IOT Based Fire Alarm System

Abstract
Here in this IOT based project we have tried to take the first steps toward making an
Environment monitoring system with its primary objective being able to predict fire hazards
both urban and wildfires based on environmental variables such as temperature , humidity etc.
in real time. The immediate goal of this particular project was to configure an IOT based fire
alert system that would measure the temperature surrounding the sensor and then compare it
with some fixed reference temperature value before it initiates an alarm or notification through
SMS to the concerned parties. Once we integrate other needed sensors such as fire, humidity,
positioning,gas, light etc. and place them in different environments at an acceptable density.
Afteraccumulating extracted data over time in cloud based database. Analyze and integrate
collected data and data trends to formulate mathematical models establishing co-relation
among the collected data and perceiving data trends. After that we could better understand
and speculate the possibilities of occurring natural fire under a given scenario. To that end we
need keep the individual environmental nodes as cost efficient and as durable as possible and
use them to design as part of small scale contained home/industrial solution where we can
access a dashboard to use and configure predefined services such as notification systems or
preexisting anomalies using the data collected from the nodes of a particular/specific area such
as home or factory.
Introduction
Due to several predicaments of our oil fueled civilization fire hazards have increased tenfold all
over the world. The increasing carbon emission driving the average temperature across the
world ever so higher. Coupled with deforestation, urbanization and extreme weather the
number of hazards fire incidents have increased both in urban environments and forests. The
recent on going wild fire emergency in the amazon the lungs of our planet is a devastating
example. There are a lot of system based on different technologies and a huge man power
working in tandem to extinguish occurring fires both urban and wild. Most of the existing
systems that is deployed to fight this hell hound is fire/smoke alert systems that triggers after
fire has started or about to start. We frankly need to do better to fight this behemoth of a
monster. We need an all engulfing environmental monitoring system and data gathering tool
that can collect the needed environmental data, analyze these data and make practical
decisions both in real time and for reference, based on its collected data.

Problem
We have had fire service for a very long time all over the world. Fire detection, fire extinction,
equipment transportation, are all in most countries based on very outdated technology.
Notification systems are still based on human anticipation, it goes off only after individual
responsible pulls a lever or notifies the fire service. By the time fire services can come to the
site can depend on the time of the day and place as traffic can be an obstacle in densely
populated areas where the scope of fire spreading and causing huge damage both monetary
and of life is larger.
We need to better manage these issues both from a logistic prospective and from time
management standpoint to an acceptable status. A lot of the urban fire causes so much damage
due to the fact that the notification system is slow and the response time is large. With a
digitized system based on the internet we can easily cut this delay significantly with ease.
Existing Solution
The existing Fire Alarm System that is broadly used is not quite digitized. The sensors usually
used are smoke and fire detectors. Notifications systems are pull stations or sound alarms.
Inputs from these various sensors are collected in fire alarm control unit where the data is
monitored. This component, the hub of the system, monitors inputs and system integrity,
controls outputs and relays information. This existing system is an electrical hardware system
that is used in a localized environment with huge deployment cost and is to a point outdated.

Fire Alarm Notification appliance

Fire alarm control unit

Primary power supply used for these fire alarms are usually the non-switched 120 or 240-volt
alternating current source supplied from a commercial power utility and batteries , generators
as backup. This component uses energy supplied from the fire alarm system or other stored
energy source, to inform the proximate persons of the need to take action, usually to evacuate.
This is done by means of a pulsing incandescent light, flashing strobe light, electromechanical
horn, siren, electronic horn, chime, bell, speaker, or a combination of these devices. Strobes are
either made of a xenon tube (most common) or recently LEDs.
Notification Appliances utilize audible, visible, tactile, textual or even olfactory stimuli (odorize)
to alert the occupants of the need to evacuate or take action in the event of a fire or other
emergency. Evacuation signals may consist of simple appliances that transmit encoded
information, coded appliances that transmit a predetermined pattern, and or appliances that
transmit audible and visible textual information such as live or pre-recorded instructions, and
illuminated message displays.

Proposed Solution
I here am proposing an IOT based Smart Fire Alarm system that has the ability to think and
recognize pattern based on experience aka collected data and algorithm that has a high
probability of resulting in unwanted hazardous fire.
Here we are proposing a three component fire alarm system.

1. Sensor Nodes.(Sensor & Actuator)

Depending on the risk we can integrate heat/humidity sensor and use commercial
power supply as power source. We will need to connect these sensors to an Arduino
Uno or a Wi-Fi module so we can send the collected data to our cloud platform using
internet.

2. Controller.
In our cloud platform we will store the data in a relational database. The reading of
temperature and humidity around every single node will be stored with a timestamp, an
identifier for each node preferably denoting their physical position.
Once a set of data is stored we will need to compare them with reference values of
these temperature and humidity readings and based on the result of said comparison
we can trigger different actions. Like calling various predefined APIs or web hooks that
can send SMS to concerned parties or sound alarm or make evacuation announcement
or start pouring/tinkling water in the particular areas.

3. Notification System.(Communicator)
We can design a dashboard for both for smart phones and web browsers that can be
accessed with different assigned roles. That will show a representation of the collected
data and their various analysis like average readings over a certain period of time or
graph of time period of the max temperature in a 24hour / 7days / monthly time period.
Through continuous research and development we can keep adding various analytics
and update these reference values that we are using to compare the collected data.
Design
For a single node design the components we are using are:
1. Arduino UNO board with Wi-Fi module
2. DS18B20digital temperature sensor
3. Resistor4.7k Ohm
4. Breadboard
5. jumper wire

We will need to plug the sensor on the breadboard then connect its pins to the Arduino using
the jumpers in the following order: pin 1 to GND; pin 2 to any digital pin (pin 2 in our case); pin
3 to +5V or +3.3V, at the end put the pull-up resistor.
Simulation
After connecting the components according to the circuit diagram. And connecting lan we can
access the Arduino Uno and will have to upload coding to collect the data from sensor. In this
case we will need to upload two libraries namely 1- Wire bus and Dallas Temperature. And
then using Arduino IDE we can upload the code into the Arduino board.

In above screenshot we can upload the code in Arduino and monitor the reading through its
serial monitor.
Once we start getting the data at an interval of our choosing we can then collect this data and
ship it to cloud and store there.
Once we start storing the data we can then compare them against pre-defined static or
dynamic reference values and make decisions according to pre-defined scenarios.
Result
At the end of it we should have a continuous flow of temperature and humidity reading from
various infrastructure points of an establishment and have the data stored. After collecting data
for a few days or weeks we should easily be able to graph the set of readings normal for this
particular establishment and discern if any of these scenarios from different parts of the 24
hour cycle and from different parts of the establishment is a matter of concern or not. For
example for a shop let’s say there is a place in the back of the shop or in a corner of the store
room the employs go to have a smoke in there braketime. So from the data collected the
sensor nodes around that part will be different than the other parts of the storeroom when
there is people smoking. In this case analysis of collected data helps quantify pre-existing fire
threats and help make appropriate adjustments to counter them.
In another scenario where the preexisting threats have been dealt with so now we have a set
of data that falls into a favorable set ,meaning now we have a data set which we like and want
our future data collected through the sensors resemble that set. So then we can configure our
reference value that is compared with the real time collected data and notify us whenever a
particular set of real time data over a certain time period is too far away from the desired data.
And those can identify anomalies that can be very broad in scope.

Advantages/Application
First of all IOT based fire alarm system compared to what is widely deployed right now will be a
lot faster , in real time . And as this system is temperature based we will primarily detect spike
in temperature rather than fire itself which will enable us to respond to a situation that
emerges prior to a fire rather than fire itself. We should and a fire/smoke sensor in case of
accidental fires to improve our model. So when we observe a temperature anomaly in our
configured infrastructure we can respond to that through an existing monitoring system either
through a security guard or cctv and can detect danger prior to a fire. So it would work as a
prevention system as well as an alert system.
As the notification system is real time that can cut the response time significantly. When the
fire has not spread significantly it gives us better access to that area to deal with the fire and
can be dealt with pre-installed fire hose or fire extinguisher. Even when the fire has spread the
heat signature collected from the data would enable us to identify how the fire spread and thus
give us better insight on how to best deal with it.
Unlike existing solutions as this can be a smart solution, continuous collection and storage of
data would give us insight on better identifying underlying causes on how to best prevent such
scenarios. Through analytics done on the collected data and research we would be able to
better correlate a fire happening to different events and thus would let us configure alert
systems based on a combination of data rather than just increase of temperature. For example
depending where the system is installed like if it has more flammable objects or fire sparks can
occur more often or does not have immediate access to water or fire extinguisher we can
configure our alert system on that so these areas can be prioritized or labeled as high risk at low
temperature increase and deploy needed monitoring system.
Every object has its own heat signature and thus every event would register in the data if it
causes significant heat variation. Once we start collecting data and associate sets of data to
different events we would be able to identify different events based on these collected data
and use it as a future reference. And so with in a margin of error we would be able to identify
these events and have real time data to monitor to look for anomalies that may result in fire.

Limitations
The main limitation to this system would be that we are only considering one sensor per node
namely temperature/humidity. If we add more sensors such as fire/smoke, gas, gps, sound,
camera we would start getting a lot more data to work with and this can become an overall
monitoring system that can be configured to detect almost anything. It can as well become a
policing device based on how it is deployed and used. Increasing sensors in the nodes would
scale up the system and would increase the cost significantly. We would also need to consider
the materials used to create these nodes as they may be subject to harsh environment and thus
need to with stand them. That would also increase cost.
This system feeds on referential data to better analyze its collected data to better associate its
value. For example let’s say there is a party in one infrastructure that has this system installed.
We would need some external data like how big was the party how many people attended this
party and different activities that happened at specific time. Once we co relate these external
data we would be able to associate collected data to these different external variables and thus
better understand these events through collected data for future reference. Once we have
enough observed data we would be able to comment on specific events or their arrangements
to make them safer as well as identify anomalies if it occurs.
Conclusion
Bottom line , we need to modernize our existing fire alarm system and digitized our fire services
as soon as possible to deal with these increasing fire hazards across the planet . Companies
would be wise to invest in such improvements as it would help better ensure the protection of
their properties. Governments should also add emphasis on the modernization of existing
firefighting system and if needed subsidized research if needed in this field. We would be
amazed on how much we can accomplish if we consider IOT to improve our existing fire alarm
system.

Reference

https://create.arduino.cc/projecthub/TheGadgetBoy/ds18b20-digital-temperature-sensor-and-
arduino-9cc806?ref=tag&ref_id=temperature&offset=2

https://en.wikipedia.org/wiki/Fire_alarm_system