Unit:
Department of PhysicsLibrary of the School of Science
Supervisors info:
ΜΑΡΓΑΡΙΤΑ ΑΣΗΜΑΚΟΠΟΥΛΟΥ, ΕΠΙΚΟΥΡΗ ΚΑΘΗΓΗΤΡΙΑ ΕΘΝΙΚΟΥ ΚΑΠΟΔΙΣΤΡΙΑΚΟΥ ΠΑΝΕΠΙΣΤΗΜΙΟΥ ΑΘΗΝΩΝ, ΤΜΗΜΑ ΦΥΣΙΚΗΣ
Original Title:
ΕΦΑΡΜΟΓΗ ΠΟΛΥΕΠΙΠΕΔΟΥ ΤΕΧΝΗΤΟΥ ΝΕΥΡΩΝΙΚΟΥ ΔΙΚΤΥΟΥ ΓΙΑ ΤΗΝ ΑΝΑΛΥΣΗ ΤΩΝ ΕΠΙΔΟΣΕΩΝ ΑΝΕΜΟΓΕΝΝΗΤΡΙΩΝ ΚΟΙΝΗΣ ΩΦΕΛΕΙΑΣ
Translated title:
IMPLEMENTATION OF A MULTILEVEL ARTIFICIAL NEURAL NETWORK FOR ANALYSING THE PERFORMANCE OF GENERAL INTEREST WIND TURBINES
Summary:
The purpose of this paper is to compare and evaluate the power output of a horizontal axis wind turbine after installing Vortex generators-Vgs. The Vgs are placed on the wings of a particular wind turbine model which will be analyzed, control flow separation, thereby influencing its performance. In order to compare it with its previous situation, under exactly the same meteorological and mechanical conditions for safe conclusions, a model was created through multilevel artificial neural networks that impersonates the intact wind turbine. The modeling was static, followed by a process in which the neural network was trained with a back propagation algorithm using supervised data. This process is used to model time-independent functions. This type of function is the power curve of the wind turbine that shows its power output at different wind speeds. Thus, wind turbine data series were collected and used before the Vgs were installed to create the model, to simulate and generate power curves that would contrast with the new curves of the modified wind turbine that would be created exactly under the same conditions. In this way, the two annual energy outputs will be compared and will therefore be noticed if the modification of the turbine generator with the vortex generators improves its performance. After the process followed, there is a reduction in flow separation that extends the operating duration of the wind turbine and reduces the noise caused by the turbulent flow, resulting in the stable operation for a much longer duration. There will be a change in wind distribution and wind turbine operation basically at average speeds (approximately 75% of the operation of the amplified wind turbine at these speeds), resulting in improved performance after the vortex generators are installed. Improvement will occur even at higher wind speeds.
Main subject category:
Science
Other subject categories:
Energy
Environment
Keywords:
Renewable energy, wind energy, wind turbine, vortex generators, neural networks
