Clyde K. Coelho

Detection of Fatigue Cracks and Torque Loss in Bolted Joints

Fatigue crack growth during the service life of aging aircraft is a critical issue and monitoring of such cracks in structural hotspots is the goal of this research. This paper presents a procedure for classification and detection of cracks generated in bolted joints which are used at numerous locations in aircraft structures. Single lap bolted joints were equipped with surface mounted piezoelectric (pzt) sensors and actuators and were subjected to cyclic loading. Crack length measurements and sensor data were collected at different number of cycles and with different torque levels. A classification algorithm based on Support Vector Machines (SVMs) was used to compare signals from a healthy and damaged joint to classify fatigue damage at the bolts. The algorithm was also used to classify the amount of torque in the bolt of interest and determine if the level of torque affected the quantification and localization of the crack emanating from the bolt hole. The results show that it is easier to detect the completely loose bolt but certain changes in torque, combined with damage, can produce some non-unique classifier solutions.

A strain amplitude-based algorithm for impact localization on composite laminates

Automated detection of damage due to low energy impacts in composite structures is very important for aerospace structural health monitoring applications. Low-velocity impact creates subsurface damage that can significantly reduce the stiffness of a component, yet show barely visible damage. This article proposes a novel methodology for impact localization based on the maximum strain amplitude measured by fiber Bragg grating (FBG) sensors during an impact event. The approach correlates the strain amplitude of each sensor pair to find the location of highest strain corresponding to the impact location. This approach requires minimal knowledge of the structure and fewer number of sensors as opposed to current localization methods. Both simulation and experimental data are used as proof of concept. Since FBG sensors measure strain in only one direction, the effect of sensor orientation on the performance of the algorithm is also studied. The algorithm is tested on graphite/epoxy composite plates and shows good localization results in all impact cases considered.

Impact localization on complex structures using FBG strain amplitude information

Localization of low energy impacts on carbon fiber composites is an important aspect of structural health monitoring since it creates subsurface damage which can significantly reduce the stiffness of a component. A novel impact localization method is proposed based on the strain amplitude measured by Fiber Bragg Grating (FBG) sensors. The algorithm is based on the relative placement of all sensors and the maximum strain amplitude measured by each sensor. This method requires minimal knowledge of the material or the structure and a minimum number of sensors. The algorithm showed good results on both simulated and experimental test cases of woven composite plates. It was found that a minimum of five FBG are necessary to accurately predict the impact location on a plate. The algorithm was also tested on a woven composite wing showing good localization along the span of the wing but higher errors along the chord length due to the nonlinearity in the measured strains.

A hierarchical classification scheme for computationally efficient damage classification

Abstract This article presents a methodology for data mining of sensor signals in a structural health monitoring (SHM) framework for damage classification using a machine-learning-based approach called support vector machines (SVMs). A hierarchical decision tree structure is constructed for damage classification and experiments were conducted on metallic and composite test specimens with surface mounted piezoelectric transducers. Damage was induced in the specimens by fatigue, impact, and tensile loading; in addition, specimens with seeded delaminations were also considered. Data were collected from the surface mounted sensors at different severities of induced damage. A matching pursuit decomposition (MPD) algorithm was used as a feature extraction technique to preprocess the sensor data and extract the input vectors used in classification. Using this binary tree framework, the computational intensity of each successive classifier is reduced and the efficiency of the algorithm as a whole is increased. The results obtained using this classification show that this type of architecture works well for large data sets because a reduced number of comparisons are required. Due to the hierarchical set-up of the classifiers, performance of the classifier as a whole is heavily dependent on the performance of the classifier at higher levels in the classification tree.

Optimal sensor placement for active guided wave interrogation of complex metallic components

With research in structural health monitoring (SHM) moving towards increasingly complex structures for damage interrogation, the placement of sensors is becoming a key issue in the performance of the damage detection methodologies. For ultrasonic wave based approaches, this is especially important because of the sensitivity of the travelling Lamb waves to material properties, geometry and boundary conditions that may obscure the presence of damage if they are not taken into account during sensor placement. The framework proposed in this paper defines a sensing region for a pair of piezoelectric transducers in a pitch-catch damage detection approach by taking into account the material attenuation and probability of false alarm. Using information about the region interrogated by a sensoractuator pair, a simulated annealing optimization framework was implemented in order to place sensors on complex metallic geometries such that a selected minimum damage type and size could be detected with an acceptable probability of false alarm anywhere on the structure. This approach was demonstrated on a lug joint to detect a crack and on a large Naval SHM test bed and resulted in a placement of sensors that was able to interrogate all parts of the structure using the minimum number of transducers.

Active Damage Localization in Anisotropic Materials Using Guided Waves

It is important to be able to accurately assess the health state of aerospace vehicles through the detection, location, and quantification of damage. Locating damage is especially difficult in anisotropic materials, as the guided wave velocity is a function of material orientation. The objective of this paper is damage detection and localization in composite materials. A methodology and framework is developed in which a time map is constructed for each actuator-sensor pair which establishes times of flight for each location on the sample. Differences in time between healthy and damaged sensor signals are then extracted and used to create a map of possible damage locations. These resulting solution maps are merged yielding a final damage position. Equations governing the behavior of the system are developed, data extraction is carried out, and several sensor schemes are evaluated. The framework is validated, and impact positions are calculated for two actuation frequencies. For the damage state, the previous state is taken as a baseline for damage time extraction from the sensor signals. The damage position is calculated within 9% when using both 50 kHz and 200 kHz actuation frequencies.

Impact localization and force estimation on a composite wing using fiber bragg gratings sensors

Automated detection of damage due to low energy impacts in composite structures is very important for aerospace structural health monitoring (SHM) applications since it creates subsurface damage which can significantly reduce the stiffness of a component. Fiber Bragg grating (FBG) sensors are showing promise in many applications since they are low weight, require minimal space, are immune to electromagnetic interference, and can support multiple sensors in a single fiber. This paper proposes a novel methodology for impact localization and prediction of loading as a function of time. The localization method proposed is based on the maximum strain amplitude measured by FBG sensors during impact and the relative placement of all sensors. The force time history of impact loading on composite wing is predicted using a support vector regression (SVR) approach. A time delay embedding feature extraction scheme is used since it can resolve the dynamics of the impact using the sensor signal from the FBGs. Experimental validation is presented on graphite/epoxy composite wings with accurate localization and estimation results along the span of the wing.

Robust feature extraction for rapid classification of damage in composites

The ability to detect anomalies in signals from sensors is imperative for structural health monitoring (SHM) applications. Many of the candidate algorithms for these applications either require a lot of training examples or are very computationally inefficient for large sample sizes. The damage detection framework presented in this paper uses a combination of Linear Discriminant Analysis (LDA) along with Support Vector Machines (SVM) to obtain a computationally efficient classification scheme for rapid damage state determination. LDA was used for feature extraction of damage signals from piezoelectric sensors on a composite plate and these features were used to train the SVM algorithm in parts, reducing the computational intensity associated with the quadratic optimization problem that needs to be solved during training. SVM classifiers were organized into a binary tree structure to speed up classification, which also reduces the total training time required. This framework was validated on composite plates that were impacted at various locations. The results show that the algorithm was able to correctly predict the different impact damage cases in composite laminates using less than 21 percent of the total available training data after data reduction.

Machine learning approach to impact load estimation using fiber Bragg grating sensors

Automated detection of damage due to impact in composite structures is very important for aerospace structural health monitoring (SHM) applications. Fiber Bragg grating (FBG) sensors show promise in aerospace applications since they are immune to electromagnetic interference and can support multiple sensors in a single fiber. However, since they only measure strain along the length of the fiber, a prediction scheme that can estimate loading using randomly oriented sensors is key to damage state awareness. This paper focuses on the prediction of impact loading in composite structures as a function of time using a support vector regression (SVR) approach. A time delay embedding feature extraction scheme is used since it can characterize the dynamics of the impact using the sensor signal from the FBGs. The efficiency of this approach has been demonstrated on simulated composite plates and wing structures. Training with impacts at four locations with three different energies, the constructed framework is able to predict the force-time history at an unknown impact location to within 12 percent on the composite plate and to within 10 percent on a composite wing when the impact was within the sensor network region.

Damage characterization of composite wing subjected to impact loading - An experimental study

Damage on woven composites is a phenomenon that is difficult to characterize due to complex weave geometry. A woven composite wing structure adds to the complexity of characterizing damage through Fiber Bragg Grating (FBG) sensors. The present paper studies the FBG response and damage characterization of foam core and hollow composite wings. Plain and twill weave wings were manufactured and subjected to low energy (52.5J) and high energy (150J) impacts. Damage was assessed using FBG sensors, flash thermography, and visual inspection of the wings. Two FBG sensors were placed along the chord length and the spanwise direction at equal distances from the impact site to measure the axial strain as a function of time. The main failure modes of foam core wings were fiber breakage and foam crushing for high energy impacts, while core crushing and delamination between the core and the composite wing was found for low energy impacts. The hollow wings had a significant reduction in stiffness, resulting in a ripple effect where the wing would go into tension, then compression. This phenomenon varied depending on the location of the sensors on the wing. Although the impact zone was near the middle of the chord length of the wing, the resulting stress has caused large damage at the leading edge and significant debonding at the trailing edge of the hollow wing. An FE model was created to validate the experimental results and showed good correlation between the high stress areas in the model, the FBG response, and the damage sites on the wing.Copyright