Vibration-based Methods Research Articles

A Stacked Neural Network Model for Damage Localization.

Traditional vibration-based damage detection methods often involve human intervention in decision-making, therefore being time-consuming and error-prone. In this study, we propose using Artificial Neural Networks (ANNs) to detect patterns in the structural response and create accurate predictions. The features extracted from the response signal are the Relative Frequency Shifts (RFSs) of the first eight weak-axis bending vibration modes, and the predictions refer to the damage location. To increase the accuracy of the predictions, we propose a novel stacked neural network approach, capable of detecting damage locations with high accuracy. The dataset used for training involves, as input data, the RFSs calculated with an original method for numerous damage locations and severities. The following models were used as building blocks for our stacked approach: Multilayer Perceptron, Recurrent Neural Network, Long Short-term Memory, and Gated Recurrent Units. The entire beam was thus split into segments and each network was trained in this stacked model on one beam segment. All results obtained with the models are also compared to a standard neural network trained on the entire beam. The results obtained show that the model that performs the best contains 14 stacked two-layer feedforward networks.

Wireless vibration testing and bridge deck damage identification using underneath maintenance walkway

Bridges will inevitably experience structural damage during long-term operation, which will decrease their structural load bearing capacity and durability. The vibration characteristics of bridge structures can be used as evaluation indicators for accurately diagnosing their deterioration. This study aimed at developing a vibration testing method for identifying bridge deck damage that does not affect traffic and is easy to implement. The vibration tests on the in-service bridge revealed it was mainly subjected to vertical vibrations, with the accelerations in the vertical direction about twice those in the horizontal direction. Based on fast Fourier transform analysis, the dominant vertical vibration frequencies were 2.34 Hz, 4.10 Hz, 10.30 Hz, 14.45 Hz and 18.75 Hz. A series of 0.1–0.3 Hz bandpass filters were used with those frequencies at the centers for modal analysis. The experimental results were then compared to the finite element analysis results. This study proposed a vibration-based damage identification method based on the differences between the bridge deck and both main steel girders. The study confirmed the convenience and effectiveness of using underneath maintenance walkways for wireless vibration testing.

A spectrum correlation matrix-based rapid damage identification method for joints in hinged slab bridges by sparse measurement

Hinged slab bridges are widely used in the medium and small-span bridges on highways and the joints are highly susceptible to damage under the action of the increasing traffic. However, current methods for damage identification of hinged joints, whether rely on static or dynamic testing, typically require substantial manpower, resources and time, which presents challenges for the rapid evaluation of numerous bridges. In view of this, a novel damage index named spectrum correlation matrix (SCM) is proposed and its feature extraction technology is established as well as the potential for sparse measurement is theoretically explained. Furthermore, a rapid, robust, and convenient damage identification methodology using SCM and Levenberg-Marquardt algorithm is proposed. An optimal strategy for the layout of sparse measuring points has been explored, the sensitivity and the effectiveness of the proposed method is validated through both numerical simulations and the in-situ test of a bridge in service. The results indicate that the SCM contains abundant vibration information and shows a high sensitivity to damage. The strategy for the layout of measuring points is highly efficient and the proposed methodology greatly reduces the number of measuring points compared to previous vibration-based methods, which eases rapid identification of damage for joints in hinged slab bridges.

Contactless femoral implant stability monitoring in cementless total hip arthroplasty, A step towards clinical implementation

The clinical implementation of currently used devices for intraoperative fixation monitoring of femoral implants via vibration-based methods in cementless total hip arthroplasty is challenging, due to practical and regulatory issues. Motivated by the effectiveness of electromagnetic excitation in similar dental applications, this study investigates the use of electromagnetic excitation for femoral implant stability monitoring during cementless total hip arthroplasty. The results obtained from electromagnetic excitation were largely consistent with reference results obtained through impact excitation, with a Pearson Correlation Coefficient of 0.79 in the 0.1–8 kHz frequency band. Moreover, the peak frequencies obtained via the two methods yielded a relative difference of 0.20 ± 0.22 %. Next, the excitation device was successfully utilized in conjunction with a laser vibrometer to monitor the stability of the femoral implant during an in vitro insertion, proving the feasibility of contactless implant stability monitoring. These results indicate the promising potential of this contactless method for clinical implementation.

Vibration Analysis of a Centrifugal Pump with Healthy and Defective Impellers and Fault Detection Using Multi-Layer Perceptron

Centrifugal pumps (CPs) are widely utilized in many different industries, and their operations are maintained by their reliable performance. CPs’ most common faults can be categorized as mechanical or flow-related faults: the first ones are often associated with damage at the impeller, while the second ones are associated with cavitation. It is possible to use computational algorithms to monitor both failures in CPs. In this study, two different problems in pumps, the defective impeller and cavitation, have been considered. When a CP is working in a faulty condition, it generates vibrations that can be measured using piezoelectric sensors. Collected data can be analyzed to extract time- and frequency-domain data. Interpreting the time-domain data showed that distinguishing the type of defect is not possible. However, indicators like kurtosis, skewness, mean, and variance can be used as input for the multi-layer perceptron (MLP) algorithm to classify pump faults. This study presents a detailed discussion of the vibration-based method outcomes, emphasizing the benefits and drawbacks of the multi-layer perceptron method. The results show that the suggested algorithm can identify the occurrence of different faults and quantify their severity during pump operation in real time.

A vibration-based method to estimate the low-wavenumber wall pressure field in a turbulent boundary layer

When an elastic structure is excited by a low-speed turbulent flow, it generates flow-induced vibration which is primarily due to the low-wavenumber components of the wall pressure field (WPF) beneath the turbulent boundary layer (TBL). Therefore, to accurately predict the vibration response of the structure, an accurate estimation of the low-wavenumber WPF is needed. While existing TBL models for WPF well predict the convective region, they exhibit significant discrepancies in predicting the low-wavenumber levels. This numerical study aims to explore the feasibility of estimation of the low-wavenumber WPF by analysing vibration data obtained from a structure excited by a TBL. An inverse method is proposed based on the relationship between the TBL forcing function and structural vibrations in the wavenumber domain. The random TBL force is simulated with deterministic loading using the uncorrelated wall plane wave technique. A model of a simply supported plate under a TBL excitation is developed to demonstrate the proposed method. The plate's acceleration data is then used to estimate the WPF in the low-wavenumber range. It is shown how using multiple discrete frequencies in the analysis can reduce the required snapshots for accurate estimation of the WPF.

Correction: A Comparative Study of Wear Debris and Vibration-Based Gear Damage Detection Methods Applied to Mild Wear in a Spur Gear System

Correction: A Comparative Study of Wear Debris and Vibration-Based Gear Damage Detection Methods Applied to Mild Wear in a Spur Gear System

Development of a Vibration-Based Pre-Alarming Method for Bolt Looseness of Seismic Sway Braces

Development of a Vibration-Based Pre-Alarming Method for Bolt Looseness of Seismic Sway Braces

A Comparative Study of Wear Debris and Vibration-Based Gear Damage Detection Methods Applied to Mild Wear in a Spur Gear System

A Comparative Study of Wear Debris and Vibration-Based Gear Damage Detection Methods Applied to Mild Wear in a Spur Gear System

Fault Detection of Rotating Machines Using poly-Coherent Composite Spectrum of Measured Vibration Responses with Machine Learning

This study presents an efficient vibration-based fault detection method for rotating machines utilising the poly-coherent composite spectrum (pCCS) and machine learning techniques. pCCS combines vibration measurements from multiple bearing locations into a single spectrum, retaining amplitude and phase information while reducing background noise. The use of pCCS significantly reduces the number of extracted parameters in the frequency domain compared to using individual spectra at each measurement location. This reduction in parameters is crucial, especially for large industrial rotating machines, as processing and analysing extensive datasets demand significant computational resources, increasing the time and cost of fault detection. An artificial neural network (ANN)-based machine learning model is then employed for fault detection using these reduced extracted parameters. The methodology is developed and validated on an experimental rotating machine at three different speeds: below the first critical speed, between the first and second critical speeds, and above the second critical speed. This range of speeds represents the diverse dynamic conditions commonly encountered in industrial settings. This study examines both healthy machine conditions and various simulated fault conditions, including misalignment, rotor-to-stator rub, shaft cracks, and bearing faults. By combining the pCCS technique with machine learning, this study enhances the reliability, efficiency, and practical applicability of fault detection in rotating machines under varying dynamic conditions and different machine conditions.

Vibration as a New Survey Method for Spiders.

Spiders have important ecological roles as generalist predators, are a significant source of food for many other species, and are bioindicators of environmental health. However, spiders are poorly studied. Given their importance, a comparison of spider survey methods used to determine differences in spider diversity and abundance is required to understand their limitations and biases. A new survey method to attract spiders, based on vibration from an idling diesel tractor, was tested and compared to the traditional methods of pitfall trapping and hand collection of spiders at night. Across the three survey methods, there were, in total, 2294 spiders in 34 families, 138 genera, and 226 species identified. Spider species diversity and richness were significantly greater for spiders collected at night than from the other two methods (spiders collected in pitfall traps and attracted to vibration). The collection of spiders using the night collection and vibration-based methods were very similar in terms of labor required and material costs. Of all spider species identified, 80% were captured during hand collection, 30% through pitfall trapping, and 30% from vibration-based collection. Most species of spiders caught in pitfall traps were species known to be primarily ground-dwelling, whereas both arboreal and ground-dwelling spiders were collected at night and as a result of being attracted and collected using the vibration-based method.

Diagnosis for railway point machines using novel derivative multi-scale permutation entropy and decision fusion based on vibration signals

Abstract Railway point machines (RPMs) are safety-critical pieces of equipment closely related to train operation safety. Due to their high failure rate, it is urgent to develop an effective diagnosis method for RPMs. Considering the easy-to-collect and anti-interference characteristics of vibration signals, this paper develops a vibration-based diagnosis method. First, to address the difficulty of multi-scale permutation entropy in characterizing the fault information contained in the derivatives of the raw signal, a novel feature named derivative multi-scale permutation entropy is designed, which can further complete the fault information of RPMs. Second, to further improve the diagnosis accuracy of support vector machines, a decision fusion strategy based on three feature sets is developed, which can further improve the diagnosis accuracy, especially in the normal-reverse direction. Finally, the effect and superiority of the proposed method are verified based on the collected vibration signals from Xi’an Railway Signal Co.,Ltd by experiment comparisons. The diagnosis accuracies of reverse-normal and normal-reverse directions reach 99.43% and 100% respectively, indicating its superiority.

An effective vibration-based feature extraction method for single and multiple damage localization in thin-walled plates using one-dimensional wavelet transform: A numerical and experimental study

By early detection of single and multiple damaged locations and repairing or replacing defective elements, it is possible to prevent damage to other areas and structural elements. The use of thin-walled steel plates is on the rise in the building industry. Damage to these elements can progressively develop and spread to other elements, resulting in overall structural damage. This article focused on detecting single and multiple damaged locations in thin-walled steel plates. An effective feature extraction method based on wavelet analysis with a one-dimensional theoretical background to analyze two-dimensional modal signals was introduced. This method calculates the peak of disturbances and irregularities in various damaged locations and their surrounding areas. The suggested method was used to identify the center of single and multiple damaged areas with an error of less than 5 %. The numerical and experimental results confirmed the effectiveness and efficiency of the proposed method in detecting various damaged locations.

Enhanced domain transfer deep fuzzy echo state network for rotating machinery fault diagnosis based on current signal

Although vibration-based fault diagnosis methods have achieved remarkable results in rotating machinery, they are still limited by various factors, such as environment noise and additional sensor installation. Meanwhile, due to inevitable distribution differences in practical engineering, deep transfer learning models are commonly used in fault diagnostic fields. Although deep architectures can extract representative domain-invariant features, they typically result in substantial computational burdens and training time. Therefore, to address the above problems, an enhanced domain transfer deep fuzzy echo state network (EDFESN) is proposed for rotating machinery fault diagnosis based on current signals. Firstly, current signals derived from the driving motor are directly collected without additional sensors. Secondly, an enhanced domain mapping method is raised to solve the problem of data distribution discrepancies between cross domain. Furthermore, a deep fuzzy echo state network (DFESN) integrated with fuzzy clustering is constructed, which utilizes layer-wise fuzzy-tuning learning paradigm instead of backpropagation step to reduce training time. Two datasets from gearbox and bearing both verify the effectiveness of the proposed method. Compared with other popular transfer methods, EDFESN not only achieves highest accuracy, but also has fastest training speed.

Vibration-based methods for local damage identification of breathing cracks in truss-like structures

A vibration-based local damage identification method is proposed for truss-like structures with breathing cracks. Firstly, a bilinear model is established to represent the characteristics of breathing cracks, and the features of such cracks are detected through reciprocity and homogeneity tests on structural input-output signals. Since the crack is influenced by tensional and compressive forces on truss members, only one damage factor is required to quantify the stiffness reduction of each member. Subsequently, a dynamic equilibrium equation concerning the damage factor is constructed to identify the new load term. The integral equation method, which relies on the virtual work principle, is employed to obtain the displacement responses and loads under all tensile conditions. Finally, the mean value of cumulative damage coefficients for each truss member can be derived as their corresponding damage factor. Numerical examples involving a plane truss structure and a truss offshore platform are utilised to illustrate the effectiveness of this damage identification method. The results demonstrate exceptional precision and efficiency in accurately identifying various damages in truss structures under different excitations and signal-to-noise ratios. A actual space truss structure confirm that this damage identification method exhibits high-level precision in identifying single-member damages as well as multi-member damages with breathing cracks locally.

Application of Modal Strain Energy Analysis to Localize an Added Mass on an Aluminium Plate.

The modal strain energy method is a vibration-based damage detection method that allows the use of passive measurements. The method is based on the measurement and analysis of the modal response of the structure under investigation. The aim of this work is to investigate the application of this method to detect and localise a local change in a structure using acoustic measurements in a fluid around the structure. Firstly, the modal strain energy is presented. The technique is based on the analysis of the modal shapes of undamaged and damaged samples obtained by operational modal analysis. The application of modal strain energy to off-contact acoustic measurements is then discussed. The method is used to detect and locate a local change (an additional mass) in an aluminium plate subjected to a mechanical force. The study begins with numerical simulation using finite element models, followed by experiments on the aluminium plate.

Vibration-based damage detection method with tunable resolution for composite laminates

Composite laminates are increasingly applied in advanced engineering structures, but they are prone to be damaged in service. However, multiple defects with different sizes challenge the current damage detection methods to identify all of them without a baseline. In this paper, a baseline-free damage identification method based on vibration is developed, and a novel damage index with tunable resolution of detection is proposed based on two-dimensional continuous wavelet transform by manipulating wavelet parameters. The relationship between the smallest detectable damage size and wavelet parameters is explored quantitatively, and the effectivity and reliability of proposed method are numerically and experimentally verified to directly detect damage in composite laminates without baselines. By a comparison with other vibration-based damage detection methods, results show that the proposed method can precisely locate and visualize multi-damage with determinate resolution of detection, and can realize multi-resolution of damage detection for identifying various sizes of defects.

A combined implicit-explicit vibration-based SHM method for damage detection of wind turbine blades

A combined implicit-explicit vibration-based SHM method for damage detection of wind turbine blades

A New Order Tracking Method for Fault Diagnosis of Gearbox under Non-Stationary Working Conditions Based on In Situ Gravity Acceleration Decomposition

Rotational speed measuring is important in order tracking under non-stational working conditions. However, sometimes, encoders or coded discs are not easy to mount due to the limited measurement environment. In this paper, a new in situ gravity acceleration decomposition method (GAD) is proposed for rotational speed estimation, and it is applied in the order tracking scene for fault diagnosis of a gearbox under non-stationary working conditions. In the proposed method, a MEMS accelerometer is locally embedded on the rotating shaft or disc in the tangential direction. The time-varying gravity acceleration component is sensed by the in situ accelerometer during the rotation of the shaft or disc. The GAD method is established to exploit the gravity acceleration component based on the linear-phase finite impulse response (FIR) filter and complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) methods. Then, the phase signal of time-varying gravity acceleration is derived for rotational speed estimations. A motor–shaft–disc experimental setup is established to verify the correctness and effectiveness of the proposed method in comparison to a mounted encoder. The results show that both the estimated average and instantaneous rotational speed agree well with the mounted encoder. Furthermore, both the proposed GAD method and the traditional vibration-based tacholess speed estimation methods are applied in the context of order tracking for fault diagnosis of a gearbox. The results demonstrate the superiority of the proposed method in the detection of tooth spalling faults under non-stationary working conditions.

Application of variational mode decomposition in vibration-based bridge damage detection

Bridges are critical infrastructure that require continuous structural health monitoring to ensure safety. Vibration-based methods use changes in dynamic response for damage detection. In this study, variational mode decomposition (VMD), a signal analysis technique, was applied to detect damage in a three-span beam bridge model subjected to moving loads. A finite-element model was developed to represent the bridge–vehicle interaction and damage was simulated by reducing the flexural stiffness of the bridge. Sensors placed on each span recorded the vibrations during vehicle passage. VMD was used to decompose the signals for analysis of damage indications. Different damage levels (30–50% stiffness reduction) and locations (each span) were evaluated. The vehicle speed was also varied to assess its effect. Damage was identified as peaks in the instantaneous energy diagrams of decomposed signals, correctly indicating the axle locations. Detection was improved with sensors nearer the damage and lower vehicle speed. The results demonstrate the ability of VMD to detect and locate damage without a baseline, indicating its potential as a non-baseline vibration-based damage detection method for bridges.