A Novel Neuro Simulated Annealing Algorithm for Detecting Proportion of Component Gases in Manhole Gas Mixture

Abstract

In present article we are exploring the design issues in development of an intelligent gas recognizer for detecting proportion of component gases in manhole gas mixture. Principally, the gas components found in manhole gas mixture are, Ammonia (NH3), Carbon Dioxide (CO2), Carbon Monoxide (CO), Hydrogen Sulfide (H2S), Methane (CH4), and Nitrogen Oxide (NOx). These gases are harmful for human health. We are focusing on the development of an intelligent sensory system which can detect the extent poisonous gases found in manhole gas mixture. A gas sensor array is used for this purpose. Sensor responses are cross-sensitive, because multiple gas sensors are simultaneously used to detect multiple gases. The cross-sensitivity is an overlapping effect of one gas on sensor of another, inducing thereby difficulty in sensing mechanism all together. We resort to artificial neural network (ANN) and simulated annealing (SA) algorithm for the development intelligent sensory system. The SA algorithm is used to search out optimized combination of synaptic weights for the ANN trained for sensing proportion of constituent gases.