Banca de DEFESA: EDISON CAMILO DE MORAES JUNIOR

Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.
DISCENTE : EDISON CAMILO DE MORAES JUNIOR
DATA : 08/09/2021
HORA: 08:30
LOCAL: videoconferência na plataforma RNP (sala PEI-UFBA)
TÍTULO:

DEVICES AND VIRTUAL ANALYZER FOR MONITORING AGRICULTURAL PRODUCTIVITY IN REAL TIME IN THE SUGAR ALCOHOLE INDUSTRY.

PALAVRAS-CHAVES:

sugarcane, dendrometer, NDT,oBrix meter, quality assessment, Internet of Things (IoT), neural networks.

PÁGINAS: 95
GRANDE ÁREA: Engenharias
ÁREA: Engenharia de Produção
RESUMO:

Brazil is the world's largest producer of sugarcane, accounting for almost 41% of production, followed by India (17%) and China (6%). Sugarcane quality is an important factor in the yield of the sugarcane industry and one of the main indicators of sugarcane quality is the sugar content (°Brix). In addition to oBrix, Total Recoverable Sugar (ATR) is a typical productivity indicator that provides the mass of sugar (kg) present in a ton of cane produced. Currently, the methods most used by the sugar and alcohol industries to measure °Brix are the °Brix refractometer, the MIR spectroscopy (middle infrared spectroscopy) and the NIR (near infrared spectroscopy). In turn, the ATR analysis is performed in the sucrose laboratory through physical processes such as clarification and filtration, among others. The technologies for measuring °Brix and ATR are based on sampling and laboratory analysis, which makes it impossible to monitor in real time the quality and productivity of sugarcane and the consequent decision-making in production management. This work presents two innovative proposals for real-time monitoring of sugarcane oBrix and ATR. In the first case (oBrix), two electronic devices (Dendometer and UltraBrix) were developed that provide a non-destructive alternative for measuring the diameter of the stem and sugar content of the cane directly in the field. In the second case (ATR), a pseudo-dynamic model based on Artificial Neural Networks (ANN) was identified to estimate the ATR throughout production, based on meteorological measurements. Stem diameter measurements obtained by the device had a mean error of ±3 mm (approximately 10% at an average diameter of 30 mm) using a standard caliper measurement as a reference. The sucrose content meter (UltraBrix) employs the continuous wave technique and the estimates provided a coefficient of determination (R2) of 0.83, taking as reference values measured by a refractometer. The neural models identified for six cultivars (C0997, RB92579, RB93509, RB845210, RB867515 and SP791011) provided an average error in estimating the final ATR (kg/ton) of the crop of 2,60 (2,01 %), 3,99 (2,86 %), 3,42 (2,49 %), 2,95 (2,09 %), 3,41 (2,53 %) and 3,37 (2,47 %), respectively. Empirical models based on ANN include input variables (accumulated solar radiation, accumulated rainfall, leaves and number of days of planting) that are easy to measure/monitor, which allows the estimation of the evolution of the sugarcane ATR throughout the harvest and the identification of the appropriate time to carry out the harvest, enabling an estimated increase of around 20% in sucrose production.

MEMBROS DA BANCA:
Externo à Instituição - DANIEL IBRAIM PIRES ATALA
Externo à Instituição - CAMILO FREDDY MENDOZA MOREJON - UNIOESTE
Presidente - 2199115 - CRISTIANO HORA DE OLIVEIRA FONTES
Externo ao Programa - 1194490 - IURI MUNIZ PEPE
Externo à Instituição - Jandecy Cabral Leite
Externo ao Programa - 2042003 - MARCUS VINICIUS AMERICANO DA COSTA FILHO