Banca de DEFESA: ANA CAROLINA DE SÁ SILVA LINS

Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.
DISCENTE: ANA CAROLINA DE SÁ SILVA LINS
DATA: 03/02/2026
HORA: 14:00
LOCAL: On-line
TÍTULO:

INNOVATIVE SMART TECHNOLOGIES FOR ASSESSING THERMAL COMFORT IN LAYING HENS: HYBRID MODELS AND AI-BASED DIAGNOSTIC TOOLS

PALAVRAS-CHAVES:

poultry farming; laying hen performance; artificial intelligence; fuzzy inference system; smart production; prediction of zootechnical responses; raspberry pi.

PÁGINAS: 118
GRANDE ÁREA: Ciências Agrárias
ÁREA: Engenharia Agrícola
SUBÁREA: Construções Rurais e Ambiência
ESPECIALIDADE: Engenharia de Construções Rurais
RESUMO:

This thesis proposes solutions to mitigate the impacts of heat stress in laying hens—one of the most critical challenges in poultry farming, particularly in tropical regions. The introductory chapter contextualizes the issues faced by Brazilian poultry production in view of climate variability and increasing demands for animal welfare, emphasizing the relevance of computational modeling and intelligent sensing technologies. By addressing the physiological effects of temperature, humidity, and ventilation on birds, the work establishes the scientific and technological foundation for the proposed innovations. Bridging concepts from engineering, artificial intelligence, and animal environment, the research developed and validated two key technological tools: a hybrid intelligent model for predicting physiological responses and the global heat transfer coefficient, and a portable heat stress detector based on fuzzy logic. The first article presents a hybrid predictive model that combines heat transfer equations, fuzzy inference systems (FISs), and regression techniques. A total of 84 FISs were evaluated to predict physiological variables—cloacal temperature, surface temperature, and respiratory rate—as well as the global heat transfer coefficient (U). Multiple combinations of input variables were tested using both internally developed FISs and data from the literature to identify the most accurate models. The resulting models outperformed traditional empirical methods, eliminating the need for invasive measurements and enabling simulations that support thermal management strategies and the design of poultry facilities. The second article describes the development of a portable heat stress detector built on a Raspberry Pi 5 microcomputer, incorporating calibrated environmental sensors (DHT22 and DS18B20) and a fuzzy logic algorithm to estimate, in real time, the respiratory rate of laying hens and classify their thermal comfort status. The system achieved high predictive accuracy and demonstrated robustness under conditions of uncertainty. Designed for integration into precision poultry farming operations, the solution is affordable, replicable, and suitable for use in environmental monitoring and decision-support systems. The thesis contributes to the advancement of artificial intelligence applications in animal environments, promoting more sustainable poultry production, improved bird welfare, and enhanced operational efficiency in modern poultry farming.

MEMBROS DA BANCA:
Interno - ALESSANDRO TORRES CAMPOS (Membro)
Externo à Instituição - DIAN LOURENÇONI - UNIVASF (Suplente)
Externo ao Programa - EDISON JOSE FASSANI - DZO/FZMV (Membro)
Externo ao Programa - GISELLE BORGES DE MOURA - DEA/EENG (Suplente)
Externo à Instituição - GREGORY MURAD REIS - UF (Membro)
Presidente - TADAYUKI YANAGI JUNIOR (Membro)
Externo à Instituição - YAMID FABIÁN HERNÁNDEZ JULIO - UNISINÚ (Membro)