Strict-Feedback Backstepping Digital Twin and Machine Learning Solution in AE Signals for Bearing Crack Identification

Bearings are nonlinear systems that can be used in several industrial applications. In this study, the combination of a strict-feedback backstepping digital twin and machine learning algorithm was developed for bearing crack type/size diagnosis. Acoustic emission sensors were used to collect normal and abnormal data for various crack sizes and motor speeds. The proposed method has three main steps. In the first step, the strict-feedback backstepping digital twin is designed for acoustic emission signal modeling and estimation. After that, the acoustic emission residual signal is generated. Finally, a support vector machine is recommended for crack type/size classification. The proposed digital twin is presented in two steps, (a) AE signal modeling and (b) AE signal estimation. The AE signal in normal conditions is modeled using an autoregressive technique, the Laguerre algorithm, a support vector regression technique and a Gaussian process regression procedure. To design the proposed digital twin, a strict-feedback backstepping observer, an integral term, a support vector regression and a fuzzy logic algorithm are suggested for AE signal estimation. The Ulsan Industrial Artificial Intelligence (UIAI) Lab’s bearing dataset was used to test the efficiency of the combined strict-feedback backstepping digital twin and machine learning technique for bearing crack type/size diagnosis. The average accuracies of the crack type diagnosis and crack size diagnosis of acoustic emission signals for the bearings used in the proposed algorithm were 97.13% and 96.9%, respectively.

Experimental Detection and Measurement of Crack-Type Damage Features in Composite Thin-Wall Beams Using Modal Analysis

An experimental proof-of-concept for damage detection in composite beams using modal analysis has been conducted. The purpose was to demonstrate that damage features can be detected, located, and measured on the surface of a relatively complex thin-wall beam made from composite material. (1) Background: previous work has been limited to the study of simple geometries and materials. (2) Methods: damage detection in the work is based on the accurate measurement of mode shapes and an appropriate design of the detection mesh. Both a method requiring information about the healthy structure and a baseline-free method have been implemented. (3) Results: short crack-type damage features, both longitudinal and transverse, were detected reliably, and the true length of the crack can be estimated from the damage signal. Simultaneous detection of two cracks on the same sample is also possible. (4) This work demonstrates the feasibility of automated damage detection in composite beams using sensor arrays.

Assessment of integrity and remaining working life of welded steel structures

Analyzing the period of exploitation of welded steel structures it can be concluded that they are predominantly exposed to the action of variable load. The welded joint as the largest stress concentrator due to the heterogeneity of structural, mechanical and operational properties is a key problem that is further complicated by the possible and realistically probable presence of crack-type faults. The assessment of integrity largely depends on a comprehensive analysis of the welded joint as the most critical place of welded steel structures. Integrity assessment is a necessary obligation for extending the working life, as well as revitalization, as a way to keep the structures in operation, despite the long period of exploitation. This paper presented an analysis of the process of fatigue crack initiation and growth, i.e. an assessment of the of the welded steel structures’ integrity and remaining service life under the influence of variable load.

Development of Low-Frequency Phased Array for Imaging Defects in Concrete Structures

The nondestructive inspection of concrete structures is indispensable for ensuring the safety and reliability of aging infrastructures. Ultrasonic waves having a frequency of tens of kHz are frequently used to reduce the scattering attenuation due to coarse aggregates. Such low frequencies enable the measurement of the thickness of concrete structures and detection of layer-type defects, such as delamination, whereas it causes a lack of sensitivity to crack-type defects. In this paper, to realize the ultrasonic phased array (PA) imaging of crack-type defects, we fabricated a low-frequency (LF) array transducer with a center frequency of hundreds of kHz. To avoid the crosstalk between piezoelectric elements and dampen the vibration of each element, we adopted soft lead zirconate titanate (soft PZT) with a low mechanical quality factor. Subsequently, we optimized the geometry of each piezoelectric element using a finite element method to generate a short pulse. After validating the design in a fundamental experiment using a single-element transducer, we fabricated a 32-element array transducer with a center frequency of 350 kHz. To show the imaging capability of the LF array transducer, we applied it to a concrete specimen with a delamination. As a result, the PA with the LF array transducer clearly visualized the delamination, which could not be visualized using the PA with a 2.5 MHz array transducer. Furthermore, we applied it to a more challenging defect, a slit, which is sometimes used to simulate crack-type defects. As a result, the PA with the LF array transducer clearly visualized a slit of 1 mm width and 40 mm height in a concrete specimen. Thus, we demonstrated the usefulness of the LF array transducer for inspecting crack-type defects.

Concrete Cracks Monitoring using Deep Learning-based Multiresolution Analysis

In this paper, we propose a new methodology for crack monitoring in concrete structures. This approach is based on a n this paper, we propose a new methodology for monitoring cracks in concrete structures. This approach is based on a multi-resolution analysis of a sample or a specimen of the studied material subjected to several types of solicitation. The image obtained by ultrasonic investigation and processing by a dedicated wavelet will be analyzed according to several scales in order to detect internal cracks and crack initiation. The ultimate goal of this work is to propose an automatic crack type identification scheme based on convolutional neural networks (CNN). In this context, crack propagation can be monitored without access to the concrete surface and the goal is to detect cracks before they are visible on the concrete surface. The key idea allowing such a performance is the combination of two major data analysis tools which are wavelets and Deep Learning. This original procedure allows to reach a high accuracy close to 0.90. In this work, we have also implemented another approach for automatic detection of external cracks by deep learning from publicly available datasets.

EVALUATION OF IMPACT OF ELEVATED TEMPERATURES ON YOUNG’S MODULUS OF GLT BEAMS

The influence of elevated temperatures on mechanical behavior of glued laminated timber beams is examined on the basis of tensile tests. Dog bone samples prepared from beams exposed to fire of variable duration were categorized with respect to the type and position of the failure crack, type and number of discontinuities such as knots, and the level of browning. The acquired experimental results suggest that the wood variability and the effect of growth discontinuities are probably more significant than the effect of elevated temperatures. To support this conclusion, further study is currently under way, exploring samples from the second series of the fire tests.

Numerical Investigation on Crack Propagation for a Central Cracked Brazilian Disk Concerning Friction

This paper concerns the effect of friction on crack propagation for the centrally cracked Brazilian disk under diametric forces by using a modified finite element method. It shows that the mode II stress intensity factor decreases obviously with the increase of friction after the crack is closed, while friction has no influence on the stress intensity factor of mode I and T-stress. Meanwhile, there are some significant influences on the crack propagation due to the change of the friction after the crack is closed with the appropriate loading angle and relative length of the crack. When T-stress is positive, the effect of friction becomes obvious and the crack propagation angle increases with a lager friction coefficient. With increasing the friction, the deviation for the crack propagation trajectory increases and the curvature of path decreases, which may lead to the change of crack type. Additionally, the larger relative crack length can amplify the effect of friction, which is similar to the loading angle.

Using a Variational Autoencoder in the Task of Overlaying Multiple Images

The possibility and expediency of using variational autoencoders when expanding training datasets of neural networks for cases when the training set consists of several dozen samples is tested. Investigations are carried out on the example of images of «crack»-type defects. Brief information on the theory of variational autoencoders is given. Practical recommendations are given for constructing training sets of variational autoencoders. It is shown that deviation from the suggested recommendations will most likely not allow generating realistic images for the case of a small dataset. For the case of a dataset of eighty images, the distribution of «crack»- type defects in the hidden space after training the variational autoencoder is demonstrated. Examples of images with defects sampled from different parts of the distribution of latent factors are given. For the case of images of cracks, the continuity of the hidden space is demonstrated, when one image is sufficiently smoothly transformed into another on the way through the space of hidden features. A method for obtaining superimposed images based on the use of variational autoencoders is proposed. This method seems to be promising, since it allows automating the process of obtaining superimposed images. Examples of generated cracked superimposed images are shown.

Research on Construction Crack Diagnosis and Treatment Technology of Double-Column Large Cantilever Pre-Stressed Bent Cap

This paper took a pre-stressed bent cap of a high pier of a bridge as the research object, expounded the crack type and construction process of the bent cap, combined with the numerical calculation and analysis of the construction process to study the crack generation mechanism and its influence during the construction period of the bent cap. Besides, it gave the repair and treatment technology of jacking and unloading combined with pressure glue, so as to provide reference for the disease diagnosis and treatment of similar bridge structures.

A CRACK DETECTION SYSTEM FOR STRUCTURAL HEALTH MONITORING AIDED BY A CONVOLUTIONAL NEURAL NETWORK AND MAPREDUCE FRAMEWORK

The quickly expanded development of artificial intelligence offers alternative ways to solve numerous civil engineering problems. The work is devoted to the development of a computer-vision-based crack detection system capable to process big data related to pathology recognition. In this study, we discuss an automated crack type classification pipeline based on CNN deep learning algorithm and MapReduce framework. The results of numerical modeling illustrate the potential of the crack detection system.