Papers by Keyword: Damage Detection

Abstract: Signal processing method based on wavelet transform used in non-linear acoustic test is presented in the paper. The method is applied for sidebands identification in response signal acquired during vibro-acoustic modulation test of impacted carbon fiber reinforced plate (CFRP). The plate was impacted with known energy using drop-weight testing machine. The modulation effect in investigated specimen results from the interaction of low and high frequency excitation with damage. The paper investigates different than mono-harmonic low-frequency excitation usually used in non-linear acoustics tests. Application of aperiodic low-frequency excitation signal allows to omit the modal test, where natural frequency of the structure are estimated. However, this requires the use of dedicated signal processing methods.

Abstract: Principal Component Analysis (PCA) and Wavelet Transform (WT) aretwo well-known signal processing tools that are widely used indifferent fields. PCA playsa vital role in statistical analysis as a dimensional reduction tool. Besides, WT has proven its abilityto overcome many of the limitation of the others among various time-frequencyanalyzers. The present work attempts to use the properties and advantagesof both methodologies together in damage detection. To achieve thisaim, PCA is applied on ridges of wavelet transform of measured signalsfrom the structure. The results show that the proposed combination improvesthe accuracy of detection comparing with PCA damage detection basedon original data captured from sensors. According to the result, when PCA uses the ridges of transformed data, theidentifications of damages are more clear and accurate. This work involvesexperiments with an aluminum beam using piezoelectrictransducers as sensors and actuators. Damages are introduced intothe structure as a cut in several steps enlarging the depthof cut.

Abstract: Damage Detection problem in Structural Health Monitoring (SHM) is widely studied by many researchers, therefore lots of damage detection algorithms can be found in the literature. Feature Selection / Extraction methods are essential in the accuracy of these algorithms, they provide the suitable data to be used. The main goal of this work is to improve the input data to be the most representative for the damage detection problem. This is done using different Feature Selection / Extraction methods (PCA, UmRMR, and a combination of both). After taking the representative features, the results are tested using a damage detection method; the NullSpace in this case. The data has been collected from a Laboratory Offshore tower model. The different results are compared (different preprocessing vs Raw data) and these show how the correct preselection of the data can improve damage detection.

Abstract: The use of changes in vibration properties for global damage detection and monitoring of existing concrete structures has received great research attention in the last three decades. To track changes in vibration properties experimentally, structures have been artificially damaged by a variety of scenarios. However, this procedure does not represent realistically the whole design-life degradation of concrete structures. This paper presents experimental work on a set of damaged reinforced concrete beams due to different loading regimes to assess the sensitivity of linear and non-linear vibration characteristics. Of the total set, three beams were subject to incremental static loading up to failure to simulate overloading, and two beams subject to 15 million loading cycles with varying amplitudes to produce an accelerated whole-life degradation scenario. To assess the vibration behaviour in both cases, swept sine and harmonic excitations were conducted at every damage level. The results show that resonant frequencies are not sensitive enough to damage due to cyclic loading, whereas cosh spectral and root mean square distances are more sensitive, yet more scattered. In addition, changes in non-linearity follow a softening trend for beams under incremental static loading and are significantly inconsistent for beams under cyclic loading. Amongst all examined characteristics, changes in modal stiffness are found to be most sensitive to damage and least scattered, but modal stiffness is tedious to compute due mainly to the difficulty of constructing restoring force surfaces from field measurements.

Abstract: The ability of a Structural Health Monitoring (SHM) system to automatically identify damage in a composite structure is a vital requirement demanded by end-users of such systems. This paper presents the demonstration of a potential method. A composite fatigue specimen was manufactured and initially tested at 1Hz for 1000 cycles. Acoustic emission (AE) signals were recorded for complete fatigue cycles periodically in order to establish a base-line associated with undamaged specimens. The specimen was then subjected to impact damage to create barely-visible impact damage (BVID) and subjected to further fatigue cycles with acoustic emission recorded until failure. The data was subsequently analysed using a range of techniques including basic RMS signal levels and frequency-based analysis. At various stages during the test, C-scanning was used to validate the results obtained. Results demonstrated that AE is capable of detecting BVID in composite materials under fatigue loading. The proposed method has wide applicability to composite structures which are subjected to cyclic loading, such as wind turbine blades.

Abstract: Composite materials are increasingly being used in a wide range of structural applications in place of metallic materials. This presents a new range of challenges when considering the monitoring of damage and failure in complex components. This paper explores these challenges and presents a potential monitoring method using airborne acoustics which is both non-contact and easily implemented. A carbon composite panel was manufactured and statically loaded in tension until failure. During the test, Digital Image Correlation (DIC) was used to measure full field surface strain in the panel. An array of microphones, placed adjacent to the panel, was used to capture airborne acoustic signals between 400Hz and 20kHz during the test. The captured sound waves potentially contain signals originating from a range of sources, such as fibre failures and matrix cracking, but also contain background noise. A range of techniques have been used to examine the signals and determine the onset of failure, including Short-Time Fourier Transforms (STFT). The detection of failure using the airborne acoustic system has been validated using the strain data from the DIC measurements. The results presented demonstrate the applicability of the airborne system to monitoring of composite components.

Abstract: Many of the bridges currently in use worldwide are approaching the end of their design lives. However, rehabilitating and extending the lives of these structures raises important safety issues. There is also a need for increased monitoring which has considerable cost implications for bridge management systems. Existing structural health monitoring (SHM) techniques include vibration-based approaches which typically involve direct instrumentation of the bridge and are important as they can indicate the deterioration of the bridge condition. However, they can be labour intensive and expensive. In the past decade, alternative indirect vibration-based approaches which utilise the response of a vehicle passing over a bridge have been developed. This paper investigates such an approach; a low-cost approach for the monitoring of bridge structures which consists of the use of a vehicle fitted with accelerometers on its axles. The approach aims to detect damage in the bridge while obviating the need for direct instrumentation of the bridge. Here, the effectiveness of the approach in detecting damage in a bridge is investigated using a simplified vehicle-bridge interaction (VBI) model in theoretical simulations and a scaled VBI model in a laboratory experiment. In order to identify the existence and location of damage, the vehicle accelerations are recorded and processed using a continuous Morlet wavelet transform and a damage index is established. A parametric study is carried out to investigate the effect of parameters such as the bridge span length, vehicle speed, vehicle mass, damage level and road surface roughness on the accuracy of results.

Abstract: High energy consumption, excessive data storage and transfer requirements are prevailing issues associated with structural health monitoring (SHM) systems, especially with those employing wireless sensors. Data compression is one of the techniques being explored to mitigate the effects of these issues. Compressive sensing (CS) introduces a means of reproducing a signal with a much less number of samples than the Nyquist's rate, reducing the energy consumption, data storage and transfer cost. This paper explores the applicability of CS for SHM, in particular for damage detection and localization. CS is implemented in a simulated environment to compress SHM data. The reconstructed signal is verified for accuracy using structural response data obtained from a series of tests carried out on a reinforced concrete (RC) slab. Results show that the reconstruction was close, but not exact as a consequence of the noise associated with the responses. However, further analysis using the reconstructed signal provided successful damage detection and localization results, showing that although the reconstruction using CS is not exact, it is sufficient to provide the crucial information of the existence and location of damage.

Abstract: The central target of this work is to provide an alternative to machine learning approaches to structural health monitoring with one of robust multivariate statistic novelty detection. Damage detection and identification is a procedure that is hierarchical in nature. At its most sophisticated, diagnosis of the damage could include localisation, classification and severity assessment and even go so far as to estimate the time-to-failure of the structure. In this paper, robust multivariate statistics were investigated focused mainly on a high level estimation of the outliers which determines only the presence or absence of novelty - something that is of fundamental interest. These methods allow a diagnosis of deviation from normality and the option of identifying the presence of masking effects caused by multiple outliers. This paper is trying to introduce a new scheme for damage detection by adopting simple measurements and exploiting robust multivariate statistics.

Abstract: Cracking is a common type of failure in machines and structures. Cracks must be detected at an early stage before catastrophic failure. In structural health monitoring, changes in the vibration characteristics of the structure can be utilized in damage detection. A fatigue crack with alternating contact and non-contact phases results in a non-linear behaviour. This type of damage was simulated with a finite element model of a simply supported beam. The structure was monitored with a sensor array measuring transverse accelerations under random excitation. The objective was to determine the smallest crack length that can be detected. The effect of the sensor locations was also studied. Damage detection was performed using the generalized likelihood ratio test (GLRT) in time domain followed by principal component analysis (PCA). Extreme value statistics (EVS) were used for novelty detection. It was found that a crack in the bottom of the midspan could be detected once the crack length exceeded 10% of the beam height. The crack was correctly localized using the monitoring data.