Session: 36-01 Preliminary Design and Structural Optimisation

Paper Number: 152092

Multimodal Rotor-Dynamic Optimization Based on Transfer Matrix Method for Different Rotor Structures 

Considering the thrust/weight ratio in the aviation industry, the most efficient engines are turbomachines. In the design of high-speed turbomachines, there are design requirements such as not having a natural frequency in the operating speed range, bearing loads not exceeding a certain value, clearances between the rotor and stator being within a certain value, and ensuring a lightweight rotor design, and taking all of these requirements into consideration is beyond human intuition requires optimization. More than one optimization algorithm that can be used for such a comprehensive design is available in the literature. However, the performance of optimization algorithms varies greatly depending on the problem type, as required by the No Free Lunch Theorem. For this reason, optimizing a mini turbojet rotor and optimizing a conventional turboshaft or turbojet rotor in preliminary design phase are quite different optimization problems in terms of design space and behavior of objective functions. In this study, metaheuristic optimization methods Genetic Algorithm (GA), Differential Evolution (DE), Simulated Anneal (SA), Gravitational Searh Algorithm (GSA), Black Hole (BH), Particle Swarm Optimization (PSO), Harris Hawk Optimization (HHO) were used. , Aritifcal Bee Colony (ABC) and non-metaheuristic Pattern Search (PS) methods, Multimodal (MM) and Unimodal (UM) fitness functions, Multi-objective functions with various design spaces including Low-Dimensional (LD) and High-Dimensional (HD) rotordynamics was applied to optimization problems and their performances were compared. In addition, unlike the Finite Element Method (FEM) used in the literature, our own developed JoeRot rotordynamic solver, based on the Complex Transfer Matrix Method (CTMM), was used to obtain the objective functions in optimization problems. CTMM, which provides much faster results than FEM, has proven its applicability for complex rotordynamics design problems. When the results obtained were examined, it was observed that PSO, one of the metaheuristic algorithms, gave better results for the UM-LD and MM-LD type optimization problem, while ABC, one of the metaheuristic algorithms, gave better results for the MM-HD type optimization problem.

Presenting Author: Hüseyin Tarık Niş TUSAS Engine Industries Inc.

Presenting Author Biography: I'm Hüseyin Tarık Niş, working as a Whole Engine Structural Engineer at TEI - TUSAŞ Engine Industries Inc. My expertise is focused on Whole Engine Structural Analysis. I've been in the aviation industry for over three years, concurrently pursuing my doctoral studies.

During my tenure as a PhD Research Engineer, I've been engaged in studies within the field of Mechanical Engineering, specifically concentrating on Rotor Dynamics. I hold bachelor's degrees in Electrical Electronics Engineering and Mechanical Engineering from Bursa Technical University. My dual major in Mechanical and Electrical Electronics Engineering provided me with a versatile perspective and functional skills.

In my current role, I aim to utilize and further develop my professional experience at TEI to generate solutions within the aviation sector.

Authors:

Hüseyin Tarık Niş TUSAS Engine Industries Inc.
Ahmet Furkan Özden TUSAS Engine Industries
Ömer Faruk Kale TUSAS Engine Industries
Ahmet Yıldız Bursa Uludag University