Electronic Nose Based on Gas Sensor Array and Neural Network for Indoor Hydrogen Gas Control System

Hydrogen gas leaks in a closed room can pose a fire hazard and poor air quality. The concentration of hydrogen gas of 4-75% in the air is highly flammable and can cause explosions. An electronic nose system consisting of gas sensors array and a neural network has been built to detect hydrogen gas leaks in a room. Data from each sensor is used as input for the classification of gases on the neural network. Proportional-integral-derivative (PID) method is applied to control the exhaust fan to eliminate hydrogen gas leaks in the room. The electronic nose and PID control are implemented on the Arduino Nano microcontroller. The experiment results showed that this system could classify several gases such as hydrogen gas, vehicle smoke, and perfume with a success rate of 86.67%. The PID control becomes active when hydrogen gas with concentrations above 100 ppm has been classified. These results can minimize and prevent hydrogen gas leaks and maintain good indoor air quality.