Date of Award
Open Access Thesis
As in service aircraft begin to age and fatigue, a method for evaluating the operational life they are currently operating under and have remaining comes into question. Structural health monitoring is (SHM) is a popular method of structural analysis with growing interest in the aerospace industry. SHM is capable of damage assessment and structural life estimations.
The ultimate goal of the research presented in this thesis is to develop a methodology of classifying the length of a fatigue crack though the use of machine learning. The thesis has three major chapters as described below.
The first chapter deals with the understanding of acoustic emissions as elastic waves, as well as how their qualities are related to the characteristics of their source mechanism. It is proved that the waveforms of acoustic emissions (AE) signals change as the length of the fatigue crack they originate from changes. It is shown that the AE signals also carry distinct patterns when they originate from a source of crack growth or crack rubbing.
The second chapter of the thesis presents a method of fatigue testing with load parameters determined by the stress intensity factor (SIF) approach. This chapter also explains the setup and results of in-situ fatigue experiments to collect AE signals.
The third chapter of this thesis shows the approach to the implementation of the collected AE data into an AI model. The approach was done with two separate model types, the first being GoogLeNet, a popular convolutional neural network (CNN) model, and the second being a custom long-short-term memory (LSTM) model. The work goes over the processes taken to process raw data into usable Choi Williams transform (CWT) spectrograms. Model performance enhancement techniques are also discussed in the forms of synthetic data generation and class balancing.
The AI models developed in this paper will have the potential future use of being applied to in service aircraft for the detection of fatigue cracking in order to avoid dangerous failure situations. These AI models can be used with passive PWAS sensors and be used for structural life analysis with minimal monitoring and upkeep.
Ennis, S. T.(2023). An Artificial Intelligence Approach to Fatigue Crack Length Estimation From Acoustic Emission Signals. (Master's thesis). Retrieved from https://scholarcommons.sc.edu/etd/7280