A smart real-time system for autonomous detection and prevention of LPG gas leaks with theoretical validation

A gas leak sensor usually only sends a warning signal, leading to a delay in treatment and increased danger. The proposed study aims to develop an autonomous liquefied petroleum gas (LPG) gas leak detection system along with containment, using an MQ-5 sensor, Arduino Uno, and a stepper motor-based shut-off valve. The proposed system can detect a gas leak in 2.3 seconds and contain it in 3.1 seconds, with a high accuracy rate of 97.5%. The proposed system also offers a major advantage over other gas leak detection systems that only rely on global system for mobile communication (GSM) or internet of things (IoT) based warning messages by using adaptive control with a proportional-integral-derivative (PID) controller, a Laplace transform model, and exponential gas diffusion analysis. The proposed system also includes a feature to sense a smoke alarm and power failure, along with abnormality detection using artificial intelligence (AI)-based analysis. The proposed system is also efficient in terms of power consumption, with an average power consumption of only 2.1 W. The proposed system thus enhances a shift from passive detection to active containment in a smart system, thereby contributing to the domain of electrical and control engineering.