K. Krishnan Nair

Time series based structural damage detection algorithm using gaussian mixtures modeling

In this paper, a time series based detection algorithm is proposed utilizing the Gaussian Mixture Models. The two critical aspects of damage diagnosis that are investigated are detection and extent. The vibration signals obtained from the structure are modeled as autoregressive moving average (ARMA) processes. The feature vector used consists of the first three autoregressive coefficients obtained from the modeling of the vibration signals. Damage is detected by observing a migration of the extracted AR coefficients with damage. A Gaussian Mixture Model (GMM) is used to model the feature vector. Damage is detected using the gap statistic, which ascertains the optimal number of mixtures in a particular dataset. The Mahalanobis distance between the mixture in question and the baseline (undamaged) mixture is a good indicator of damage extent. Application cases from the ASCE Benchmark Structure simulated data have been used to test the efficacy of the algorithm. This approach provides a useful framework for data fusion, where different measurements such as strains, temperature, and humidity could be used for a more robust damage decision.

Dynamic studies of railtrack sleepers in a track structure system

Abstract This paper discusses the dynamic response of a typical prestressed concrete railtrack sleeper due to wheel–track interaction dynamics, involving wheel and rail imperfections, under various parametric conditions. The interaction dynamics of the vehicle and track is first carried out in the time domain using MSC/NASTRAN. Using the resulting load time histories on an isolated sleeper, a detailed finite element model of the sleeper is used to analyze its dynamic behaviour. The dynamic amplification factors for deflection, ballast pressure and bending moments have been evaluated at the critical section (rail-seat and centre) for various exciting frequencies under different vehicle–track parametric conditions. The results provide a basis for improved and rational design of the sleeper.

Use of Wavelet-Based Damage-Sensitive Features for Structural Damage Diagnosis Using Strong Motion Data

This paper introduces three wavelet-based damage-sensitive features (DSFs) extracted from structural responses recorded during earthquakes to diagnose structural damage. Because earthquake excitations are nonstationary, the wavelet transform, which represents data as a weighted sum of time-localized waves, is used to model the structural responses. These DSFs are defined as functions of wavelet energies at particular frequencies and specific times. The first DSF (DSF1) indicates how the wavelet energy at the original natural frequency of the structure changes as the damage progresses. The second DSF (DSF2) indicates how much the wavelet energy is spread out in time. The third DSF (DSF3) reflects how slowly the wavelet energy decays with time. The performance of these DSFs is validated using two sets of shake-table test data. The results show that as the damage extent increases, the DSF1 value decreases and the DSF2 and DSF3 values increase. Thus, these DSFs can be used to diagnose structural damage. The r...

The application of statistical pattern recognition methods for damage detection to field data

Recent studies in structural health monitoring have shown that damage detection algorithms based on statistical pattern recognition techniques for ambient vibrations can be used to successfully detect damage in simulated models. However, these algorithms have not been tested on full-scale civil structures, because such data are not readily available. A unique opportunity for examining the effectiveness of these algorithms was presented when data were systematically collected from a progressive damage field test on the Z24 bridge in Switzerland. This paper presents the analysis of these data using an autoregressive algorithm for damage detection, localization, and quantification. Although analyses of previously obtained experimental or numerically simulated data have provided consistently positive diagnosis results, field data from the Z24 bridge show that damage is consistently detected, however not well localized or quantified, with the current diagnostic methods. Difficulties with data collection in the field are also revealed, pointing to the need for careful signal conditioning prior to algorithm application. Furthermore, interpretation of the final results is made difficult due to the lack of detailed documentation on the testing procedure.

Time synchronization algorithms for wireless monitoring system

Wireless health monitoring schemes are innovative techniques, which effectively remove the disadvantages associated with current wire-based sensing systems, i.e., high installation and upkeep costs. However, recorded data sets may have relative time-delays due to the blockage of sensors or inherent internal clock errors. In this paper, two algorithms are proposed for the synchronization of the recorded asynchronous data measured from sensing units of a wireless monitoring system. In the first algorithm, the input signal to a structure is measured. Time-delay between an output measurement and the input is identified based on the minimization of errors of the ARX (auto-regressive model with exogenous input) models for the input-output pair recordings. The second algorithm is applicable when a structure is subject to ambient excitation and only output measurements are available. ARMAV (auto-regressive moving average vector) models are constructed from two output signals and the time-delay between them is evaluated based on the minimization of errors of the ARMAV models. The proposed algorithms are verified by simulation data and recorded seismic response data from multi-story buildings.

EVALUATION OF DYNAMIC LOAD ON RAILTRACK SLEEPERS BASED ON VEHICLE-TRACK MODELING AND ANALYSIS

The present study reports the results of a rigorous dynamic interaction analysis that accounts for the vehicle-track characteristics and rail imperfections. In order to perform a rigorous dynamic analysis, a model involving all components of the track structure and vehicle parameters are required. A vehicle model (conforming to Indian Railways) with 17 degrees of freedom has been considered. The track model consists of rail, rail-pad, sleeper, ballast, sub-ballast, subsoil and a track length encompassing 12 prestressed concrete sleepers. The dynamic interactive analysis is carried out between the vehicle and track in the time domain using a finite element software. The results of the interactive analysis give responses in the form of reaction time histories at the rail-seat locations during the passage of vehicle. A parametric study is carried out to assess the influence of different track parameters on the dynamic load on the railtrack sleepers. Based on this study, suitable load amplification factors are arrived at to facilitate an improved design basis for an equivalent static analysis in practical designs of sleepers.

Application of statistical pattern classification methods for damage detection to field data

The field of Structural Health Monitoring (SHM) has received considerable attention for its potential applications to monitoring civil infrastructure. However, the damage detection algorithms that form the backbone of these systems have primarily been tested on simulated data instead of full-scale structures because of the scarcity of real structural acceleration data. In response to this deficiency in testing, we present the performance of two damage detection algorithms used with ambient acceleration data collected during the staged demolition of the fullscale Z24 Bridge in Switzerland. The algorithms use autoregressive coefficients as features of the acceleration data and hypothesis testing and Gaussian Mixture Modeling to detect and quantify damage. While experimental or numerically simulated data have provided consistently positive results, field data from real structures, the Z24 Bridge, show that there can be significant false positives in the predictions. Difficulties with data collection in the field are also revealed pointing to the need for careful signal conditioning prior to algorithm application.