Modal Analysis for Heat Exchangers in Electrified Aviation

Kurzfassung

This thesis provides an overview of vibration analyses to be used for heat exchangers in electrified aviation. Information is given on the relevance of heat exchangers in electrified aviation as well as an explanation on the need for electrified aviation. Electrified aviation comes in many different configurations, many of which use hydrogen fuel cells as energy source. These fuel cells release heat which actively needs to be removed from the system, requiring a new generation heat exchanger for application in aviation. Hydrogen fuel cells convert chemical energy from hydrogen into electrical energy, producing water and heat as by-products, making them a common battery alternative. All fuel cells operate on the same basic principle: two electrodes (a cathode and an anode) are separated by an electrolyte. The anode gets fuel, and the cathode gets air. In a hydrogen fuel cell, a catalyst at the anode splits hydrogen into protons and electrons. Protons move through the electrolyte while electrons create an electric current. At the cathode, protons and electrons combine with oxygen, forming water and heat. Polymer electrolyte membrane and solid oxide fuel cells are the most promising types for use in electrified aviation. Heat exchangers are used to transfer heat between two different bodies. In electrified aviation, lightweight heat exchangers are required to remove the excess heat released by the fuel cell. A promising candidate for a heat exchanger structure is a triply periodic minimal surface (TPMS) structure. TPMS is a nature inspired curve function, which allows for maximal surface area at minimal mass. Thus, TPMS can be used in a heat exchanger with application in aviation with strict mass restrictions. Vibrations impose a great danger on the structural integrity of any object. When a vibration exerts a frequency on a heat exchanger which equals one of the resonance frequencies of the heat exchanger, it can be heavily damaged, causing leakage or other catastrophic failure. The methodology used in this thesis to obtain the eigenfrequencies using three different methods is laid out. Additionally, information on the classification of this work, including prior works, contents of this work, and following works is provided. Eigenfrequencies of multiple test bodies are determined though analytical, simulative, and experimental modal analyses. The analytical eigenfrequency solution has been found using test subjects in the shape of beams, fulfilling the Euler-Bernoulli criterion for slender beams. This allows for calculation of eigenfrequencies of beams with free-free boundary condition and homogeneous properties using the normal mode method. MSC Patran is used to simulate the eigenfrequencies and eigenmodes of the beams using FEM. The setup of the models is laid out and results are provided. An experimental modal analysis has been performed using the impact hammer excitation method. In this method, an impact hammer is used to excite a structure, whilst a laser vibrometer is used to measure the frequency response of the structure. To model the free-free boundary condition, the bodies have been placed on soft foam padding. The eigenfrequencies of four plastic, 3D printed test strips have been determined analytically and simulative, yielding an error of only 0.02% on the first eigenfrequency. This error increases with increasing eigenmode. The eigenfrequencies of a 1.5 mm thick aluminium plate have been solved for as well, showing small errors of less than 1% on the first four modes. Lastly, the eigenfrequencies of a 3 mm thick beam have been compared analytically, simulative, and experimental, yielding a rather consistent error of around 7% for all but the first two modes, due to modal interference of the first two modes. The consistent error can be explained by a frequency shift due to damping. For the continuation of this project, it is recommended to make some adjustments to the procedure: starting with a simple structure to eliminate as many errors from the start, using a more advanced FEM software, creating more suitable test objects, as well as implementing a feedback loop between experimental results and FEM.