A comparative machine learning approach for entropy-based damage detection using output-only correlation signal

Hamidian, Pouria; Soofi, Yasaman J.; Bitaraf, Maryam

doi:10.1007/s13349-022-00587-z

Cited by 9 publications

(3 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Subsequently, sample entropy (SampEn) was developed as an improvement over AE, offering independence from embedding dimension and similarity coefficient in measuring time series irregularity [24]. SE has been widely embraced as a health indicator in both mechanical fault diagnosis [25] and SHM [26]. Permutation entropy (PE) was subsequently introduced to assess physical systems, considering the nonlinear states within the sequence of a time series [27].…”

Section: Introductionmentioning

confidence: 99%

Multi-weighted symbolic sequence entropy: a novel approach to fault diagnosis and degradation monitoring of rotary machinery

Wu,

Yuan,

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

Structural health monitoring relies heavily on measurements. Entropy theory is emerging as a critical quantitative analysis technique for interpreting measured data for both health monitoring and fault identification of structural systems. This paper introduces the algorithms of weighted symbolic sequence entropy (WSSE) and its multi-weighted derivation, termed multi-weighted symbolic sequence entropy (MWSSE). WSSE optimizes the existing algorithms, including symbolic sequence entropy (SSE) and improved symbolic sequence entropy (ISSE), by introducing a weighting factor in the computation process of the algorithm. WSSE significantly contributes to enlarging the difference between normal and fault sequences of signals in rotary machinery. The multi-weighted derivation of WSSE, namely MWSSE, more effectively depicts the dynamical characteristics of rotary machinery by utilizing entropy values of multiple weighting factors as health indicators. The MWSSE-based degradation monitoring and fault identification approaches developed by the authors are validated through application research and comparative analysis. These novel entropy algorithms offer innovative quantitative analysis techniques for fault diagnosis in rotary machinery and structural health monitoring and hold potential for application in broader research fields.

show abstract

Section: Introductionmentioning

confidence: 99%

Multi-weighted symbolic sequence entropy: a novel approach to fault diagnosis and degradation monitoring of rotary machinery

Wu,

Yuan,

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…Using the time-series response, the auto/cross correlation function becomes a viable tool in vibration-based damage detection. [17][18][19][20][21][22][23][24][25][26][27][28][29] Gradzki et.al 20 proposed a rotor fault detection approach by using auto-correlation and power spectral density. Both a numerical and experimental test of the rotor verified high sensitivity and reliability of the method.…”

Section: Introductionmentioning

confidence: 99%

“…Wang and Chen 22 showed that the data fusion of the correlation function among different types of vibration measurements could significantly improve the accuracy of the identification. Hamidian and Soofi 23 proposed an output-only damage classification method using correlation function and information entropy, which has 93.98% accuracy for the test data. Taking the idea of Cross Correlation Function Amplitude Vector (CorV) 24 and Inner Product Vector (IPV), 25 based on the theory of natural excitation technique (NExT), 26 Zhang and Schmidt [27][28][29] proposed the Auto correlation function at Maximum point value Vector (AMV) using the auto correlation function.…”

Section: Introductionmentioning

confidence: 99%

Damage Assessment Using the Auto-Correlation-Function-Based Damage Index

Zhang,

Li,

Peng

et al. 2023

The International Journal of Acoustics and Vibration

View full text Add to dashboard Cite

A damage index describing damage severity based on the Auto correlation function at Maximum point value Vector (AMV) is proposed for damage assessment as well as locating the damage is described in this study. The detectability of the AMV-based damage detection method is compared with Mean Strain Energy (MSE) and the Generalized Flexibility Matrix (GFM) method, to show its efficiency in localizing the damage and its simplicity to conduct the analysis. Moreover, a procedure to estimate the damage severity using the sensitivity analysis is introduced. The stiffness reduction detection of a twelve-story shear frame structure is provided to show that the estimated damage severity by the AMV-based method is in close agreement with the simulated damage for small damage even when the measurement noise exists or is just part of the available modal parameters.

show abstract

Prediction of vertical displacement for a buried pipeline subjected to normal fault using a hybrid FEM-ANN approach

Jalali,

Yeganeh Khaksar,

Mohammadzadeh S.

et al. 2024

Front. Struct. Civ. Eng.

View full text Add to dashboard Cite

A comparative machine learning approach for entropy-based damage detection using output-only correlation signal

Cited by 9 publications

References 48 publications

Multi-weighted symbolic sequence entropy: a novel approach to fault diagnosis and degradation monitoring of rotary machinery

Multi-weighted symbolic sequence entropy: a novel approach to fault diagnosis and degradation monitoring of rotary machinery

Damage Assessment Using the Auto-Correlation-Function-Based Damage Index

Prediction of vertical displacement for a buried pipeline subjected to normal fault using a hybrid FEM-ANN approach

Contact Info

Product

Resources

About