Please use this identifier to cite or link to this item:
http://hdl.handle.net/1843/BUOS-B8UG5Z
Type: | Tese de Doutorado |
Title: | Prediction of solar/wind energy in a hybrid renewable energy system using artificial intelligence |
Authors: | Ali Khosravi |
First Advisor: | Luiz Machado |
First Referee: | Raphael Nunes de Oliveira |
Second Referee: | Tatiana Paula Alves |
Third Referee: | Ralney Nogueira de Faria |
metadata.dc.contributor.referee4: | Rudolf Huebner |
metadata.dc.contributor.referee5: | Antonio Augusto Torres Maia |
Abstract: | . |
Abstract: | The case study is Unit 132 of the second refinery of the South Pars, Bushehr in the south of Iran. This unit has a refrigeration system that is responsible to deliver 39 kg/s cold water for cooling the equipment of the refinery. In the first part, two refrigeration systems are designed for this target, which are heat exchanger refrigeration system (HXRS) and ejector expansion refrigeration system (EERS). R134a, R407C and R410A refrigerants are evaluated to identify the most suitable one for the proposed systems. Energy, exergy, economic and environmental analyses are investigated for each system. Sizing of the evaporator, condenser and ejector device are determined. In the second part, artificial intelligence methods are developed to predict solar and wind energy. These predicted data are employed to calculate the produced power by wind turbines and photovoltaic panels. In the third part, a hybrid renewable energy with hydrogen energy storage system is designed to provide the electrical energy for this refrigeration unit. The target of this part is to define and assess an off-grid hybrid renewable energy with hydrogen storage system. The system combines solar energy, wind energy, hydrogen production unit and fuel cell. Energy, exergy, and economic analyses are carried out for the proposed system. The results demonstrated that from energy, exergy, environmental and economic point of views R134a EERS is more efficient than HXRS with different working fluids. For prediction of solar radiation data, multilayer feed-forward neural network (MLFFNN) and support vector regression (SVR) performed better than the other developed models. For wind speed prediction, SVR outperformed the other developed models for all time intervals. Also, for hybrid renewable energy system, the amount of energy and exergy efficiencies for photovoltaic system (in the case study region) were obtained as 12% and 16%, respectively. In addition, for wind turbine system, the values of energy and exergy efficiencies were achieved 32% and 26%, respectively. The payback period of the proposed renewable energy system was obtained around 11 years |
Subject: | Energia eólica Geração de energia fotovoltaica Inteligência artificial Engenharia mecânica |
language: | Inglês |
Publisher: | Universidade Federal de Minas Gerais |
Publisher Initials: | UFMG |
Rights: | Acesso Aberto |
URI: | http://hdl.handle.net/1843/BUOS-B8UG5Z |
Issue Date: | 21-Jun-2018 |
Appears in Collections: | Teses de Doutorado |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
thesis_final.pdf | 6.07 MB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.