Dorotea Sigurdardottir

Neutral axis as damage sensitive feature

Structural health monitoring (SHM) is the process of continuously or periodically measuring structural parameters and the transformation of the collected data into information on real structural conditions. The centroid of stiffness is a universal parameter and its position in a cross-section can be evaluated for any load-carrying beam structure as the position of the neutral axis under conveniently chosen loads. Thus, a change in the position of the neutral axis within a cross-section can indicate a change in the position of the centroid of stiffness, i.e., unusual structural behaviors. This paper proposes a novel monitoring method based on deterministic and probabilistic determination of the position of the neutral axis under conveniently chosen conditions. Therefore, the method proposed in this paper is potentially applicable to a large variety of beam-like structures. Data from two existing structures were used to validate the method and assess its performance: Streicker Bridge at Princeton University and the US202/NJ23 highway overpass in Wayne, NJ. The results show that the neutral axis location is varying even when damage is not present. Reasons for this variation are determined and the accuracy in the evaluation assessed. This paper concludes that the position of the neutral axis can be evaluated with sufficient accuracy using static and dynamic strain measurements performed on appropriate time-scales and indicates its potential to be used as a damage sensitive feature.

Detecting minute damage in beam-like structures using the neutral axis location

The neutral axis is the curve within a cross-section of a beam under loading at which the normal stress and strain vanish. The neutral axis is correlated with the centroid of stiffness which depends on material and geometric properties of the beam cross-section. The centroid is sensitive to changes in area and stiffness of the cross-section, i.e. damage. The neutral axis is a universal parameter in beam-like structures, and since the majority of structural systems is based on beams, the neutral axis location can be used as a damage sensitive feature in a variety of structures. In this paper the sensitivity of the neutral axis location to minute damage is studied through a test structure with induced artificial damage. The structure was equipped with long-gauge fiber-optic strain sensors in three cross-sections, one with partial delamination in the concrete, one with a small crack, and one with no damage (healthy). The structure was monitored during lifting and transport, and dynamic tests were performed in close-to-real conditions. A detailed uncertainty analysis was performed both on the evaluations of the centroid location and the neutral axis measurements. A Z-score hypothesis test was used to set thresholds and determine whether the damage was detected or not. The results show that the neutral axis location is sensitive to minute damage, and that the latter can be detected within the uncertainty limits in on-site conditions.

Damage detection and characterization using long-gauge and distributed fiber optic sensors

Abstract. Fiber optic strain sensors have significantly evolved and have reached their market maturity during the last decade. Their widely recognized advantages are high precision, long-term stability, and durability. In addition to these benefits, fiber optic (FO) techniques allow for affordable instrumentation of large areas of civil structures and infrastructure enabling global large-scale monitoring based on long-gauge sensors, and integrity monitoring based on distributed sensors. The FO techniques that enable these two approaches are based on fiber Bragg-gratings and Brillouin optical time-domain analysis. The aim of this paper is to present both FO techniques and both structural assessment approaches, and to validate them through large-scale applications. Although many other currently applied methods fail to detect the damage in real, on-site conditions, the presented approaches were proven to be suitable for damage detection and characterization, i.e., damage localization and, to certain extent, quantification. This is illustrated by two applications presented in detail in this paper: the first on a post-tensioned concrete bridge and the second on segmented concrete pipeline.

Damage detection and characterization using fiber optic sensors

Fiber optic sensors (FOS) have significantly evolved and have reached their market maturity during the last decade. Their widely recognized advantages are high precision, long-term stability, and durability. But in addition to these advantageous performances, FOS technologies allow for affordable instrumentation of large areas of structure enabling global large-scale monitoring based on long-gauge sensors and integrity monitoring based on distributed sensors. These two approaches are particularly suitable for damage detection and characterization, i.e., damage localization and to certain extent quantification and propagation, as illustrated by two applications presented in detail in this paper: post-tensioned concrete bridge and segmented concrete pipeline. Early age cracking was detected, localized and quantified in the concrete deck of a pedestrian bridge using embedded long-gauge FOS. Post-tensioning of deck closed the cracks; however, permanent weakening in a bridge joint occurred due to cracking and it was identified and quantified. The damage was confirmed using embedded distributed FOS and a separate load test of the bridge. Real-size concrete pipeline specimens and surrounding soil were equipped with distributed FOS and exposed to permanent ground displacement in a large-scale testing facility. Two tests were performed on different pipeline specimens. The sensors bonded on the pipeline specimens successfully detected and localized rupture of pipeline joints, while the sensors embedded in the soil were able to detect and localize the failure plane. Comparison with strain-gauges installed on the pipeline and visual inspection after the test confirmed accurate damage detection and characterization.

On predicting monitoring system effectiveness

While the objective of structural design is to achieve stability with an appropriate level of reliability, the design of systems for structural health monitoring is performed to identify a configuration that enables acquisition of data with an appropriate level of accuracy in order to understand the performance of a structure or its condition state. However, a rational standardized approach for monitoring system design is not fully available. Hence, when engineers design a monitoring system, their approach is often heuristic with performance evaluation based on experience, rather than on quantitative analysis. In this contribution, we propose a probabilistic model for the estimation of monitoring system effectiveness based on information available in prior condition, i.e. before acquiring empirical data. The presented model is developed considering the analogy between structural design and monitoring system design. We assume that the effectiveness can be evaluated based on the prediction of the posterior variance or covariance matrix of the state parameters, which we assume to be defined in a continuous space. Since the empirical measurements are not available in prior condition, the estimation of the posterior variance or covariance matrix is performed considering the measurements as a stochastic variable. Moreover, the model takes into account the effects of nuisance parameters, which are stochastic parameters that affect the observations but cannot be estimated using monitoring data. Finally, we present an application of the proposed model to a real structure. The results show how the model enables engineers to predict whether a sensor configuration satisfies the required performance.

Performance and damage evolution of plain and fibre-reinforced segmental concrete pipelines subjected to transverse permanent ground displacement

Abstract This paper presents the results of three full-scale experiments performed on segmental concrete pipelines subjected to permanent ground displacement. The first pipeline was made of reinforced concrete pipes and the second pipeline was made of steel fibre-reinforced concrete pipes. The third pipeline was made of a combination of fibre-reinforced and reinforced concrete pipes. An array of sensing techniques was used to assess the damage evolution in pipelines and their overall performance. Three stages of damage were observed. In the first stage, damage was concentrated in the joints near the fault line. In the second stage, the damage occurred in all joints along the pipeline. While in the first two stages damage was mainly concentrated at the bell and spigot joints of the pipe segments, the third stage of damage was characterised by severe damage and rupture of the body of pipe segments located in the immediate vicinity of the fault line. The modes of failure for the plain and fibre-reinforced concrete pipelines were similar in the first and second stages of damage. However, in the pipeline constructed using both plain and fibre-reinforced concrete pipe segments, the damage was concentrated in the standard reinforced concrete pipe segments.

Challenges in Damage Detection Based on Finite Element Analyses and Monitoring of Dynamic Curvature of Concrete-steel Composite Structures

This work researches damage detection in composite concrete-steel structures that are typical for highway overpasses and bridges. The techniques herein proposed assume that typical damage in the deck occurs in form of delamination and cracking, and that it affects the peak PSD of curvature. The investigation is performed by combining results of measurements collected by long-gauge fiber optic strain sensors installed on monitored structure and finite element analysis. The finite element model has been prepared by using design parameters and validated by comparing it with dynamic frequencies obtained from the monitoring system. The method was able to detect unusual behavior through probabilistic study of the peak PSD. doi: 10.12783/SHM2015/35

Evaluation of the use of fiber optic sensors in identification of fresco fracturing patterns

Ageing of materials and extreme events tend to damage structures, and ancient historical monuments are particularly vulnerable due to their age and long-term exposure to adverse events and influences. As an example, the wall paintings (frescoes) from the seventeenth century BCE found at the archaeological site of Akrotiri (Santorini, Greece) were recovered from volcanic ash in fragments with dimensions ranging from a few centimeters to a few decimeters. Identification of the fracturing patterns is helpful to the process of piecing together the fragments of frescos. Previous work has involved looking at fracturing patterns in frescos that have been reassembled. Recent work has looked at the process by which fractures develop. Current identification techniques involve experimental study of fracture development on plaster molds using a high-speed camera combined with sophisticated algorithms for pattern recognition. However, the use of a high-speed camera is challenging due to very demanding data processing and analysis and some inaccuracies in identification of fracture initialization generated by light conditions. This paper aims to evaluate whether or not short-gauge fiber optic sensors (FOS) based on Fiber Brag-Gratings (FBG), can be used to help identify the fracturing patterns of falling frescoes as a complement to high-speed cameras. In total four tests were performed using surface and embedded sensors on various plaster molds. The data taken by sensors installed on the surface of the mold were more complex to analyze and interpret than the data taken by embedded sensors, since the former reflected combined influence from fracture and bending. While their practicality is challenged by cost, moderately dense arrays of embedded FOS are found to be a plausible complement to the high speed-camera in the experiments.