Damage Detection of the Jabiru's Aircraft Wing under Operational Fuel Loading Conditions using Neural Network

Damage detection and structural health monitoring (SHM) of an aircraft wing exposed of changing fuel load can lead to a false alarm if the loading effects are not intelligently discriminated. This is because the loading effects can alter the vibration response and be misinterpreted as damage effects. This study proposed the Principal Component Analysis (PCA)-Artificial Neural Network (ANN) for detecting damage of on aircraft wing under the effects of varying fuel tank loading conditions. A vibration test is performed on Jabiru wing which the measured signal is applied with Principal Component Analysis (PCA) to reduce the high dimensionalities and extract the features. ANN is then utilized to map the principal component indices into various damage severities and loading classes using multi -layer perceptron ANN. The results from the study show promising results when incorporating PCA with the ANN to predict various damage severities of the aircraft wing under changing fuel load conditions.