Ferramentas para análise de autogestão e sustentabilidade do uso da água em perímetros irrigados

• Ano de defesa: 2015
• Autor(a) principal: Gonçalves, Fabricio Mota
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Engenharia agrícola; Análise multivariada; Inteligência artificial; Programação linear; Multivariate analysis
• Link de acesso: http://www.repositorio.ufc.br/handle/riufc/16880

This work aims to characterize the current stage of Irrigated Perimeters Federal government with a view to self-management process and present alternative of allocating water distribution in secondary irrigation canals. The research was divided into two themes. The first addressed the development of a methodology for evaluating the performance of Irrigated Perimeters from the creation of a statistical model Multivariate discriminant and an Artificial Neural Network using the performance indicators of irrigated public areas of the National Department of Works Against Drought (Dnocs) and Development Company of the São Francisco and Parnaíba (Codevasf) as a way to evaluate the prospect of self-management of the same. The second dealt with the optimization of water use, a case study at the Experimental Farm Curu Valley, belonging to the Federal University of Ceará, in the area adjacent to the irrigated Curu Pentecost were accomplished. Based on information provided by the National Department of Works Against Drought (Dnocs) and Development Company of the São Francisco and Parnaíba (Codevasf), the key performance indicators relating to Self-Management of Irrigated Perimeters were evaluated. The Multivariate and discriminant analysis (AMD) technique Artificial Neural Networks (ANN) were used to separate the standards relating to the performance of Irrigated Perimeters linear character or not. RNA yielded the automatic identification of the pattern that belongs to each perimeter over time. Based on the results obtained in the multivariate discriminant analysis, we observed the Generation Revenue per Hectare (HRM) as the most important indicator in discriminatory process between Irrigated Perimeters regarding self-management. The perimeters with the best performance in relation to self-management were: Nilo Coelho, Curaçá I Pirapora and Maniçoba. Regarding the operationalization of water use, we used a mathematical model of linear programming to determine the most rational way to release water for irrigated areas. The allocation defined by mathematical modeling proved adequate for the needs of established cultures, showing the most rational use of water.