Aplicação de redes neurais em análise de viabilidade econômica de co-geração de energia elétrica
Carregando...
Data
2005-06-11
Autores
Nagaoka, Marilda da Penha Teixeira [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
A co-geração de energia elétrica excedente por meio do aproveitamento do bagaço de cana-de-açúcar tem sido considerada uma alternativa importante na diversificação de fontes de geração de energia elétrica no Brasil, considerando-se as vantagens em relação à grande produção de matéria prima, menores custos de geração de energia e a possibilidade de reduzir o ônus dos investimentos em geração de energia do setor público. Apesar do grande potencial apresentado por esta fonte alternativa de energia, o mercado para a energia elétrica co-gerada está ainda hoje, sujeito a um ambiente de grande risco e incerteza, seja decorrente de condições do mercado de energia ou da produção. Este trabalho teve por objetivos analisar a viabilidade econômica de um projeto de investimento em co-geração de energia elétrica em uma usina sucroalcooleira na região Oeste do estado de São Paulo,com vistas à comercialização de excedentes, sob condições de risco, utilizando o algoritmo de Redes Neurais Artificiais. Procurou-se também testar a convergência dos resultados obtidos por este método com outro mais tradicionalmente utilizado em análise de risco para a determinação dos indicadores de viabilidade econômica do investimento. Os indicadores utilizados foram Valor Atual Líquido (VAL); Taxa Interna de Retorno (TIR); Relação Benefício - Custo (RBC); Payback Simples (PBS) e Payback Econômico (PBE). A análise foi realizada considerando seis cenários, considerando a possibilidade ou não de obtenção de financiamento e diferentes níveis de eficiência de queima do bagaço. No método de Redes Neurais Artificiais, as redes foram alimentadas com as seguintes variáveis de entrada: valor do investimento; despesas com juros e amortização; despesa com aquisição e transporte do bagaço e receita operacional, tendo como variável de saída o fluxo líquido de caixa.
The co-generation of surplus electrical energy by means of the use of sugar-cane bagasse has been considered as an important alternative in the diversification of sources of electrical energy in Brazil. Its advantages in relation to the production of raw material are: smaller costs of generation of energy and the possibility to reduce the costs of the investments in the generation of energy in the public sector. In spite of the great potential presented by this alternative source of energy, the market for the co-generation of electrical is still today subject to an atmosphere of great risk and uncertainty, be it due to conditions of the energy or of the production market. The objective of this research study was to analyze the economic viability of an investment project of cogeneration of electrical energy in an alcohol and sugar mill based on the Western area of the state of São Paulo having in view the commercialization of surpluses, under risk conditions, using the algorithm of Artificial Neural Networks. It was also tried to test the convergence of the results obtained by this method with a more traditionally method used in analysis of risk for the determination of the indicators of economic viability of the investment. The indicators used were Liquid Current Value (LCV); Internal Rate of Return (IRR); Benefit - Cost Relationship (BCR); Simple Payback (SPB) and Economic Payback (EPB). The analysis was performed into six different scenarios, having into consideration the possibility or not availabity of financing, and the different levels of efficiency in the burning of bagasse. In the method of Artificial Neural Networks the nets were supplied with entrance variables, such as, the value of the investment; expenses with interests and amortization; expense with acquisition and transport of the bagasse; operational revenue, and the exit variable included the liquid cash flow.
The co-generation of surplus electrical energy by means of the use of sugar-cane bagasse has been considered as an important alternative in the diversification of sources of electrical energy in Brazil. Its advantages in relation to the production of raw material are: smaller costs of generation of energy and the possibility to reduce the costs of the investments in the generation of energy in the public sector. In spite of the great potential presented by this alternative source of energy, the market for the co-generation of electrical is still today subject to an atmosphere of great risk and uncertainty, be it due to conditions of the energy or of the production market. The objective of this research study was to analyze the economic viability of an investment project of cogeneration of electrical energy in an alcohol and sugar mill based on the Western area of the state of São Paulo having in view the commercialization of surpluses, under risk conditions, using the algorithm of Artificial Neural Networks. It was also tried to test the convergence of the results obtained by this method with a more traditionally method used in analysis of risk for the determination of the indicators of economic viability of the investment. The indicators used were Liquid Current Value (LCV); Internal Rate of Return (IRR); Benefit - Cost Relationship (BCR); Simple Payback (SPB) and Economic Payback (EPB). The analysis was performed into six different scenarios, having into consideration the possibility or not availabity of financing, and the different levels of efficiency in the burning of bagasse. In the method of Artificial Neural Networks the nets were supplied with entrance variables, such as, the value of the investment; expenses with interests and amortization; expense with acquisition and transport of the bagasse; operational revenue, and the exit variable included the liquid cash flow.
Descrição
Palavras-chave
Energia eletrica e calor - Cogeração, Energia elétrica, Redes neurais (Computação), Risk, Investment, Artificial neural networks, Monte-Carlo method, Cogeneration
Como citar
NAGAOKA, Marilda da Penha Teixeira. Aplicação de redes neurais em análise de viabilidade econômica de co-geração de energia elétrica. 2005. xii, 124 f. Tese (doutorado) - Universidade Estadual Paulista, Faculdade de Ciências Agronômicas, 2005.