Inka Buethe

Online Simultaneous Reconstruction of Wind Load and Structural Responses—Theory and Application to Canton Tower

The actual wind load information is helpful in evaluating the health status of high-rise structures. However, as a type of distributed load, the wind load is very difficult to be measured directly. A possible solution is to reconstruct it from the structural response measurements. This is often an ill-posed inverse problem. In this article, such ill posedness is solved by using a stable input estimator. With the help of the proposed application-oriented algorithm selection guidance, a type of state and input estimator is formulated. This type of estimator is designed based on the Kalman filter scheme, and is capable of estimating the unknown inputs and the system states within one sampling time. This actually facilitates the online simultaneous reconstruction of the wind load and the structural responses. The 600 m tall Canton Tower is situated in a typhoon active area, and a structural health monitoring system has already been integrated onto this tower. These two points make the Canton Tower an ideal test bed for validating the above illustrated online reconstruction strategy for the wind load and the structural responses. An operational modal analysis (OMA) is first performed to identify the modal properties of the Canton Tower under the Typhoon Nanmadol in 2011. Then a reduced-order finite element model (FEM) of the Canton Tower is updated according the OMA results. Finally, the equivalent fluctuating lateral loads and moments, which act on the nodes of the reduced-order FEM are reconstructed using the acceleration measurements recorded during Tyhoon Kai-tak in 2012. The reconstruction results are validated by comparing the simultaneously reconstructed structural acceleration with the corresponding sensor measurements. The mean component of the loads and moments are calculated using the real-time wind speed measurements and the available aerodynamic force coefficients. It is noted here that the focus of this article is not to develop a totally new theory, but rather to explore the application of a state and input estimator in the foreground to a practical complex structure.

Applications of Self-Organizing Maps in Structural Health Monitoring

The Structural Health Monitoring process includes several steps like feature extraction and probabilistic decision making, which need some form of data fusion and information condensation. These take place after data acquisition and before being able to decide, if a monitored structure has faced damage. Although feature selection is an important step, the processing and suitable preparation of these data are significant, influencing the potential of decision making in various ways. With Self-Organizing Maps (SOM) a multi-purpose instrument for these tasks of pattern recognition and data interpretation is presented here. Self-Organizing Maps belong to the group of artificial neural networks and by using the special map character provide the opportunity of additional visualization. Especially when monitoring a structure over a long period of time, environmental changes often occur, which can mask the effects of damage on the dynamic behavior of the structures. As one potential application of SOM, the possibility of distinguishing between environmental changes and damage of the structure is shown. In this application a self-organizing network is trained with data of the undamaged structure and via calculation of the distance to the map a damage indicator is developed. Moreover, the distinction between different damage modes of piezoelectric sensors is presented using SOM as a tool of pattern recognition and visualization. This application uses data recorded from different damage modes extracted from one specimen of a piezoelectric element. The trained network can be compared with other piezoelectric elements mounted in a similar way to be able to detect possible sensor damage. This helps avoiding false alarms even under changing environmental conditions. Both applications have been successfully used to analyze experimental data on coupon level showing the applicability of the presented concepts.

Model-based detection of sensor faults under changing temperature conditions:

Nowadays, a variety of civil, aeronautic or mechanical engineering structures need regular inspections. While non-destructive testing has become state of the art, there is a trend towards online testing on demand using built-in sensors. A vast amount of in-service monitoring, as part of structural health monitoring, uses piezoelectric elements. Their deployment requires a control of the sensor performance in order to prevent false alarm. Several kinds of damage can influence the sensor performance, for example, degradation of piezoceramic or adhesive, debonding or breakage of the element. Therefore, this paper aims to detect damage of circular piezoelectric elements and their bonding layers during in-service monitoring. For this purpose, a local method, using the coupled electro-mechanical admittance, is proposed. However, changing environmental and operational conditions such as temperature have an effect on the measured quantities, masking the damage. To cope with this, the dynamic behaviour of an attached piezoelectric element including temperature trends is modelled to enable physics-based temperature compensation. This novel combination of an improved analytical model, updated according to impedance measurements, realizes a method less dependent on experimental baseline data, than the sole comparison of experimental data, established in many monitoring systems. By comparison with experimentally obtained electro-mechanical susceptance the physical model of circular piezoelectric elements is validated. The feasibility of the proposed method is presented employing experimental data of piezoelectric elements mounted on aluminium coupons, partly damaged with degradation and sensor breakage. The application produces promising results, detecting the created defects.

Vibration-Based Damage Detection under Changing Environmental and Operational Conditions

Structural Health Monitoring (SHM) allows to perform a diagnosis on demand which assists the operator to plan his future maintenance or repair activities. Using structural vibrations to extract damage sensitive features, problems can arise due to variations of the dynamical properties with changing environmental and operational conditions (EOC). The dynamic changes due to changing EOCs like variations in temperature, rotational speed, wind speed, etc. may be of the same order of magnitude as the variations due to damage making a reliable damage detection impossible. In this paper, we show a method for the compensation of changing EOC. The well-known null space based fault detection (NSFD) is used for damage detection. In the first stage, a training is performed using data from the undamaged structure under varying EOC. For the compensation of the EOC-e ects the undamaged state is modeled by different reference data corresponding to different representative EOC conditions. Finally, in the application, the influences of one or other EOC on each incoming data is weighted separately by means of a fuzzy-classiffcation algorithm. The theory and algorithm is successfully tested with data sets from a real wind turbine and with data from a laboratory model.

Sensor Performance Assessment Based on a Physical Model and Impedance Measurements

The employment of a large number of embedded sensors in advanced monitoring systems becomes more common, enabling in-service detection, localization and assessment of defects in mechanical, civil and aerospace structures. These sensors could be optical fibre sensors, accelerometers, strain gauges or piezoelectric wafer active sensors (PWAS). As the latter are quite popular, due to its multipurpose application as actuators and sensors and its low cost, this type will be investigated. Within this paper a possible approach of sensor performance is presented. The method uses the coupled electro-mechanical admittance to detect damage of the PWAS and its bonding layer. The help of a temperature dependent theoretical model provides for influences of changing environmental and operational conditions. The model will be compared with FEM-results, before showing the successful application on experimental results.

Path-Based MAPOD Using Numerical Simulations

Probability-based methods for the consideration of detection rates and associated damage sizes have been state of the art in NDT. For structural health monitoring (SHM) systems, the quantification of detection capabilities needs to be addressed to enable the industrial implementation. Due to the fixed mounting of SHM systems on structures, experimentally based investigation is particularly difficult and resource consuming. Therefore, the use of numerical simulations is suggested to generate additional data for probability studies. Within this paper, two methods of model-assisted probability of detection (MAPOD) are presented. For the use case of carbon fibre-reinforced plastics (CFRP) panels, tested with acousto-ultrasonics, a path-based analysis was chosen. After a short description of the underlying numerical models, the used probability-based methods are explained. Their application is shown in detail using a 3D and a 2D model for a CFRP panel.

Integration of Knowledge-Based Approach in SHM Systems

This paper discusses the potentials for integration of knowledge-based techniques in Structural Health Monitoring (SHM). Knowledge-based techniques and methods reinforce health assessment and influence on predictive maintenance of structures. A concept of the knowledge-based approach is developed. In particular, toolboxes for a simple numerical 3-degree of freedom (dof)-model and for force reconstruction at Canton Tower are implemented, respectively. The case studies deepen the insight into identifying needs in the field of SHM to employ knowledge-based approaches, especially in the reasoning process. The proposed concept lays the ground for future research in the field of SHM for utilizing knowledge-based methods in correlation with SHM algorithms and analysis of feedbacks obtained from sensors, engineering expertise and users former experience. The foresight is to broaden the scope for applying Knowledge Management (KM) techniques and methods towards developing a decision-making component for supporting SHM systems, and in turn fostering the detection, localization, classification, assessment and prediction.

Continuous wavelet transform application in diagnostics of piezoelectric wafer active sensors

Piezoelectric wafer active sensors (PWAS) are employed in a variety of structural health monitoring (SHM) applications. Failure of these might lead to significant problems, so monitoring of actuators themselves is necessary. While totally debonded PWAS can be detected easily, small debondings could still occur. In that case PWAS is still capable of generating ultrasound waves, but might lead to false diagnostic results since the underlying baseline measurements are not valid anymore. Therefore an experimental setup with a specimen of 16 partially debonded actuators has been used. Phenomena accompanying wave excitation by debonded actuators are examined. Collected knowledge is analyzed in order to identify existence, location and shape of a debonded part of the actuator. For a sufficiently debonded PWAS some interesting abnormalities have been detected for high frequencies. Wavelet analysis has revealed that the velocities of the motion and carrier frequencies depend on the shape of the debonded part of the PWAS.

Quality Assessment for the EMI-based Inspection of PWAS

This paper presents a new procedure to evaluate a method for piezoelectric wafer active sensor (PWAS) inspection, when it is combined with a given structural health monitoring (SHM) system for structural damage detection. A data-driven method is used, applying a damage indicator DIPWAS based on the electro-mechanical impedance spectrum (EMI), to distinguish between faulty and healthy PWAS. To evaluate the quality of this method, the consequences of PWAS damage are investigated. In this case, these are the changes within the generated wave field. We propose that the damage of the PWAS should be detected before it leads to false alarm by the SHM system. Checking this condition results in a novel type of quality assessment for the investigated EMI-based method to detect damaged PWAS. The methodology can also be transferred to other methods of detecting faulty PWAS. It is explained in detail followed by an application, showing excellent feasibility. doi: 10.12783/SHM2015/84

Damage Identification in Composite Panels - Methodology and Visualisation

A methodology for the identification of an impact damage using guided waves on a composite structure is implemented. Both numerical and experimental results are used, and a graphical user interface is developed to visualise the potentially damaged area. The latter allows, on top of detection, an assessment of the location and severity of the damage. The input can be experimentally based or calculated with the help of numerical models. Within this work, two numerical models are presented, based on stacked-shell finite element approach and on spectral element approach in time domain. The graphical interface allows the user to choose the most suitable approach from various damage identification methods using pitch-catch acousto-ultrasonics. The numerical models allow us to test a variety of damage locations with variable extents. The quality of the models is shown by a comparison of simulated and experimental data in time domain and respective damage indices. Finally, the visualisation allows to focus on specific areas, enhancing the analysis of multiple damages in a structure. The damage identification tool is a powerful tool in understanding the effects of various damage scenarios on the time response data and together with the numerical model provides a valuable input for model-assisted probability of detection (MAPOD).