Adam Santos

A novel unsupervised approach based on a genetic algorithm for structural damage detection in bridges

This paper proposes a novel unsupervised and nonparametric genetic algorithm for decision boundary analysis (GADBA) to support the structural damage detection process, even in the presence of linear and nonlinear effects caused by operational and environmental variability. This approach is rooted in the search of an optimal number of clusters in the feature space, representing the main state conditions of a structural system, also known as the main structural components. This genetic-based clustering approach is supported by a novel concentric hypersphere algorithm to regularize the number of clusters and mitigate the cluster redundancy. The superiority of the GADBA is compared to state-of-the-art approaches based on the Gaussian mixture models and the Mahalanobis squared distance, on data sets from monitoring systems installed on two bridges: the Z-24 Bridge and the Tamar Bridge. The results demonstrate that the proposed approach is more efficient in the task of fitting the normal condition and its structural components. This technique also revealed to have better classification performance than the alternative ones in terms of false-positive and false-negative indications of damage, suggesting its applicability for real-world structural health monitoring applications.

A global expectation-maximization based on memetic swarm optimization for structural damage detection

During the service life of engineering structures, structural management systems attempt to manage all the information derived from regular inspections, evaluations and maintenance activities. However, the structural management systems still rely deeply on qualitative and visual inspections, which may impact the structural evaluation and, consequently, the maintenance decisions as well as the avoidance of collapses. Meanwhile, structural health monitoring arises as an effective discipline to aid the structural management, providing more reliable and quantitative information; herein, the machine learning algorithms have been implemented to expose structural anomalies from monitoring data. In particular, the Gaussian mixture models, supported by the expectation-maximization (EM) algorithm for parameter estimation, have been proposed to model the main clusters that correspond to the normal and stable state conditions of a structure when influenced by several sources of operational and environmental variations. Unfortunately, the optimal parameters determined by the EM algorithm are heavily dependent on the choice of the initial parameters. Therefore, this paper proposes a memetic algorithm based on particle swarm optimization (PSO) to improve the stability and reliability of the EM algorithm, a global EM (GEM-PSO), in searching for the optimal number of components (or data clusters) and their parameters, which enhances the damage classification performance. The superiority of the GEM-PSO approach over the state-of-the-art ones is attested on damage detection strategies implemented through the Mahalanobis and Euclidean distances, which permit one to track the outlier formation in relation to the main clusters, using real-world data sets from the Z-24 Bridge (Switzerland) and Tamar Bridge (United Kingdom).

Applicability of linear and nonlinear principal component analysis for damage detection

The goal of this work is to detect structural damage using vibration-based damage identification approaches even when the damage-sensitive features are camouflaged by the presence of operational and environmental conditions. For feature classification purposes, four machine learning algorithms are applied based on the principal component analysis (PCA), nonlinear PCA, kernel PCA and greedy kernel PCA. Time-series data from an array of accelerometers under several structural state conditions were obtained from a well-known base-excited three-story frame structure. The main contribution of this work is the applicability of those PCA-based algorithms, for damage detection, in the presence of operational and environmental effects. For these specific data sets, one can infer that the greedy kernel PCA algorithm is more appropriate when one wants to minimize false-positive indications of damage without increasing, significantly, the false-negative indications of damage.

Clustering Studies for Damage Detection in Bridges: A Comparison Study

Improved and more continuous condition assessment of bridges has been demanded by our society to better face challenges presented by aging civil infrastructure. In the last two decades, bridge condition assessment techniques have been developed in order to improve the maintenance of bridges in a systematic way. The Structural Health Monitoring (SHM) has given the ability to provide information, in real time, about the performance of the structural system. However, the reliability of that information depends highly on the quality of the data analysis, as the operational and environmental variability introduces changes into the system that can mask those changes related with damage. Therefore, this paper intends to evaluate the performance of several algorithms for clustering strategies in vibration-based damage detection under operational and environmental effects. For statistical modeling and feature classification are proposed K-means, Gaussian Mixture Models (GMM), Support Vector Clustering (SVC) and Self-Organizing Maps (SOM) algorithms. The study is performed on standard data sets from the Tamar Suspension Bridge, in England, and the Z-24 Bridge in Switzerland. The contribution of this work is the applicability of the proposed algorithms for clustering strategies in damage detection as well as the comparison of the classification performance between these algorithms. doi: 10.12783/SHM2015/146

A Global Expectation–Maximization Approach Based on Memetic Algorithm for Vibration-Based Structural Damage Detection

This paper proposes a novel unsupervised damage detection approach based on a memetic algorithm that establishes the normal or undamaged condition of a structural system as data clusters through a global xpectation–maximization technique, using only damage-sensitive features extracted from output-only vibration measurements. The health state is then discriminated by considering the Mahalanobis squared distance between the learned clusters and a new observation. The proposed approach is compared with state-of-the-art ones by taking into account real-world data sets from the Z-24 Bridge (Switzerland), where several damage scenarios were performed. The results indicated that the proposed approach can be applied in structural health monitoring applications where life safety, economic, and reliability issues are the most important motivations to consider.

Deep principal component analysis: An enhanced approach for structural damage identification

The structural health monitoring relies on the continuous observation of a dynamic system over time to identify its actual condition, detect abnormal behaviors, and predict future states. The regular changes in environmental factors have been reported as one of the main challenges for the application of structural health monitoring systems. These influences in the structural responses are in general nonlinear, affecting the damage-sensitive features in the most varied forms. The usual process to remove these normal changes is referred to as data normalization. In that regard, principal component analysis is probably the most studied algorithm in structural health monitoring, having numerous versions to learn strong nonlinear normal changes. However, in most cases, not all variability is properly accounted for via the existing nonlinear principal component analysis approaches, resulting in poor damage detection and quantification performances. In this article, a new paradigm based on deep principal component analysis, rooted in the deep learning field, is presented to overcome these limitations. This approach extracts the most salient underlying feature distributions by stacking multiple feedforward neural networks trained to learn an identity mapping of the input variables, where the network inputs are reproduced into the outputs. Similar to the traditional nonlinear principal component analysis–based approach, our approach identifies a nonlinear output-only model of an undamaged structure by comprising modal features into an internal bottleneck layer, which implicitly represents the independent environmental factors. The proposed technique is validated through the application on a progressively damaged prestressed concrete bridge and a three-span suspension bridge. The experimental results demonstrate that capturing the most slight nonlinear variations in the data can lead to improved data normalization and, consequently, better damage detection and quantification performances.

A semi-autonomous particle swarm optimizer based on gradient information and diversity control for global optimization

Abstract The deterministic optimization algorithms far outweigh the non-deterministic ones on unimodal functions. However, classical algorithms, such as gradient descent and Newtons method, are strongly dependent on the quality of the initial guess and easily get trapped into local optima of multimodal functions. On the contrary, non-deterministic optimization methods, such as particle swarm optimization and genetic algorithms perform global optimization, however they waste computational time wandering the search space as a result of the random walks influence. This paper presents a semi-autonomous particle swarm optimizer, termed SAPSO, which uses a gradient-based information and diversity control to optimize multimodal functions. The proposed algorithm avoids the drawbacks of deterministic and non-deterministic approaches, by reducing computational efforts of local investigation (fast exploitation with gradient information) and escaping from local optima (exploration with diversity control). The experiments revealed promising results when SAPSO is applied on a suite of test functions based on De Jongs benchmark optimization problems and compared to other PSO-based algorithms.

A Generalized Approach to Integrate Machine Learning, Finite Element Modeling and Monitoring Data for Bridges

In the last decades, the structural health monitoring (SHM) of civil structures has been performed arguably based on two approaches: model- and data-based. The former approach tries to identify damage by relating the measured data from the structure to the prediction of physics-based numerical models tailored for the same structure. The latter one is a data-driven modeling approach, where measured data from a given state condition is compared to the baseline condition. The data-based approach has been rooted in the machine learning field, where machine learning algorithms are essential to learn the structural behavior from the past data, and to perform pattern recognition for damage identification. In the SHM field, this approach has been known as the statistical pattern recognition paradigm. Basically, in both approaches, the identification of damage requires data comparison between two state conditions, the baseline and a damaged condition; thus in a general sense, those two approaches make use of pattern recognition techniques. This paper intends to step forward through the combination of machine learning, finite element modeling and monitoring data from the Z-24 Bridge in one unique damage detection approach. To achieve this combination, data from simulated undamaged and damaged scenarios can be introduced into the learning process using predictions from finite element models.

Effect of periodic cable nonuniformities on transmission measurements

This paper presents a likely explanation for dips observed in transmission coefficient measurements (S21) for twisted-pair copper cables. We show via computer simulations that periodic cable nonuniformities behave like selective filters, prohibiting electromagnetic wave to propagate at certain frequencies. Similar effect is achieved in optics using fiber Bragg grating (FBG) and may be analyzed as such. The simulation results indicate that periodic cable nonuniformities acting like a FBG is a plausible reasoning to explain the presence of dips.

Data Management System for Structural Health Monitoring

Optical sensors have found application in many fields, such as in Civil Engineering, Aeronautics, Energy and Oil & Gas Industries. Monitoring solutions based on this technology have proven particularly cost effective and can be applied to large scale structures where hundreds of sensors must be deployed for long term measurements of different mechanical and physical parameters. Sensors based on Fiber Bragg Gratings (FBGs) are the most common solution used in Structural Health Monitoring (SHM) and the measurements are performed by instruments known as optical interrogators. Acquisition rates increasingly higher have been possible using the latest optical interrogators, which gives rise to a large volume of data whose processing and storage can demand special softwares. This work presents the Interrogator Abstraction (InterAB) software for these purposes. The results obtained during tests in laboratory and real environment demonstrate the efficiency and flexibility of this software for different types of sensors and optical interrogators.