TY - GEN
T1 - ANFIS-2D wavelet transform approach to structural damage identification
AU - Escamilla-Ambrosio, P. J.
AU - Liu, X.
AU - Lieven, N. A.J.
AU - Ramirez-Cortes, J. M.
PY - 2011
Y1 - 2011
N2 - In this paper, a structural damage identification approach is proposed combining adaptive network-based fuzzy inference system (ANFIS) and 2D wavelet transform (2D WT) technologies. The approach is referred to as ANFIS-2D-WT. First, measured structure vibration response signals from multiple sensors are arranged as a 2D image signal. Then, 2D WT is applied with a twofold objective, perform sensor data fusion and work as a feature extractor. After 2D WT is applied, the energy distribution in different frequency bands of the resultant sub-2D signals is calculated. Based on its energy percentage contribution, selected elements of the obtained feature vector are taken as inputs for the ANFIS. The output of the ANFIS is a condition index, which can be a Boolean value (0 or 1) for level 1 damage assessment use (damage detection), or a number of values for level 2 damage assessment use (damage localisation). Provided an ANFIS model is well trained by the available data, it can be used for health monitoring and damage localisation. The proposed approach was applied to the data obtained from an experiment involving a cantilever beam for damage detection and localisation. The testing results show that the method is successful in detecting and classifying structural damage even in the presence of noise.
AB - In this paper, a structural damage identification approach is proposed combining adaptive network-based fuzzy inference system (ANFIS) and 2D wavelet transform (2D WT) technologies. The approach is referred to as ANFIS-2D-WT. First, measured structure vibration response signals from multiple sensors are arranged as a 2D image signal. Then, 2D WT is applied with a twofold objective, perform sensor data fusion and work as a feature extractor. After 2D WT is applied, the energy distribution in different frequency bands of the resultant sub-2D signals is calculated. Based on its energy percentage contribution, selected elements of the obtained feature vector are taken as inputs for the ANFIS. The output of the ANFIS is a condition index, which can be a Boolean value (0 or 1) for level 1 damage assessment use (damage detection), or a number of values for level 2 damage assessment use (damage localisation). Provided an ANFIS model is well trained by the available data, it can be used for health monitoring and damage localisation. The proposed approach was applied to the data obtained from an experiment involving a cantilever beam for damage detection and localisation. The testing results show that the method is successful in detecting and classifying structural damage even in the presence of noise.
KW - 2D Wavelet transform
KW - ANFIS
KW - Structural damage detection
KW - Structural damage identification
UR - http://www.scopus.com/inward/record.url?scp=79955907070&partnerID=8YFLogxK
U2 - 10.1109/NAFIPS.2011.5751912
DO - 10.1109/NAFIPS.2011.5751912
M3 - Contribución a la conferencia
AN - SCOPUS:79955907070
SN - 9781612849676
T3 - Annual Conference of the North American Fuzzy Information Processing Society - NAFIPS
BT - 2011 Annual Meeting of the North American Fuzzy Information Processing Society, NAFIPS'2011
T2 - 2011 Annual Meeting of the North American Fuzzy Information Processing Society, NAFIPS'2011
Y2 - 18 March 2011 through 20 March 2011
ER -