respect, the power of the source and the increase in the dangerous
volume of gas are not directly proportional. This dependence can be widely neglected.
The amount of gas in the hemisphere under the ceiling is proportional to the volume and
Sensors 2023, 23, x FOR PEER REVIEW 10 of 18
Sensors 2023, 23, 8349 10 of 17

density of the gas. The assumption that the gas propagates in a hemisphere of the same
density is
density ofsufficient
the gas. The assumption
to ensure a levelthat the gas
of safety, propagates
since in a hemisphere
it is a stronger of thewould
condition than same
density
be is sufficient
necessary to ensure
if the density of athe
level
gasofdecreases
safety, since
as ititmoves
is a stronger condition
away from than would be
the source.
necessary if the density of the gas decreases as it moves away from the source.
2.3.3. Selecting Sensor Locations
2.3.3. Selecting Sensor Locations
It is necessary to choose the location of sensors in the protected area, which will
ensureIt is necessary
safety and willto be
choose the location
economically of sensors
feasible. in the protected
The hexagonal area, which
arrangement will
of the sen‐
ensure
sors safety and
provides the will be economically
densest coverage of the feasible. The hexagonal
protected area withoutarrangement
gaps; the of the sensors
distances be‐
provides the densest coverage of the protected area without gaps; the distances between the
tween the sensors form equilateral triangles, with the side that we denote through A.
sensors form equilateral triangles, with the side that we denote through A. Figure 2 shows
Figure 2 shows two variants of the hexagonal arrangement of sensors: on the left with a
two variants of the hexagonal arrangement of sensors: on the left with a lower frequency,
lower
with a frequency, with a sensors
distance between distanceA between
1 ; on the right withАa1;higher
sensors on thefrequency,
right withwith
a higher fre‐
a distance
quency,
A2 . The with
blackacircle
distance А2. The black
corresponds to the circle corresponds
boundary toarea,
of the gas the boundary
which hasof the to
risen gas area,
the top
which has risen
and spreads to the
along the top and spreads along the ceiling.
ceiling.

Figure
Figure 2.2. Views
Views from
from above.
above. Variants
Variants ofof hexagonal
hexagonal arrangement
arrangement ofof sensors
sensors with
with aa distance
distance between
between
them
them AА1 and
and AА2.. The
Theblack
blackcircle
circlerepresents
represents the
the boundary
boundary of of area
the the area
of ofthat
gas gashas
that hastorisen
risen the to the
ceiling.
1 2
ceiling.
With a more frequent arrangement of sensors, gas is detected earlier, with a smaller
With
volume of agas
more frequentThe
emissions. arrangement
ceiling areaofissensors,
coveredgaswithissensors
detected earlier,gaps,
without with except
a smaller
for
volume
areas nearof gas emissions.
the walls of theThe
room.ceiling area is covered
For complete with
closure, sensors sensors
additional withoutcangaps, except
be placed
for
nearareas near At
the walls. thethe
walls
same oftime,
the room. Forgas
when the complete closure,
rises along additional
the walls, sensors
the pressure willcan be
more
placed near the walls. At the same time, when the gas rises along the walls,
actively displace its molecules into the inner regions of the room, into the zones of action ofthe pressure
will more actively
the sensors. Thus, displace its molecules
at the walls and in theinto the inner
corners of theregions of the
room, the gasroom, intothe
reaches thesensors
zones
of action of the sensors. Thus,
faster than in the centre of the room. at the walls and in the corners of the room, the gas reaches
the sensors
Figurefaster
3 showsthan in theof
a circle centre
gas onof the ceiling
room. of the room, which is described around
Figureof3sensors
a triangle shows aarranged
circle of gas
in aon the ceiling
hexagonal of the room,
pattern. This iswhich is described
an element of the around
situationa
triangle
shown in ofFigure
sensors 2. arranged
The figure inshows
a hexagonal
the mostpattern. This is an
unfavourable elementthat
situation of the situation
occurs when
shown
the gas in Figure
region 2. Thethe
reaches figure shows
ceiling thecentre
in the mostof unfavourable
the triangle, situation that occurs
at the vertices of which when
the
the gas region
sensors reaches the ceiling in the centre of the triangle, at the vertices of which the
are located.
sensors are located.
 2023, 23,
Sensors 2023, 23, 8349
x FOR PEER REVIEW 1111of
of 18
17

Figure 3.
Figure 3. Gas
Gas propagation area touches
propagation area touches sensors
sensors on
on the
the ceiling,
ceiling, top
top view.
view.

position of
The position of the
thegas
gassource
sourceunder
underthe thesensors
sensorscorresponds
corresponds toto
thethe largest
largest volume
volume of
gas accumulated in the room. Therefore, if the gas reaches the level of the
of gas accumulated in the room. Therefore, if the gas reaches the level of the sensors at sensors at any
other point,
any other then then
point, the gas theleak
gas notification will be
leak notification received
will with smaller
be received volumes
with smaller of gas in
volumes of
the
gas room.
in the The
room.sensors form anform
The sensors equilateral triangle,
an equilateral the sidethe
triangle, of side
which of iswhich A—the
equalistoequal to
distance betweenbetween
A—the distance the sensors. The radius
the sensors. Theofradius
the described circle L, which
of the described circle L, determines
which deter‐ the
area
minesofthe
gasarea
propagation along the ceiling,
of gas propagation along theis ceiling, is
√√3
3
rn 𝑟= A А (13)
(13)
33
The mass
The mass m of aa given
mg of given volume
volumeofofgas,
gas,asasshown
showninin(1),
(1),with
with guaranteed
guaranteed registration
registration at
g
at a constant density q (which can be considered the maximum, i.e., the
a constant density q (which can be considered the maximum, i.e., the upper estimate upper estimate of
of the
the mass
mass of gas,
of the the gas, taking
taking intointo account
account diffusion
diffusion and and pressure
pressure change)
change)
mg max = Vg q (14)
mg max = Vg q (14)
In this case, the minimum volume of gas Vmin = Vup for registration occurs when the
In thisreaches
gas cloud case, the
theminimum volume
ceiling height of gas
H, (the Vmin
level Vup
of =the for registration
sensors). Its mass occurs
is equalwhen
to the
gas cloud reaches the ceiling height H, (the level of the sensors). Its mass is equal to
mg min = Vmin q = 1/3 π rn2 H q (15)
2
When the gas reaches one mg min
of =the
Vmin q = 1/3aπmessage
sensors, rn H q will appear indicating (15)the
presence of gas. The message will appear as a signal on the operator’s monitor, which
When the gas reaches one of the sensors, a message will appear indicating the presence
shows the location of all sensors. The localization of the signals will indicate the location
of gas. The message will appear as a signal on the operator’s monitor, which shows the
of the danger. At the same time, the signals are sent to the automated security system. In
location of all sensors. The localization of the signals will indicate the location of the danger.
this system, the coordinates of the sensors that gave the signals are used to calculate the
At the same time, the signals are sent to the automated security system. In this system, the
coordinates of
coordinates of the
the sensors
locationthatof a gave
possible gas source,
the signals and measures
are used arethe
to calculate taken to prevent
coordinates of the
the
threat or eliminate the fire.
location of a possible gas source, and measures are taken to prevent the threat or eliminate
As the distance between sensors decreases, the number of sensors increases, and
the fire.
therefore,
As the the cost of the
distance safetysensors
between system decreases,
increases. At thethe same time,
number the volume
of sensors of flam‐
increases, and
mable gas that will have time to spread before it is registered is reduced.
therefore, the cost of the safety system increases. At the same time, the volume of flammable
Let will
gas that us consider
have time thetoXspread
and Ybefore
coordinates of the placement
it is registered is reduced.of sensors under the ceil‐
ing, and
Let us consider the X and Y coordinates of the placement of filling
the dependence of the volume of gas on the density of sensorsbyunder
the sensors. Let
the ceiling,
the room have dimensions T x on the X axis and Ty on the Y axis; i.e., the area of the room S
and the dependence of the volume of gas on the density of filling by the sensors. Let the
is equal to
𝑆 𝑇 𝑇.
Sensors 2023, 23, 8349 12 of 17

room have dimensions Tx on the X axis and Ty on the Y axis; i.e., the area of the room S is
equal to
S = Tx × Ty .
Let us calculate the coordinates of the sensors at a hexagonal location. The distance
between the sensors is A. The sensor is located in the centre of a regular hexagon with
side C. Let us start placing the sensors from the origin located in the lower left corner of
the room.
The coordinates of the first sensor D11 , located in the first row, are equal to
√
3 A
D11 : x11 = C = A ; y11 = ; (16)
3 2
The coordinates of the second sensor in the first row D12 are
√
3 A
D12 : x12 = x11 + 3C = C + 3C = 4A ; y12 = (17)
3 2
The coordinates of the n-th sensor D1n in the first row are
√ A
D1n : x1n = x1,n−1 + 3C = x1,n−1 + A 3; y1,n−1 = (18)
2
The first row is filled, and it is necessary to consider the possibility and necessity of
filling the remainder to the wall of the room, which is provided by

x1,n−1 ≤ Tx . (19)

The coordinates of the first sensor D12 , located in the second row, are
√
3 5 3
D12 : x21 = C + C = A ; y21 = A (20)
2 2 3
The coordinates of the n2 -th sensor D2n , located in the second row, are
√
D2n : x2,n = x2,n−1 + 3C = x2,n−1 + A 3; y2,n = A (21)

The second row is filled, and it is necessary to consider the need to fill the space up to
the wall of the room, which is provided by

A
y2,n−1 + ≤ Ty . (22)
2
Having filled the rows of sensors to the right and up, we solve the issues of closing the
areas adjacent to the walls with sensors. Practice shows that no special control is required
at the borders. The corners and walls direct the gas to the central part of the protected
room, where the sensors are located.
The sum of the intervals between the sensors in each row, taking into account the
initial distance, should not exceed the length of the room.

Tx Tx
dx = 1 + = 1+ √ . (23)
3C A 3

The number of such rows is dy = 2Ty /A, where A/2 is the distance between the rows.
Thus, it is possible to obtain an estimate of the total number of sensors d
 
Tx 2Ty
d = dx × dy = 1 + √ . (24)
A 3 A
Sensors 2023, 23, 8349 13 of 17

The considered example demonstrates the application of the developed method of gas
sensor location to ensure the fire safety of industrial premises.

3. Results and Discussion

According to the developed method, practical calculations of examples of choosing the
location
Sensors 2023, 23, x FOR PEER REVIEW of sensors and determining their number were carried out. The calculations 13 were
of 18
carried out using a program for working with Microsoft Excel spreadsheets. Considering
the choice of the location of the sensors depending on the permissible volume of gas in the
protected room,inthe
As shown maximum
Table allowable
1, the volume volume
of the of gaswhen
gas cloud in thethe
room is denoted
height of 4 m is Vm .
byreached
As shown in Table 1, the volume of the gas cloud when the height of 4
will be Vm = 67.02 m .3 The radius of the base of the cone, with the assumptions made, will
3 m is reached
will
be be V
4 m. m = volume
This 67.02 m can
. The
be radius of the
taken as the maximum
base of the permissible
cone, with the assumptions
value, made,limit
since at lower will
be 4 m. This volume can be taken as the
values, the gas will not yet reach the sensors. maximum permissible value, since at lower limit
values, the gas will not yet reach the sensors.
Let the dimensions of the room be 20 m by 10 m: Tx = 20 m, Ty = 10 m; the height of
Let the dimensions of the room be 20 m by 10 m: Tx = 20 m, Ty = 10 m; the height of
the room is H = 3 m.
the room is H = 3 m.
Under these conditions, according to Formulas (16)–(24) and Figure 4, the location of
Under these conditions, according to Formulas (16)–(24) and Figure 4, the location of
the sensors is determined. This arrangement will provide full coverage of the surface of
the sensors is determined. This arrangement will provide full coverage of the surface of
the room, and provides control of potential gas emissions. If rn = 3 m, then, according to
the room, and provides control of potential gas emissions. If rn = 3 m, then, according to
Formula (13), A = 5.2 m. Let the sensors be located no further than 4.34 m from each other.
Formula (13), A = 5.2 m. Let the sensors be located no further than 4.34 m from each other.
Let us calculate the coordinates of the sensors and their number.
Let us calculate the coordinates of the sensors and their number.

Figure
Figure4.
4.Hexagonal
Hexagonalarrangement
arrangementof
ofsensors
sensorswith
with the
the distance
distance between
between them
them A
A1.1 .

The coordinates
The coordinates of
of the
the first
first sensor D1111, ,located
sensor D locatedin
inthe
thefirst
firstrow,
row,are
areequal
equal to
to

D11: 𝑥 𝐶 2.5; y A 2.17 (25)

D11 : x11 = C = 2.5; y11 = = 2.17 (25)
2
The coordinates of the second sensor in the first row D12 are
The coordinates of the second sensor in the first row D12 are
D12: 𝑥 𝑥 3𝐶 4 𝐶 10; y 2.17 (26)
A
The coordinatesDof : x12
12 the = xsensor
third 11 + 3CD= 4 C = 10; y12 = are = 2.17
13 in the first row 2
(26)

D13: 𝑥 𝑥 3𝐶 7𝐶 17.5 ; y 2.17 (27)

Since Tx − x13 = 20 − 17.5 = 2.5, which is less than 3C = 7.5, i.e., the step of placement of
sensors, the first row is filled. You can consider the need to fill the remainder to the wall
of the room. There are three sensors here.
The coordinates of the first sensor D21, located in the second row, are equal to
Sensors 2023, 23, 8349 14 of 17

The coordinates of the third sensor D13 in the first row are

A
D13 : x13 = x2 + 3C = 7C = 17.5; y13 = = 2.17 (27)
2
Since Tx − x13 = 20 − 17.5 = 2.5, which is less than 3C = 7.5, i.e., the step of placement
of sensors, the first row is filled. You can consider the need to fill the remainder to the wall
of the room. There are three sensors here.
The coordinates of the first sensor D21 , located in the second row, are equal to

3 5
D21 : x21 = C + C = C = 6.25; y21 = A = 4.34 (28)
2 2
The coordinates of the 2nd sensor D22 , located in the second row, are

D22 : x22 = x21 + 3C = 13.75, y22 = A = 4.34 (29)

Since Tx − x22 = 20 − 13.75 = 6.25, which is less than 3C = 7.5, i.e., the step of placement
of sensors, the second row is also filled, and it is necessary to consider the need to fill the
remainder to the wall of the room. There are two sensors here.
The coordinates of the sensors in the third row differ only vertically, i.e., along the
y-axis. The coordinates of the first sensor D31 , located in the third row, are equal to

3
D31 : x31 = C = 2.5; y31 = A = 2.17 + 4.34 = 6.51; (30)
2
The coordinates of the second sensor in the third row D32 are equal to

3
D32 : x32 = x31 + 3L = 4 C = 10; y32 = A = 6.51 (31)
2
The coordinates of the third sensor D33 in the third row are

3
D33 : x33 = x2 + 3C = 7 C = 17.5; y33 = A = 6.51 (32)
2
Since Tx − x33 = 20 − 17.5 = 2.5, which is less than 3C = 7.5, i.e., the step of placement
of sensors, the third row is filled. You can consider the need to fill the remainder to the wall
of the room. There are three sensors here.
The coordinates of the sensors in the fourth row differ only along the y-axis. The
coordinates of the first sensor D41 , located in the fourth row, are equal to

3 5
D41 : x41 = C + C = C = 6.25; y41 = 2A = 8.68 (33)
2 2
The coordinates of the second sensor D42 , located in the fourth row, are equal to

D42 : x42 = x41 + 3C = 13.75, y42 = 2A = 8.68 (34)

Since Tx − x42 = 20 − 13.75 = 6.25, which is less than 3C = 7.5, i.e., the step of placement
of sensors, the fourth row is also filled, and it is necessary to consider the need to fill the
remainder to the wall of the room. There are two sensors here.
Since Ty − y42 = 10 − 8.68 = 1.32, which is less than A/2 = 2.17, i.e., the step of
placement of sensors along the y-axis, the ceiling surface is filled and the problem is solved.
In principle, it is possible to consider the need to apply sensors to the rest of the ceiling
and to the wall of the room.
A
y2,n−1 + ≤ Ty . (35)
2
Having filled the rows of sensors to the right and up, until the boundaries of the room
are reached, we solve the issues of closing the areas adjacent to the walls with sensors.
Sensors 2023, 23, 8349 15 of 17

The number of sensors in a room is determined by the size of the room and the distance
between the sensors. The number of sensors in two rows of three and in two rows of two
equal a total of 10 detector sensors needed.
Sensors 2023, 23, x FOR PEER REVIEW The arrangement of sensors in four rows on the ceiling of the room is schematically
15 of 18
shown in Figure 5. This can be considered as a framework that, if necessary, can be
completely moved to give the system greater symmetry.

Figure 5.
Figure 5. Location
Location of
of sensors
sensors on
on the
the ceiling
ceiling of
of aa room
room 20
20 m
m long,
long, 10
10 m
m wide,
wide, with
with aa height
height of
of 33 m.
m.
Overhead view.
Overhead view.

The number of sensors in the room is determined by the size of the room, the the per‐
per-
missible volume of hazardous gas, and and thethe sensitivity
sensitivity of of the
the sensors.
sensors. These
These parameters
parameters
determine the distancedistance between
between the the sensors
sensors to to the
the greatest
greatest extent.
extent.
The considered example represents the rational placement of sensors that register
the appearance
appearance of of aadangerous
dangerousgas. gas.The Thedeveloped
developed method
method makes
makes it possible
it possible to adjust the
to adjust
placement
the placement of sensors
of sensors depending
depending ononthethemaximum
maximumpermissiblepermissiblevolume volumeof of gas
gas and the
sensitivity of the sensorssensors used.used. TheThe method
method makes makes it possible to control the layout of
sensors depending
depending on the volume of protectedpremises,
on the volume of protected premises, types
types of of
production,
production, thethe
creation
crea‐
of new
tion sensors,
of new sensors,changes
changesin the
in composition
the composition of hazardous
of hazardous gasgasin different
in different technological
technolog‐
processes.
ical processes.
Next,
Next, itit isisnecessary
necessarytotocalculate
calculatethe localization
the localization (coordinates
(coordinates of aofpossible
a possible place) of the
place) of
gas emission
the gas emission according
according to the coordinates
to the coordinates of the
of thesensors
sensors that signal
that signal thetheregistration
registration ofofa
hazardous substance in the room. This method provides a calculation
a hazardous substance in the room. This method provides a calculation of the coordinates of the coordinates of
the gasgas
of the source
source fromfromthe the
coordinates
coordinatesof theoftriggered
the triggered sensors. Knowing
sensors. Knowingthe coordinates
the coordinatesof all
sensors, the coordinates of the sensors that gave the alarm, and
of all sensors, the coordinates of the sensors that gave the alarm, and the formula for de‐ the formula for determining
the volumethe
termining of the gas cloud
volume of theunder theseunder
gas cloud sensors, theseit issensors,
possibleit to is assess
possible thetolocalization
assess the lo‐of
the source of
calization of thehazardous
source of gas in the protected
hazardous gas in the premises.
protected This increasesThis
premises. the increases
efficiency theof the
ef‐
fire safety
ficiency ofsystem.
the fire safety system.
The
The method
methodprovides providesthethe calculation
calculation of the
of rational
the rationalplacement
placementof sensors when chang-
of sensors when
ing conditions: the height and size of the room, increasing
changing conditions: the height and size of the room, increasing the sensitivity the sensitivity of the sensors,
of the
the
requirements for the maximum
sensors, the requirements for the permissible
maximumvolumes permissible of recorded
volumesgases, etc. gases, etc.
of recorded
The
The considered
considered method method assumes
assumes that that ideal
ideal conditions
conditions are are created
created in in the
the room
room for
for
monitoring
monitoring the the airair without
without the the influence
influence of of wind
wind caused
caused by by the
the fan
fan (air
(air conditioner).
conditioner). The The
force of the
force of the wind
wind (ventilation)
(ventilation)creates
createsthe themovement
movementofof airairandandgasgas in the
in the horizontal
horizontal di‐
direction. The lifting force acts on the gas (of the type in question) vertically. The forces act
rection. The lifting force acts on the gas (of the type in question) vertically. The forces act
independently, so the gas molecules will rise to the ceiling, thus, to the sensors. The lifting
time will not change significantly, but the place where the ceiling is reached will change.
The problem can be solved by entering into the calculation a correction factor that
shifts the “gas cone” and the layout of the sensors in the direction of the wind. At a cer‐
Sensors 2023, 23, 8349 16 of 17

independently, so the gas molecules will rise to the ceiling, thus, to the sensors. The lifting
time will not change significantly, but the place where the ceiling is reached will change.
The problem can be solved by entering into the calculation a correction factor that
shifts the “gas cone” and the layout of the sensors in the direction of the wind. At a certain
high-air-flow rate, gas molecules can be outside the room before they reach the ceiling.
Then, you need to add sensors at the outlet of the air flow from the room.
Production rooms of a large size and complex configuration can be divided into small
rooms. For each part, apply the developed method of sensor placement. The problem is to
connect the boundaries of the individual parts. To do this, you can apply the method of
network models, for which the author has developed algorithms for calculating in parts,
presented in [24,25]. This will require, of course, additional work, such as the creation of a
network model of this subject area. In the future, it is planned to conduct such a study.

4. Conclusions
One of the important problems of maintaining the operation of industrial enterprises
is to ensure fire safety. It is important to detect a potential fire before ignition, at the
stage of when deviation beings from the normal technological process. Indicators of an
emergency situation are various gases, an increase in the concentration of which indicates
the beginning of a dangerous process. The creation and use of gas sensors makes it possible
to detect small changes in the concentrations of hazardous gases. The article presents
a mathematical method for choosing the location and number of sensors that solve the
problem of gas control. The natural physical approach will change from the standpoint
of analyzing the gas flow from the source to the locations of sensors in the premises
of enterprises.
The developed mathematical model of the rational location of gas detectors on H2
and CO for early detection of fire (accident) is presented. The model-based method for
calculating the location of gas analyzers makes it possible to diagnose and predict fire
hazard at technological facilities based on gas control, and compare the obtained values
with the maximum permissible values. The calculation of the conditions for the rational
placement of gas detectors for different heights of technological premises is carried out.
The method made it possible to determine the rational placement of sensors that
ensure timely registration of gas, taking into account economic feasibility. The resulting
solution can be algorithmically converted for use in other conditions, when changing the
size, height, shape of the room, and when changing the norms of permissible volumes of
hazardous gas. The method provides a change in decisions with the appearance of more
sensitive sensors. The method is also applicable for technologies where other types of gases
are indicators. It is planned to develop a mathematical program for an automated fire
protection control system with an early fire detection function based on H2 and CO gas
control in technological design.

Funding: This research received no external funding.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.
Conflicts of Interest: The author declares that there are no known competing financial interest or
personal relationship that could affect the work presented in this paper.

References
1. Bywater, D. Detection of Real Fires by Carbon Monoxide Detectors—Foreign Experience. The Results of 10 Years of Research
Lead to a Leap in Fire Detection Technology. Available online: https://www.aktivsb.ru/statii/obnaruzhenie_realnykh_pozharov
_detektorami_ugarnogo_gaza_zarubezhnyy_opyt.html (accessed on 5 October 2023).
2. Gas Detector. Available online: https://en.wikipedia.org/wiki/Gas_detector (accessed on 5 October 2023).
3. Bakanov, V. Fire detectors with a gas sensor in the light of current regulatory requirements. Prot. Technol. 2014, 4, 71–78.
Sensors 2023, 23, 8349 17 of 17

4. EN 50291-1 2010-04 + A1 2012-06; Electrical Apparatus for the Detection of Carbon Monoxide in Domestic Premises—Part 1: Test
Methods and Performance Requirements. iTeh: Etobicoke, ON, Canada, 2012.
5. Fedorov, A.; Bytcinskaya, T.; Lukyanchenko, A.; Hung, T.D. Trends in the development of automatic fire detectors. Technol.
Technosphere Saf. 2009, 23, 111–114. Available online: https://cyberleninka.ru/article/n/tendentsii-razvitiya-avtomaticheskih-p
ozharnyh-izveschateley-1 (accessed on 5 October 2023).
6. Fire Alarm Systems and Automation of Fire Protection Systems. Designing and Regulations Rules. CΠ 484.1311500.2020.
Available online: https://docs.cntd.ru/document/566249686 (accessed on 5 October 2023).
7. Siting Gas Detectors. Available online: https://www.internationalgasdetectors.com/wp-content/uploads/2016/04/SITING-GA
S-DETECTORS-BS023.pdf (accessed on 5 October 2023).
8. Gas Detector Handbook. Available online: https://hsseworld.com/wp-content/uploads/2022/01/Gas-Detection-Handbook.pd
f (accessed on 5 October 2023).
9. Fixed Gas Detector. Available online: https://www.safetygas.com/media/productattach/e/n/en-gas_and_flame_detection_gaz
detect_2021.ed01-web.pdf (accessed on 5 October 2023).
10. Xgard Gas Detectors. Available online: https://evikontroll.ee/pdf/crowcon/Xgard_UM_EN.pdf (accessed on 5 October 2023).
11. AP 1027: Cleaning and Sanitizing Portable Gas Detectors. Available online: https://www.gfgsafety.com/fileadmin/templates/i
mg/GfG-Branches/GfG-USA/Application_Notes/AP1027_cleaning_and_sanitizing_portable_gas_detectors_14_MAY_20_lo
wer_res.pdf (accessed on 5 October 2023).
12. Available online: https://www.canarysense.com/pdfs/resources/brochures/Honeywell-Gas-Detection-Book2.pdf (accessed on
5 October 2023).
13. Guidelines for Placing Sensors on Flammable Gases. Available online: https://www.seitron.ru/assets/docs/dokumenty/razme
shenie%20sensorov.pdf (accessed on 5 October 2023).
14. Basic Guidelines for Selection of Fire and Gas Detectors Inst Tools. Fire & Gas System. Available online: https://instrumentationt
ools.com/selection-of-fire-and-gas-detectors (accessed on 5 October 2023).
15. Brattery, A.; Gavelly, F.; Hansen, O.R.; Scott, S.G. Using CFD to analyze gas detector location in process facilities. In Proceedings
of the Mary Kay O’Connor Process Safety Center Symposium 2011, College Station, TX, USA, 25–27 October 2011.
16. Okeke, R.O.; Ehikhamenle, M. Design and simulation of gas and fire detector and alarm system with water sprinkle. Int. J. Eng.
Res. Gen. Sci. 2017, 5, 216–225.
17. Zhen, T.; Klise, K.A.; Cunningham, S.; Marszal, E.; Laird, C.D. A mathematical programming approach for the optimal placement
of flame detectors in petrochemical facilities. Process Saf. Environ. Prot. 2019, 132, 47–58. [CrossRef]
18. Güllüce, Y.; Çelik, R.N. FireAnalyst: An effective system for detecting fire geolocation and fire behavior in forests using
mathematical modeling. Turk. J. Agric. For. 2020, 44, 127–139. [CrossRef]
19. Kaliyev, D.I.; Shvets, O.Y. Convolutional neural networks for solving fire detection problems based on aerial photography.
Program Syst. Theory Appl. 2022, 13, 195–213. (In Russian)
20. Zhdanova, A.O.; Kopylov, N.P.; Kropotova, S.S.; Kuznetsov, G.V. Characteristics of a Typical Indoor Seat of Fire. J. Eng. Phys.
Thermophy 2023, 96, 143–149. [CrossRef]
21. Gupta, A.K.; Kumar, R.; Yadav, P.K.; Naveen, M. Fire Safety through Mathematical Modeling. Available online: https://www.re
searchgate.net/publication/228763095_Fire_safety_through_mathematical_modelling (accessed on 5 October 2023).
22. Lukyanchenko, A.A.; Teterin, N.M.; Topolskii, N.G.; Lukyanchenko, A.A.; Fedorov, A.V. Integrated automated system for early
fire detection and environmental monitoring “Kassandra”. Nauchn. Tech. Catalog 2014, 52.
23. Yutyaev, A.E.; Iakunchikov, E.N.; Oganesyan, A.S.; Agafonov, V.V. Evaluation of design solutions and technological systems of
coal mines taking into account the risk. Ugol 2019, 7, 52–57. [CrossRef]
24. Petrov, A.E.; Fedorov, A.V.; Kochegarov, A.V.; Lomaev, E.N.; Preobrazhenskiy, A.P. The Analysis of Network Models for the
Design of Industrial and Fire Safety Systems for Oil Refineries. IOP Conf. Ser. Earth Environ. Sci. 2021, 808, 012024. [CrossRef]
25. Petrov, A.E. Tensor Method and Dual Networks in Electrical Engineering. Russ. Electr. Eng. 2008, 79, 645–654. [CrossRef]
26. Luk’yanchenko, A.A.; Petrov, A.E.; Fedorov, A.V.; Denisov, A.N. Method of Rational Location of Gas Sensors for Early Fire
Detection Based On Gas Control Technology. J. Adv. Res. Dyn. Control. Syst. 2020, 12, 1293–1306. [CrossRef]
27. Petrov, A.; Fedorov, A.; Mintsaev, M.; Ilyukhin, A.; Marsov, V. Theoretical Foundations and Methods for the Rational Location of
Gas Fire Sensors Based on Gas Control Technology. Mathematical Modelling of Gas Fire Sensors Location for Early Fire Detection.
In Proceedings of the XV International Scientific Conference “INTERAGROMASH 2022”, Rostov-on-Don, Russia, 25–27 May 2022;
Lecture Notes in Networks and Systems; Beskopylny, A., Shamtsyan, M., Artiukh, V., Eds.; Springer: Cham, Switzerland, 2023;
Volume 574, pp. 1658–1667. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.