GCRIS

Now showing 1 - 2 of 2

Aerodynamic Optimization of a Transonic Aero-Engine Fan Module
(Izmir Institute of Technology, 2016) Kor, Orçun; Özkol, Ünver
Aerodynamic design of an aero-engine fan blade is a multi-step process with multi-variables. The general purpose in aerodynamic design is to obtain proper blade angles and flowpath geometry providing the necessary pressure ratio with maximum efficiency, while respecting the structural and aerodynamic constraints. The throughflow design in aerodynamic design procedure is a key step where one can obtain a basic aero-design which generally fixes 80% to 90% of the final fan geometry, by adjusting parameters like blade exit angle distribution, solidity, hub and shroud contour, meridional chord length, etc. Throughout this procedure, the aim of the designer is to obtain an optimum (i.e. light, reliable and robust) system with highest efficiency. Among optimization methods, zero order methods are reported to fit best for turbomachinery problems, due to their good performance in discrete and non-differentiable problems and their ability to find the global optimum. Genetic algorithm is the most widely used optimization method in turbomachinery optimization. Methods inspired by swarm intelligence are reported as promising global optimizers, whereas, to the author’s knowledge, there are no reported studies that employs such algorithms in turbomachinery throughflow optimization. These methods can find the neighborhood that provides the globally optimum design, rather than exactly finding the global design. This drawback is overcome by hybridizing genetic/swarm inspired algorithms by first order (gradient based) methods. Within this aspect, the present study focuses on developing genetic and swarm inspired algorithms hybridized with gradient based algorithms to find the optimum throughflow design of a transonic aero-engine fan module.
Development of a Streamline Curvature Throughflow Design Method for Fan Module of Turbofan Engines
(Izmir Institute of Technology, 2015) Acarer, Sercan; Özkol, Ünver
Through-flow modeling of turbomachinery flows is the principle tool for inverse design, off-design analysis and post-processing of test data, due to its capability to simulate the principal aspects of turbomachinery flows, swirling flow with rotors and stators, in the axisymmetric meridian plane with minimum two orders of magnitude smaller computational time compared to three-dimensional analysis methods. Turbomachine energy equation and empirical models for incidence, deviation, pressure loss and blockage are used to define source terms for an axisymmetric compressible flow solution. Even though the subject has been studied in numerous aspects for compressors and turbines, open literature on fully coupled fan and splitter design of turbofan engines is still limited. The present study addresses this void by developing a new split-flow method for inverse streamline curvature flow solution methodology in the course of this thesis. Hybridized empirical models that are compiled from the literature are implemented as a baseline to be calibrated. The method is validated both experimentally and numerically on a total of six different test cases within a three-step validation strategy. Firstly, split-flow solutions of the developed method for three representative duct geometries, but without a turbomachinery, are validated. Secondly, two different single-stream transonic fans, NASA 2-stage fan and a custom-designed fan stage are used to experimentally and numerically validate the empirical models, respectively. Thirdly, experimental data of GE-NASA by-pass fan is used to validate the complete models. It is shown that the accuracy of solutions in the tested cases are within less than 1.6% in pressure ratio, 2.3% in efficiency, 8% in velocity and 1.8 degree in flow angle. With this accuracy level, the proposed method is shown to be valid and can be implemented into existing compressor streamline curvature methodologies with minimal numerical effort.

Phd Degree / Doktora

Browse

Filters

Settings

Sort By

Results per page

Search Results