In the past few years, intelligent structural damage identification algorithms based on machine learning techniques have been developed and obtained considerable attentions worldwide, due to the ad...

https://journals.sagepub.com/doi/pdf/10.1177/1475921718804132

A novel deep learning-based method for damage identification of smart building structures:

Structural health monitoring using modal strain energy damage indicator coupled with teaching-learning-based optimization algorithm and isogoemetric analysis

Size-dependent isogeometric analysis of functionally graded carbon nanotube-reinforced composite nanoplates

Fast simulations for solving fracture mechanics inverse problems using POD-RBF XIGA and Jaya algorithm

Modal identification and damage detection in beam-like structures using the power spectrum and time-frequency analysis

Damage assessment based on the vibration response of structures has been, in the last decades, an important research area in various engineering branches. It integrates knowledge from a variety of disciplines, such as mechanical and structural engineering, material science, computer science and signal processing, and data management. This special issue is dedicated to fundamental understanding of the vibrational behavior of damaged structures and to the development of methods and techniques that are able to tackle these specific problems. A number of 23 papers have been submitted, from which a total of 7 papers have been finally selected to integrate the present issue. Two among them are focused on failures in machineries: W. Wen et al. in the paper “Planetary Gearbox Fault Diagnosis Using Envelope Manifold Demodulation” extract fault characteristics from the vibration signals, since in “Undecimated Lifting Wavelet Packet Transform with Boundary Treatment for Machinery Incipient Fault Diagnosis” by L. Duan et al. a Volterra series assisted undecimated lifting wavelet packet transform is investigated for machinery incipient fault diagnosis. The presented approaches dramatically enhance the weak defect feature extraction. Damages in structural elements are the target of research presented in “Development of a Vehicle-Bridge-Soil Dynamic Interaction Model for Scour Damage Modelling” by L. Prendergast et al., “Feature Recognition forMine HoistingWire Rope Based on Magnetic Flux Leakage Signals” by Z. Zhao et al., and “Experimental Study on Progressive Collapse Performance of Frame with Specially Shaped Columns Subjected to Middle Column Removal” by T. Wang et al. A model updating strategy for bolt-ball joint stiffness characterization for health monitoring is presented in “Finite Element Model Fractional Steps Updating Strategy for Spatial Lattice Structures Based on Generalized Regression Neural Network” by C. Liu et al. An interesting method to perform efficient fatigue tests is proposed by Y. Jiang et al. in the paper “Experimental Design and Validation of an Accelerated Random Vibration Fatigue Testing Methodology.” They succeed to significantly reduce the test time and the sample size. The papers reflect, without doubt, significant researches and developments of this subject. We hope that readers will find all articles of the special issue useful and exciting and that the articles will stimulate further research activities in the area of damage assessment.

/pdf/damage-models-and-assessment-methods-2a6v6ghkwm.pdf

Damage Models and Assessment Methods

Non-destructive testing of structures involving vibration-based damage detection methods implies knowing the beam's boundary conditions. For perfect boundary conditions, numerous damage detection methods are developed and ensure more or less accurate estimation of the crack type, position, and severity. On the contrary, for imperfect boundary conditions, which are the real ones, there are only a few dedicated works. This paper presents a methodology that allows the identification of the weak clamping and the crack if both exist. The weak clamping is modelled as a defect that produces an identical relative frequency shift (RFS) for all vibration modes. Therefore, to find the two defects namely the real crack and defect simulating the weakly clamped end, we apply the principle of superposition. The method is implemented as an application written in the Python programming language. Tests show that defects are successfully identified even if there are uncertainties about the fixing of the beam.

A method to detect cracks in the beams with imperfect boundary conditions

PyFEST - a Python code for accurate frequency estimation

The current paper presents an algorithm for locating transverse cracks present in Euler–Bernoulli steel beams, by considering the relative frequency shifts (RFSs) measured for the weak-axis vibration modes and the mode shape curvatures. To apply the algorithm, we first determine, by means of the finite element method (FEM), the natural frequencies for the first four bending vibration modes for the healthy beam fixed-free and fixed-fixed end conditions. Afterward, we introduced transverse cracks with known depth and location and determined again the natural frequencies. As a next step, we use an original algorithm to plot the potential severity curves (PSCs) for the four modes and find the intersection, which indicates the crack severity and location. It was found that the method is reliable for determining the location of transverse cracks just by using the measured natural frequency and the squared normalized modal curvatures (SNMCs) for both imposed boundary conditions.

Crack Assessment Based on the Use of Severity-Adjusted Modal Curvatures of the Healthy Beam

We present in this paper a simple approach to estimate the frequency components of a signal acquired from a freely vibrating structure. A non-contact excitation technique with a controlled bandwidth is involved, which is obtained by employing a loudspeaker producing for a short time a sine signal. This excitation ensures a strong harmonic component in the frequency range of interest and low components outside this range. In this way, the component of interest is not polluted by noise. The relevant portion of the response signal is extracted for spectral analysis. First, an interpolation by applying zero-padding in order to double the number of the spectral bins is performed. In this way, the main lobe in the spectrum contains four points. Based on three of these points a polynomial interpolation is applied and the function crossing the points is found. The maximum is presumed to be the magnitude, while its ordinate is considered the natural frequency. Simulations involving a generated pure sine confirm the precision achieved by this method.

Gilbert-Rainer Gillich

Papers

Damage Models and Assessment Methods

A method to detect cracks in the beams with imperfect boundary conditions

PyFEST - a Python code for accurate frequency estimation

Crack Assessment Based on the Use of Severity-Adjusted Modal Curvatures of the Healthy Beam

Two-Stage Interpolation Algorithm for Precise Frequency Estimation of Vibration Signals