Aditi Sinha

Smart Approach to Fire Prevention and Detection

A cool system that warns us of anomalous temperature, detects fire, sends alert signals using sound and light, notifies us on our phone.

IntermediateFull instructions provided2 hours520
Smart Approach to Fire Prevention and Detection

Things used in this project


Read more


Backend Code for Fire Prevention and Detection System

#import the necessary libraries
from boltiot import Bolt, Sms
import json, time, math, statistics

#store all credentials related to Twilio and Bolt within double quotes
sid = "" # sid given on Twilio dashboard
auth_token = "" # authorization token given on Twilio dashboard
From = "" #phone number given by Twilio
To = "" #your mobile number that receives notification  
a = "XXXXXXXXXXXXXXXXX" # your api_key given on Bolt Cloud dashboard
d = "BOLTXXXXXX" # your device_id given on Bolt Cloud dashboard

maxim = 300 #probable analog equivalent of temperature at which fire starts
frame_size = 6 #used for Z-score analysis
mul_factor = 3 #used for Z-score analysis

mb = Bolt(a,d) #create object of Bolt class
sms = Sms(sid,auth_token,From,To) #create object of Sms class
prev_data=[] #stores sensor values

#function to calculate bounds for the temperature sensor using Z-score
def cal_bounds(prev_data,frame_size,mul_factor):
  #check if enough data is collected for calculation of Z-score
  if len(prev_data)<frame_size :
    return None
  #deletion of older data
  if len(prev_data)>frame_size :
   	del prev_data[0:len(prev_data)-frame_size]
  mean=statistics.mean(prev_data) #calculate mean
  for i in prev_data :
    var+= math.pow((i-mean),2) #calculate variance
  Zn = mul_factor * math.sqrt(var/ frame_size) #calculate Z-score
  Hb= prev_data[frame_size-1]+Zn #calculate upper bound
  Lb = prev_data[frame_size-1]-Zn #calculate lower bound
  return [Hb,Lb]

#iterate over a loop to sense temperature in theinterval of 10s
while True:
  print("\nReading sensor value")
  response = mb.analogRead('A0') #read values of A0 analog pin 
  data = json.loads(response) #convert json data to python object
  if data['success'] != 1:
    print("Error in the data collected: "+data['value']) #print error
  print("Sensor value = "+str(data['value']))
  bounds = cal_bounds(prev_data,frame_size,mul_factor) #create object of cal_bounds
    s_val = int(data['value'])	
    #check if sensor value exceeds the temperature at which fire starts
    if s_val>=maxim:
      print("Starting Alarm!")
      print("Alarm is ON "+mb.digitalWrite('0','HIGH')) #Buzzer rings
      print("Led is ON "+mb.digitalWrite('1','HIGH')) #Led glows
      sms.send_sms("Sensor value is " + str(s_val)) #send Twilio alert message  
      print("Alarm is ON "+mb.digitalWrite('0','LOW')) #Buzzer is silent
      print("Led is ON "+mb.digitalWrite('1','LOW')) #Led doesn't glow
    #check if enough data has been collected
    if not bounds:
      print("Need ",(frame_size-len(prev_data))," more data points for Zscore")
      #check if sensor value is within the bounds computed by Z-score 
      if s_val > bounds[0] :
        print("Temperature increased suddenly")
        #notify anomaly using Twilio
        sms.send_sms("Sensor value increased suddenly to " + str(s_val))
    except Exception as e:
      print("Error in detecting anomaly: ",e) #print error
  except Exception as e:
    print("Error: ",e) #print error


Aditi Sinha

Aditi Sinha

1 project • 1 follower
A coding enthusiast and an avid learner. I use the web occassionally to enhance my knowledge in Machine Learning, Deep Learning and IoT.