Figure 1 : The six components that make up a smart city ( Liu & Zhang , 2021 ).
concentrations reflected only part of the entire dispersion behaviour . To completely reproduce the atmospheric dispersion of gases , a very large number of sensors will be required .
Sensors are already essential in various industries . Industrial plants employ sensors for process control , monitoring , and safety . They measure and detect changes in temperature , concentration , pressure , and other relevant parameters , during operation and transmit the information to controllers for regulation of process conditions or to alarm systems for indication of any abnormal conditions and problems . In the event of a hazardous gas leak , a sensor network is expected to detect the gas as fast as possible for a quick and correct emergency response to avoid secondary damages .
Sensors play a central role in smart systems as they provide the necessary information for systems to make decisions . The coupling of sensors and IoT is often highlighted , but their optimal placement is neglected . The sensors are normally placed near the known source location . The challenge arises in the case of sensor placement for an unknown source and the need to reduce the number of
sensors . Optimal placement and number of sensors are two important design aspects that need to be considered when deploying sensors for hazardous gas monitoring and detection . An increase in the number of gas sensors will improve hazardous gas detection performance in a particular plant . However , an over abundance of sensors may lead to higher replacement and maintenance costs as well as an increase in the number of false alarms .
The performance of a sensor network is also highly dependent on the location of the sensors . Poor placement of sensors can lead to false alarms or blind spots , where some areas are not properly monitored .
The standard approach to sensor layout placement is via qualitative methods , typically arising from engineering experience , which can be subjective and biased . On the other hand , quantitative methods rely on quantifiable values , which are objective and measurable . For hazardous gas detection , quantitative values include concentration and volume of gas . Quantitative models organise and interpret the measured datasets using mathematical equations
27