Linear regression based statistical approach for detecting proportion of component gases in manhole gas mixture

Abstract

The present article proposes the issues in designing an intelligent recognizer for detecting proportion of component gases in manhole gas mixture. The major components found in manhole gas mixture are Hydrogen Sulfide (H2S), Ammonia (NH3), Methane (CH4), Carbon Dioxide (CO2), Nitrogen Oxide (NOx), and Carbon Monoxide (CO). The manhole gas is formed after the decomposition of waste products, domestic garbage etc. into the sewer pipelines which are built for exhausting these waste products out of our cities and towns. The manholes are built across these pipelines for cleaning purpose. Thus safety for the people working in this field is a matter of concern because all the above mentioned gases are harmful gases and they are potent to loss of human lives. Also detection of these gas components is of primary concern today as because a short exposure of these components with human physiology results endanger to their lives. So our focus is on developing an intelligent gas recognition system which can recognize multiple gases simultaneously. A gas sensor array is an array of sensors, consisting of two or more electrical type semiconductor gas sensors. Response of electrical type semiconductor gas sensors in presence of gases are either the change in resistance or change in voltage of the sensor. At an instant a gas sensor array contains as many sensors as many individual gases we are targeting to detect. Use of multiple gas sensors and presence of multiple gases together results cross-sensitivity. The cross-sensitivity is an overlapping effect of one gas on another sensor. We adopt linear regression based statistical approach to deal with issues of simultaneous detection of multiple gases notwithstanding cross-sensitivity issue.