A Novel Approach for Fire Safety

Abstract

Liquids and solvents in industrial products produce flammable vapor which, when mixed with air, can ignite or explode. The ease by which those liquids produce flammable vapors depends on their flashpoints which allow them to be categorized according to the fire hazard they exhibit in their normal use. In this chapter, a novel approach for the classification of liquids is proposed. The proposed approach relies on the use of a vector network analyzer, a patch antenna, and a group of classifiers. In this study, random forest and REPTree algorithms are preferred as classifers. As proven in the study, random forest algorithm can provide higher accuracy than REPTree algorithm in the classification of hazardous liquids. A prototype system is currently under development in order to integrate the components of the proposed approach into a single unit. It is expected that the prototype system will quickly and reliable make a non-contact classification of liquids in different kinds of bottles.