SSI-LSTM network for adaptive operational modal analysis of building structures

In order to identify time-varying transient modal parameters only from nonstationary vibration response measurement signals for slow linear time-varying (SLTV) structures which are weakly damped, a moving window differential evolution (DE) independent component analysis- (ICA-) based operational modal analysis (OMA) method is proposed in this paper. Firstly, in order to overcome the problems in traditional ICA-based OMA, such as easy to go into local optima and difficult-to-identify high-order modal parameters, we combine DE with ICA and propose a differential evolution independent component analysis- (DEICA-) based OMA method for linear time invariant (LTI) structures. Secondly, we combine the moving widow technique with DEICA and propose a moving window differential evolution independent component analysis- (MWDEICA-) based OMA method for SLTV structures. The MWDEICA-based OMA method has high global searching ability, robustness, and complexity of time and space. The modal identification results in a three-degree-of-freedom structure with slow time-varying mass show that this MWDEICA-based OMA method can identify transient time-varying modal parameters effectively only from nonstationary vibration response measurement signals and has better performances than moving window traditional ICA-based OMA.

Operational modal analysis (OMA) is a procedure that allows the modal parameters of a structure to be extracted from the measured response to an unknown excitation generated during operation. Nonlinearity is inevitably and frequently encountered in OMA. The problem: The traditional OMA method based on linear modal theory cannot be applied to a nonlinear oscillation system. The solution: This paper aims to propose a nonlinear OMA method for nonlinear oscillation systems. The new OMA method is based on the following: (1) a self-excitation phenomenon is caused by nonlinear components; (2) the nonlinear normal modes (NNMs) of the system appear under a single-frequency harmonic excitation; and (3) using forced response data, the symbolic regression method (SR) can be used to automatically search for the NNMs of the system, whose modal parameters are implicit in the expression structure expressing each NNM. The simulation result of a three-degree-of-freedom (3-DOF) nonlinear system verifies the correctness of the proposed OMA method. Then, a disc-rod rotor model is considered, and the proposed OMA method’s capability is further evaluated.

Conventionally, modal monitoring of Wind Turbine Rotor Blades is primarily based on the evaluation of eigenfrequencies. Beyond this, combining a sensor network with the Operational Modal Analysis (OMA) method, mode shape and parallely a local component are utilized here. In addition it is expected that the damping, which is also determined by the OMA method, will give a lead on damage development at the rotor already at an early stage. Modal monitoring by means of measurement is combined with FEM simulation and with the comparison of results obtained from measurement and simulation. Moreover, this will establish a connection between the engineer and the design data of a rotor blade, which also are based on FEM analyzes. A further significant increase regarding error resolution is possible by combining the global modal methods with locally sensitive monitoring methods, based on guided elastic waves. These assume plate-like structures through which elastic waves propagate in the low-frequency ultrasonic range (10 - 100 kHz) in certain modes. These different wave modes interact distinctively with inner structural damages such as web fractures and delaminations. It is differentiated between piezoelectrically excited waves (acousto ultrasonics), and waves produced by energy released at fractures, delamination etc. (acoustic emission). Applying a moderate number of sensors, the combination of both methods can allow an effective monitoring of the global structure.

Structural Health Monitoring (SHM) is nowadays common in many branches of engineering since it allows to have a continuous or periodic report of the structural conditions and therefore to promptly intervene if there are incipient damages. The first step to perform a SHM is the identification of the dynamic parameters, i.e. natural frequencies, damping ratios and modal shapes, and it is a crucial step since a modification of the structural parameters can be a direct consequence of structural damages. Among the structural identification methods, Operational Modal Analysis (OMA) methods have received increasing attention from the researchers since they do not require the knowledge of the structural excitation that is due to ambient vibrations and that is usually modeled as a white noise. This aspect makes this kind of methods cheaper and simpler than the classical Experimental Modal Analysis (EMA) methods. In this paper an innovative OMA method is proposed. It is a semi - automated method that allows to identify natural frequencies, damping ratios and modal shapes of a structural system and that can be used also from users that have not knowledge in stochastic dynamics and signal analysis. First of all, the modal shapes are estimated through the use of signal filtering techniques applied on the stochastic properties of the output process and then natural frequencies and damping ratios can be estimated from the mono - component analytical signals obtained by performing a decomposition of the analytical signals matrix. The proposed method has been used to perform the dynamic identification of a real historic building situated in Palermo, i.e. Chiaramonte palace, and the results obtained have been compared with those obtained by using other OMA methods.

A new operational modal analysis (OMA) method is developed for estimation of modal parameters (MPs) of a rotating structure (RS) subject to random excitation using a nonuniform rotating beam model, an image processing method, and an improved demodulation method. The solution to the governing equation of a nonuniform rotating beam is derived, which can be considered as the response of the beam measured by a continuously scanning laser Doppler vibrometer (CSLDV) system. A recently developed tracking CSLDV system can track and scan the RS. The image processing method determines the angular position of the RS so that the tracking CSLDV system can sweep its laser spot along a time-varying path on it. The improved demodulation method obtains undamped mode shapes (UMSs) of the RS by multiplying its measured response by sinusoids whose frequencies are its damped natural frequencies (DNFs) that are obtained from the fast Fourier transform of the measured response. Experimental investigation of the OMA method using the tracking CSLDV system is conducted, and MPs of a rotating fan blade (RFB), including DNFs and UMSs, with different constant speeds and its instantaneous MPs with a non-constant speed are estimated. Estimated first DNFs and UMSs of the stationary fan blade and RFB are compared with those from the lifting method that was previously developed by the authors.

Safety assessment of structures for near-field blast-induced ground excitations using operational modal analysis

In order to identify the modal parameters of time invariant three-dimensional engineering structures with damping and small nonlinearity, a novel isometric feature mapping (Isomap)-based three-dimensional operational modal analysis (OMA) method is proposed to extract nonlinear features in this paper. Using this Isomap-based OMA method, a low-dimensional embedding matrix is multiplied by a transformation matrix to obtain the original matrix. We find correspondence relationships between the low-dimensional embedding matrix and the modal coordinate response and between the transformation matrix and the modal shapes. From the low-dimensional embedding matrix, the natural frequencies can be determined using a Fourier transform and the damping ratios can be identified by the random decrement technique or natural excitation technique. The modal shapes can be estimated from the Moore–Penrose matrix inverse of the low-dimensional embedding matrix. We also discuss the effects of different parameters (i.e., number of neighbors and matrix assembly) on the results of modal parameter identification. The modal identification results from numerical simulations of the vibration response signals of a cylindrical shell under white noise excitation demonstrate that the proposed method can identify the modal shapes, natural frequencies, and ratios of three-dimensional structures in operational conditions only from the vibration response signals.

This paper investigates the use of conventional operational modal analysis (OMA) techniques for modal parameters identification of a rotating system supported by roller bearings. Although largely applied in civil engineering, in-depth studies on different types of systems are still limited in the literature. The novel of the paper is to address such issue by applying conventional OMA methods, such as enhanced frequency domain decomposition (EFDD) and stochastic subspace identification (SSI-data), to identify the modal parameters of a rotating machine, investigating the challenging particularities of these systems due to their inherent operating conditions, especially regarding the presence of harmonic forces, eventually closed-spaced modes, and non-proportional damping due to the bearings. The results presented in the experimental tests showed that, with the use of specific tools, in comparison with traditional experimental modal analysis (EMA), the used OMA methods have managed to successfully identify the modal parameters of a roller bearing supported rotor.

Abstract. In the present study we investigate development of operational modal analysis (OMA), different OMA techniques as well as relevant issues and experimental studies. Furthermore, we classify previously performed studies have been conducted so far and consider the most important one. So, first we review the fundamental concepts. Then, the history and application of the OMA method will be examined comprehensively to observe their role in the development of the technique. Three basic methods of FDD, SSI, and Next will be discussed and the research works associated to them will be analyzed.

Operational modal analysis (OMA) enables the identification of modal characteristics under operational loads and conditions. Traditional frequency-domain methods cannot directly capture modal changes over time, while existing time-frequency representations are not sufficiently interpretable due to spurious modes and implicit parameter design. This paper develops a new OMA method in time-frequency representation based on frequency-domain decomposition (FDD). Short-time FDD and a convolution-based strategy are proposed to obtain singular values and local mode shape similarity, respectively, which are further fused into mode indicators by a fuzzy-based strategy mimicking the modal assurance criterion. The method provides not only a global view of the modal characteristics over time and frequency but also estimates of the modal parameters. It is applicable to strongly nonstationary responses under time-varying loads and conditions. All the parameters explicitly affect the time-frequency representation, and the interpretability is enhanced by including physical information from the user’s prior knowledge in selecting parameters and peak bands. The proposed method is validated based on a study of railway sleepers under train passage. The rigid-body motions and bending modes are identified at frequencies up to 6,500 Hz in laboratory tests and 4,500 Hz in field tests at speeds up to 200 km/h. The identified natural frequencies and mode shapes agree with experimental modal analysis (EMA). The proposed method outperforms EMA in terms of broad frequency range and low measurement cost and can be potentially applied to structural health monitoring under operational conditions.

Frequency Domain Decomposition (FDD) is an operational modal analysis method by which the dynamic parameters of structures are obtained using response signals recorded on several locations of structures. In recent years, many researchers have employed the method to estimate the modal parameters of several Multi-Degree-of-Freedom (MDOF) mechanical and civil systems. The accuracy of the results depends on recording and signal processing parameters such as sampling frequency, windowing, filtering etc. As a result, investigation on the influence of these parameters on the accuracy of the estimates is a good practice. In this research, the uncertainty of modal frequencies obtained from FDD method is analyzed. In order to achieve this goal, the exact values of modal parameters must be available to be compared with the results from FDD method. For this purpose, synthetic signals with random characteristics the same as the ambient vibrations’ are produced and structures with known dynamic parameters are simulated and loaded. The effect of several parameters are investigated in the accuracy of results and proper values and settings are proposed to minimize the errors.

Operational modal analysis (OMA) has been widely used for large structures. However, measured signals are inevitably contaminated with noise and may not be clean enough for identifying the modal parameters with proper accuracy. The traditional methods to estimate modal parameters in noisy situation are usually absorbing the “noise modes” first, and then using the stability diagrams to distinguish the true modes from the “noise modes.” However, it is still difficult to sort out true modes because the “noise modes” will also tend to be stable as the model order increases. This study develops a noise reduction procedure for polyreference complex exponential (PRCE) modal analysis based on ambient vibration responses. In the procedure, natural excitation technique (NExT) is first applied to get free decay responses from measured (noisy) ambient vibration data, and then the noise reduction method based on solving the partially described inverse singular value problem (PDISVP) is implemented to reconstruct a filtered data matrix from the measured data matrix. In our case, the measured data matrix is block Hankel structured, which is constructed based on the free decay responses. The filtered data matrix should maintain the block Hankel structure and be lowered in rank. When the filtered data matrix is obtained, the PRCE method is applied to estimate the modal parameters. The proposed NExT-PDISVP-PRCE scheme is applied to field test of a jacket type offshore platform. Results indicate that the proposed method can improve the accuracy of OMA.

Excessive withdrawal of groundwater for agricultural irrigation can cause seawater intrusion into coastal aquifers. Such a case will in turn results in deterioration of irrigation water quality. Determination of irrigation water quality with traditional methods is a time-consuming and costly process. However, machine learning algorithms can be useful tools for modeling and estimating groundwater quality used for irrigation water purposes. In this study, TDS, PS, SAR, and Cl parameters of groundwater were estimated with models based on EC and pH variables. For this purpose, prediction performances of two different deep learning methods (convolutional neural network (CNN) and deep neural network (DNN)) and two different classical machine learning (Random Forest (RF) and extreme gradient boosting (XGBoost)) methods were compared. In addition, predictive uncertainty of the models was determined by quantile regression (QR) analysis. Performance criteria and results of uncertainty analysis revealed that CNN (in testing phase, NSE = 0.95 for TDS, NSE = 0.96 for PS, NSE = 0.67 for SAR and NSE = 0.93 for CI) and DNN (in testing phase, NSE = 0.91 for TDS, NSE = 0.91 for PS, NSE = 0.57 for SAR and NSE = 0.94 for Cl) models had quite a close performance in estimation of TDS, PS, SAR, and Cl parameters and higher than the other two classical machine learning methods. As a result, the CNN model can be considered the best performing model in estimating all quality parameters due to the highest NSE and lowest RMSE values. In addition, the Taylor diagram showed that the values estimated using the CNN model had the highest correlation with the measured data. It was determined that the model with the lowest uncertainty based on the PICP statistics was DNN, followed by the CNN model. However, the CNN model has predicted outliers more accurately. Present findings proved that deep learning models could offer efficient tools for predicting irrigation water quality parameters.

With the explosive growth in seismic data acquisition and the successful application of deep convolutional neural networks (DCNN) to various image processing tasks within multidisciplinary fields, many researchers have begun to research DCNN based automatic seismic interpretation techniques. Due to the vast number of parameters considered in deep neural networks, deep learning methods usually require a large amount of data for training. However, collecting a large number of expert interpretations is very time consuming, so related research usually uses synthetic datasets and ignores the practical problems of field datasets. In this paper, we open-source a multi-gigabyte expert-labelled field dataset in response to the challenge of accessing large-scale expert-labelled field datasets. We show that 2D fault recognition within this dataset is an image segmentation or edge detection problem in the computer vision field, that can be expressed as a pixel-level fault/non-fault binary classification. Both types of DCNNs are compared, and we propose a novel fault recognition workflow, which involves processing and screening of seismic images and labels, training DCNNs and automatic numerical evaluation. We have also demonstrated for three case study datasets that effective image augmentation methods can reduce the required labelled crosslines while maintaining satisfactory performance. Our experimental results show that our workflow not only outperforms two state-of-the-art DCNN solutions but also achieves performance comparable to humans on an expert labelled image dataset, even predicting subtle faults that an expert interpreter did not annotate. We suggest that the proposed workflow could reduce the fault interpretation life cycle from months to hours and improve the quality, and define the confidence, of fault interpretation results.

Self-mixing interferometry (SMI) is superior to other laser interferometry methods due to its simplicity and compactness. However, SMI signals are often complex and difficult to process due to interference in the form of variations in the effective reflectivity of the target, noisy signals, complex signal shapes, and other dependencies. Deep neural networks have been a very popular area of research in computer artificial intelligence in recent years, allowing more implicit features to be uncovered than traditional shallow machine learning. It has been shown that convolutional and back-propagation neural networks can be used for SMI signal processing. There are also studies that have used machine learning genetic algorithms for absolute distance measurement. Based on the above, this study used convolutional neural networks to form a deep neural network for absolute distance measurement based on SMI technology. We first trained the deep convolutional neural network at different feedback strengths. The results of the Convolutional Neural Network (CNN) model showed a coefficient of determination of 0.9987. which is consistent with the required model. The trained network was then used to estimate absolute distances with and without the addition of noise. The comparison proves that the proposed method is noise-proof and has high adaptability for measurements under different conditions.