Modelagem fuzzy para avaliação do desenvolvimento da cultura do rabanete irrigado com água tratada magneticamente sem estresse hídrico
Carregando...
Data
2018-02-26
Autores
Choueri, Matheus
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Normalmente a horticultura utiliza de irrigação artificial para suprir as demandas hídricas das plantas, mas os produtores, em geral, não são criteriosos na utilização das lâminas de reposição, o que acarreta desperdício de água e energia elétrica aumentando os custos de produção. Outro fato é a crescente quantidade de pesquisas sobre a utilização de água tratada magneticamente (ATM) empregada nas atividades agrícolas. Estas pesquisas têm mostrado aprimoramentos produtivos para diversas culturas. Este trabalho apresenta um Sistema Baseado em Regras Fuzzy (SBRF) para avaliação da cultura do rabanete para a lâmina de irrigação a 100% da evapotranspiração. Foram utilizados para o SBRF dados experimentais obtidos por experimento que avaliou a influência da água tratada magneticamente no desenvolvimento de rabanetes. O experimento foi realizado durante os meses de setembro a novembro de 2013, em uma casa de vegetação nas dependências do Departamento de Engenharia Rural da UNESP, Faculdade Ciências Agronômicas, Fazenda Experimental Lageado, localizada no município de Botucatu, São Paulo. Para a elaboração do SBRF, definiu-se um processador de entrada (fuzzificador), um conjunto de regras linguísticas, um método de inferência fuzzy e um processador de saída (defuzzificador), gerando um número real como saída. Este SBRF representa uma função F:X_1×X_2⊆R⟶R^10. O contradomínio R^10 representa as dez variáveis de saída avaliadas: Número de Folhas (NF), Comprimento da Raiz (CR), Diâmetro do Bulbo (DB), Comprimento do Bulbo (CB), Peso Verde da Raiz (PVR), Peso Verde da Folha (PVF), Peso Verde do Bulbo (PVB), Peso Seco da Raiz (PSR), Peso Seco da Folha (PSF) e Peso Seco do Bulbo (PSB). O método de inferência utilizado foi o de Mandani. Para a verificação da adequação e validação do modelo fuzzy frente ao modelo de regressão de grau 2, foram aplicados os seguintes testes e cálculos de índices em ambos modelos: Teste qui-quadrado (x2); Coeficiente de determinação (R2); Teste da falta de ajuste (lack of fit); Desvio do quadrado médio da raiz (RMSE) ou erro quadrado médio da raiz (MAE). Como ferramentas de auxílio para análise foram utilizados os seguintes procedimentos: Análise de variância; Análise de coeficientes de variação e regressão polinomial e; Análise de Conglomerados. O capítulo I descreve a metodologia para obtenção do SBRF de avaliação da cultura do rabanete irrigados com água convencional e lâmina de 100% da ETc. O capítulo II descreve a metodologia para obtenção do SBRF de avaliação da cultura do rabanete irrigado com ATM e lâmina de 100% da ETc.
Normally the horticulture uses artificial irrigation to meet the water demands of the plants, but the producers, in General, are not careful in the use of replacement blades, which carries waste of water and electricity by increasing the production costs. Another fact is the growing amount of research on the use of magnetically treated water (MTW) employed in agricultural activities. These researches have shown production enhancements to diverse cultures. This paper presents a Fuzzy rule-based System (FRBS) for evaluation of the culture of radish in two production cycles for the blade of the 100% irrigation of evapotranspiration. Were used for the FRBS the experimental data obtained by experiment that evaluated the influence of magnetically treated water in the development of radishes. The experiment was conducted during the months of September to November 2013, in a greenhouse on the premises of the Department of Rural Engineering of UNESP, Agronomic Sciences College, the Experimental Farm Lageado, located in the city of Botucatu , São Paulo. For FRBS, defined an input processor (fuzzificator), a set of linguistic rules, a fuzzy inference method and an output processor (defuzzificator), generating a real number as output. This FRBS represents a function F: X × X ⊆ ℝ⟶ℝ . The codomain ℝ , represents the ten evaluated output variables: number of leaves (NF), length of root (CR), diameter of the bulb (DB), length of the bulb (CB), Green root Weight (PVR), Green Leaf Weight (PVF), Green bulb Weight (PVB), Dry root Weight (PSR), dry weight of foliage (PSF) and dry weight of the bulb (PSB). The inference method used was that of Mandani.For the verification of the appropriateness and fuzzy model validation outside the grade 2 regression model were applied the following tests and calculations of indices in both models: Chi-square Test (x2); Coefficient of determination (R2); Lack of fit test; Deviation of the mean square of the root (MAE) or the root mean square error (RMSE). As an aid to analysis tools were used the following: analysis of variance; Analysis of variance and regression polynomial coefficients and; Analysis of conglomerates. Chapter I describes the methodology for obtaining the SBRF radish culture assessment irrigated with conventional water and 100% blade ETc. Chapter II describes the methodology for obtaining the SBRF radish culture assessment irrigated with ATM and blade of 100% of ETc.
Normally the horticulture uses artificial irrigation to meet the water demands of the plants, but the producers, in General, are not careful in the use of replacement blades, which carries waste of water and electricity by increasing the production costs. Another fact is the growing amount of research on the use of magnetically treated water (MTW) employed in agricultural activities. These researches have shown production enhancements to diverse cultures. This paper presents a Fuzzy rule-based System (FRBS) for evaluation of the culture of radish in two production cycles for the blade of the 100% irrigation of evapotranspiration. Were used for the FRBS the experimental data obtained by experiment that evaluated the influence of magnetically treated water in the development of radishes. The experiment was conducted during the months of September to November 2013, in a greenhouse on the premises of the Department of Rural Engineering of UNESP, Agronomic Sciences College, the Experimental Farm Lageado, located in the city of Botucatu , São Paulo. For FRBS, defined an input processor (fuzzificator), a set of linguistic rules, a fuzzy inference method and an output processor (defuzzificator), generating a real number as output. This FRBS represents a function F: X × X ⊆ ℝ⟶ℝ . The codomain ℝ , represents the ten evaluated output variables: number of leaves (NF), length of root (CR), diameter of the bulb (DB), length of the bulb (CB), Green root Weight (PVR), Green Leaf Weight (PVF), Green bulb Weight (PVB), Dry root Weight (PSR), dry weight of foliage (PSF) and dry weight of the bulb (PSB). The inference method used was that of Mandani.For the verification of the appropriateness and fuzzy model validation outside the grade 2 regression model were applied the following tests and calculations of indices in both models: Chi-square Test (x2); Coefficient of determination (R2); Lack of fit test; Deviation of the mean square of the root (MAE) or the root mean square error (RMSE). As an aid to analysis tools were used the following: analysis of variance; Analysis of variance and regression polynomial coefficients and; Analysis of conglomerates. Chapter I describes the methodology for obtaining the SBRF radish culture assessment irrigated with conventional water and 100% blade ETc. Chapter II describes the methodology for obtaining the SBRF radish culture assessment irrigated with ATM and blade of 100% of ETc.
Descrição
Palavras-chave
Lógica difusa, Rabanete - irrigação, Modelos matemáticos, Análise de conglomerados, Radish