Early fire detection based on gas sensor arrays: Multivariate calibration and validation

Smoldering fires are characterized by the production of early gas emissions that can include high levels CO and Volatile Organic Compounds (VOCs) due to pyrolysis or thermal degradation. Nowadays, standalone sensors, smoke detectors, a combination these, standard components for fire alarm systems. While sensor arrays together with pattern recognition techniques valuable alternative detection, in practice they have certain drawbacks—they detect emissions, but show low immunity nuisances, time drift render calibration models obsolete. In this work, we explore performance array detecting smoldering plastic while ensuring rejection set nuisances. We conducted variety nuisance experiments validated room (240 m 3 ). Using PLS-DA SVM, evaluate different multivariate dataset. remain predictive after several months, perfect is not achieved. For example, 4 months calibration, model provides 100% specificity 85% sensitivity since system has difficulties fires, whose signatures close scenarios. Nevertheless, our results systems based on able provide faster response than conventional smoke-based alarms. also propose use small-scale increase number conditions at reduced cost. Our an effective way model, even when evaluated room. Finally, acquired datasets made publicly available community (doi: 10.5281/zenodo.5643074).