- Author
- Nakajima, T. | Miyagawa, I.
- Title
- Discrimination Method Among Various Kinds of Single-Component and Mixed Gases.
- Coporate
- Junior College of Nihon Univ., Chiba, Japan Japan Industrial Safety and Health Consultant Assoc., Tokyo, Japan
- Journal
- Bulletin of Japanese Association of Fire Science and Engineering, Vol. 41, No. 1, 29-40, 1992
- Keywords
- gas mixtures | semiconductor devices | alcohol
- Identifiers
- gas sensors; gas leak alarm; characteristics curves
- Abstract
- Semiconductor gas sensors are said to lack the ability to discriminate among gas kinds, and consequently, it is regarded quite difficult for a single gas sensor unit to discriminate among different gas kinds. Therefore, when a semiconductor gas sensor is used with a gas leak alarm, gas containing alcohol causes a false alarm. In order to improve gas-king discrimination, it is necessary to use several semiconductor gas sensor units, each sensing different kinds of gas. As for single-component gases, discrimination is enabled by processing the output voltage of these sensor units. However, an effective method for discriminating among mixed gas kinds containing alcohol, such as between one gas containing alcohol and methane and another alcohol and propane, has not been established yet. This is because alcohol has a higher sensitivity than methane or propane; a mixed gas containing alcohol shows characteristics similar to that of single-component alcohol gases. In this paper, we explain a method that compare the differences of the characteristics curves of various mixed gas kinds. The characteristics curves is the output voltage of the semiconductor gas sensor by control of the heater voltage of the sensor. Even mixed gases that contain the same gas components, the characteristics curves of mixed gases varies with the mixture ratio. Thus, by processing several patterns, discrimination becomes possible. Furthermore, when comparing the characteristics curves of mixed gases that resemble each other, e.g., one containing alcohol and methane and another alcohol and propane, we have worked out a method to enlarge the distinctive portions of the characteristics curves in order to define the gas kind. As a result, this discrimination method has been successful and is highly reliable.