Banca de DEFESA: MURILO SANTOS FREIRE

Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.
DISCENTE: MURILO SANTOS FREIRE
DATA: 19/07/2023
HORA: 14:00
LOCAL: Webconferência
TÍTULO:

A NOVEL SMART THERMAL ENVIROMENT CONTROLLER FOR LAYING HEN HOUSES

PALAVRAS-CHAVES:

Precision livestock farming, instrumentation and control, poultry

PÁGINAS: 84
GRANDE ÁREA: Ciências Agrárias
ÁREA: Engenharia Agrícola
SUBÁREA: Construções Rurais e Ambiência
RESUMO:

In poultry farming, the advances obtained in relation to genetics, nutrition, sanity, and management have still been limited by the production environment, especially the thermal environment. Therefore, it is necessary to monitor and to control in real-time the thermal environment of commercial aviaries. Despite the available technologies, there is still a lack of products and services accessible to producers and researchers. In this context, the objective of this research was to develop a controller based on artificial intelligence for the control of climatization systems in laying hens aviaries. In addition, the controller developed has the capacity for remote monitoring, from a wireless sensor network, in which data collected and processed supports the control of actuators of environmental systems in aviaries. A database containing thermal variables and physiological variables of laying hens was used for the development and simulation of a climate control system. An Algorithm based on artificial intelligence (artificial neural networks - ANNs) was embedded in the control system, using thermal variables as input data in the model and using physiological variables as output data. The performance of the ANNs was quantified by mean absolute error (MAE), mean absolute percentage error (MAPE), mean squared error (MSE), root mean squared error (RMSE), coefficient of determination (R²) and the efficiency index of Nash-Sutcliffe (NSE). The ANN18 model has the best performance for predicting respiration rate (RR), which uses the ANN-BR method. The response surface demonstrates that increasing black-globe temperature (t_bg) directly affects the RR response, mostly between 32 ºC and 36 ºC. The functional relationship between values observed and the values simulated using the chosen ANN model had an R² of 0.85. The functional relationship between values predicted and the values simulated using the chosen ANN model with literature chronic thermal data was R² of 0.87. These results showed that the ANN model developed was able to predict efficiently RR using t_bg and relative humidity (RH) as inputs. The embedded ANN model was able to predict RR and to control the actuation module, according to control states settings. The user interface was able to manage the smart control system remotely. The controller performed adequately when subjected to operating simulation.

MEMBROS DA BANCA:
Externo à Instituição - ALLAN ALVES FERNANDES - Unipampa (Suplente)
Externo à Instituição - DANIEL DOS SANTOS COSTA - UNIVASF (Membro)
Externo à Instituição - DIAN LOURENÇONI - UNIVASF (Membro)
Externo ao Programa - WILIAN SOARES LACERDA - DAT/EENG (Membro)
Externo ao Programa - LEONARDO SCHIASSI - DEA/EENG (Suplente)
Interno - ALESSANDRO TORRES CAMPOS (Membro)
Presidente - TADAYUKI YANAGI JUNIOR (Membro)