Proposta de sistema de detecção de incêndio em estrutura urbana com uso de tecnologia embarcada e recursos renováveis de água e energia

Abstract

Fire risks in cities are increasing and the demand for water and energy resources for fire safety activities grows as population increases and population density intensifies.In this context, this research seeks to develop an efficient and sustainable fire detection system (SDI), based on embedded technologies and renewable water and energy resources, so that it can detect fire early and provide renewable resources for the activity of detecting and controlling the sinister.A the first moment, it seeks to estimate the supply capacity of the technical fire reserve (RTI) provided for the PRODEMA/UFPB building through rainwater, as well as predict the energy generation of a Hybrid Power Generation System (SHGE), composed of a photovoltaic panel and a wind turbine, to supply the energy demand of the proposed SDI. For this purpose, information contained in satellite databases on rainfall, solar irradiance and wind speed in the city of João Pessoa, from 2015 and 2020, was used. With regard to SDI, a set of software and hardware was structured based on the Arduino platform, in which gas, temperature and flame sensors were connected, using the decision tree method and submitting the data obtained by the sensors. to a Matrix Laboratory (MATLAB) routine, a heuristic capable of early identification of fire patterns was determined. The results showed that it is feasible to set up a rainwater catchment system at PRODEMA/UFPB to supply the RTI.As for energy generation for the SDI, both the predictions of the mathematical model of the SHGE and the physical model proved to be able to meet the energy requirement demanded by the proposed detection system. Finally, regarding the performance of the SDI, calibrated from the analysis of response tests given by sensors to controlled burning of paper and Medium Density Fiberboard (MDF), carried out in a physical simulator with dimensions proportional to a residential building, the system was able to satisfactorily detect the fire , being able to identify it between 15 to 40 seconds after the start of the fire, presenting a response time to the fire less than that of the automatic shower, when submitted to the same test situation. As for efficiency, after comparison tests between what was seen through visual inspection and what was predicted by the implemented heuristic, the level of assertiveness was around 94%.