Gas leaks in Lima and Ica (Peru) increase every year, causing accidents and irreparable damage to the population. In this article, a controlled leak was produced using a two-burner kitchen and an array of MQ-6 sensors positioned at different angles (45°, 0° and 30°) with respect to the kitchen. The results show that, if the kitchen is in a high position (87 cm), the detected concentration is lower, but the detection is faster (6,419 s) if the arrangement is located 50 cm from the origin of the leak. The detection time is between 13,515 s and 21,740 s and the maximum concentration detected is 98 ppm. The best adapted learning model is Support Vector Machine, with an RMSE of 4,61 ppm. It is concluded that the best position for gas detection was at a height of 47 cm above the ground, at 50 cm from the sensor and at an angle of 0°. The detection time is 13,84 s. Finally, it is concluded ...

Gas leaks in Lima and Ica (Peru) increase every year, causing accidents and irreparable damage to the population. In this article, a controlled leak was produced using a two-burner kitchen and an array of MQ-6 sensors positioned at different angles (45°, 0° and 30°) with respect to the kitchen. The results show that, if the kitchen is in a high position (87 cm), the detected concentration is lower, but the detection is faster (6,419 s) if the arrangement is located 50 cm from the origin of the leak. The detection time is between 13,515 s and 21,740 s and the maximum concentration detected is 98 ppm. The best adapted learning model is Support Vector Machine, with an RMSE of 4,61 ppm. It is concluded that the best position for gas detection was at a height of 47 cm above the ground, at 50 cm from the sensor and at an angle of 0°. The detection time is 13,84 s. Finally, it is concluded ...

Gas leaks in Lima and Ica (Peru) increase every year, causing accidents and irreparable damage to the population. In this article, a controlled leak was produced using a two-burner kitchen and an array of MQ-6 sensors positioned at different angles (45°, 0° and 30°) with respect to the kitchen. The results show that, if the kitchen is in a high position (87 cm), the detected concentration is lower, but the detection is faster (6.419 s) if the arrangement is located 50 cm from the origin of the leak. The detection time is between 13,515 s and 21,740 s and the maximum concentration detected is 98 ppm. The best adapted learning model is Suport Vector Machine, with an RMSE of 4,61 ppm. It is concluded that the best position for gas detection was at a height of 47 cm above the ground, at 50 cm from the sensor and at an angle of 0°. The detection time is 13.84 s. Finally, it is concluded t...