An Examination of Gas Leakage Judgment Methods with Semiconductor Gas Sensors
The authors are conducting a study in order to reduce accidents caused by gas leaks, including poisoning, explosions and fires. As countermeasures to prevent accidents caused by gas leakage, up to now the authors have studied methods to improve the gas selectivity of semicouductor gas sensors used in the detection section of gas leak alarm devices. As a result, they have proposed a method for distinguishing the gas type by using a single semiconductor gas sensor and controlling the sawtooth wave configuration of the heater voltage in the gas sensor in order to distinguish the type of gas from trends in the output voltage. This paper proposes a method to carry out judgments on whether there is a gas leak or not by applying a two-stage type hierarchical neural network for the output response of the gas sensor as obtained by carrying out control of the sawtooth wave configuration for the heater voltage of a semiconductor gas sensor. Using this method, it was possible to carry out gas leak judgments with few false alarms.
gas leakage semiconductor gas sensors neural network gas-lcak alarm fire explosion
NAKAJIMA Takashi ONO Takashi MIURA Hikaru
Junior College,Nihon University,7-24-1Narashinodai,Funabashi,Chiba,274-8501,Japan College of Science and Technology,Nihon University,1-8-14,Kanda Surugaclai,Chiyodaku,Tokyo,101-8308,
国际会议
The 2008 International Symposium on Safety Science and Technology(2008年安全科学技术国际会议)
北京
英文
1135-1140
2008-09-24（万方平台首次上网日期，不代表论文的发表时间）