Sound Parameter Analysis for Early Detection and Prevention of Home Fire Outbreak

Abstract

Due to tragic losses caused by preventable home fires, it is imperative to have technological advancement toward more fire safety measures. Cooking fires are one of the most prevalent types of house fires, accounting for more of all residential fires; and cooking left unattended, is by far the most common cause of home fires. This paper proposes an early detection of home fire outbreaks by sound parameter analysis. The sounds produced by cooking - boiling, frying, simmering is a result of the dynamics of the cooking components. By automatically detecting the state of cooking liquids by their sounds, such changes as occurring can be used to diagnose the condition of the cooking item before a possible onset of fire. This work made use of water, a common cooking liquid, for an empirical study. Python programming with google colab was the software tool used to display and analyze key parameters obtained from sound signals of boiling water; and sound signals of water that have boiled but still heated until the water dried out completely (heated water-dried-out). The analysis made in the time-domain view showed a marked difference in sound signal between boiling water and a heated water-dried-out. Relatively, the signal levels (amplitude) of boiling water are higher than that of a heated water-dried-out. Hence, we conclude that sounds made from cooking, if collected by embedded systems and analysed in real-time, is one safety measure to averting the incidences of home fire outbreak.