Structural Health Monitoring Impact Classification Method Based on Bayesian Neural Network.

This paper proposes a novel method for multi-class classification and uncertainty quantification of impact events on a flat composite plate with a structural health monitoring (SHM) system by using a Bayesian neural network (BNN). Most of the existing research in passive sensing has focused on deterministic approaches for impact detection and characterization. However, there are variability in impact location, angle and energy in real operational conditions which results in uncertainty in the diagnosis. Therefore, this paper proposes a reliability-based impact characterization method based on BNN for the first time. Impact data are acquired by a passive sensing system of piezoelectric (PZT) sensors. Features extracted from the sensor signals, such as their transferred energy, frequency at maximum amplitude and time interval of the largest peak, are used to develop a BNN for impact classification (i.e., energy level). To test the robustness and reliability of the proposed model to impact variability, it is trained with perpendicular impacts and tested by variable angle impacts. The same dataset is further applied in a method called multi-artificial neural network (multi-ANN) to compare its ability in uncertainty quantification and its computational efficiency against the BNN for validation of the developed meta-model. It is demonstrated that both the BNN and multi-ANN can measure the uncertainty and confidence of the diagnosis from the prediction results. Both have very high performance in classifying impact energies when the networks are trained and tested with perpendicular impacts of different energy and location, with 94% and 98% reliable predictions for BNN and multi-ANN, respectively. However, both metamodels struggled to detect new impact scenarios (angled impacts) when the data set was not used in the development stage and only used for testing. Including additional features improved the performance of the networks in regularization; however, not to the acceptable accuracy. The BNN significantly outperforms the multi-ANN in computational time and resources. For perpendicular impacts, both methods can reach a reliable accuracy, while for angled impacts, the accuracy decreases but the uncertainty provides additional information that can be further used to improve the classification.

Influence of Composite Thickness on Ultrasonic Guided Wave Propagation for Damage Detection.

In this paper, the propagation properties of ultrasonic guided waves (UGWs) in different-thickness composites (i.e., 2, 4 and 9 mm) were critically assessed, and their effectiveness for damage detections and localisations under varying temperatures was demonstrated. A diagnostic film with phased-array lead zirconate titanate (PZT) transducers based on the ink-jet printing technique was used in the experiments. Initially, the dispersion curves for these composites were compared. Next, the effects of the composite thickness on the A0 and S0 mode amplitudes and the group velocity were investigated by active sensing. Next, the behaviours of UGWs under varying temperatures in different-thickness plates were also investigated. Finally, surface-mounted artificial damage and impact damage were detected and located in different composites.

Experiments and modelling of ultrasonic waves in composite plates under varying temperature.

Guided wave (GW) structural health monitoring (SHM) systems offer an attractive solution as an in-situ quasi real-time assessment of structural damage, but their sensitivity and efficiency may be impaired under varied environmental and operational conditions. Thus, virtual tests, such as that based on the Finite Element (FE) method, represent a valid approach for simulating and investigating SHM systems, enabling a substantial reduction in experimental campaigns. In this work, GW propagation characteristics in a carbon fibre-reinforced composite plate are studied under a varying temperature condition, representative of the aeronautics application. At first, GW SHM system was physically tested at room temperature (20°C), and the results were used to calibrate and assess the proposed FE modelling approaches, characterised by different element types and mesh sizes. A temperature independent averaged time compensation factor is proposed to mitigate the numerical data dependency on excitation frequency and propagation angle. Two temperature variations (from 20°C to -50°C, and 20°C to 65°C) were experimentally and numerically considered to investigate the effect of varying temperature on the GW. For all test cases, the compensated numerical data was compared to the experimental results, and discussed in terms of dispersion curves, focusing on the zero-order symmetric, S0, and antisymmetric, A0, modes. Results show that both 2D and 3D FE approaches can accurately predict the changes in GW due to varying temperature, with the group velocity of the A0 mode being less sensitive to temperature variations.

Deep Learning Approach to Impact Classification in Sensorized Panels Using Self-Attention.

This paper proposes a new method of impact classification for a Structural Health Monitoring system through the use of Self-Attention, the central building block of the Transformer neural network. As a topical and highly promising neural network architecture, the Transformer has the potential to greatly improve the speed and robustness of impact detection. This paper investigates the suitability of this new network, confronting the advantages and disadvantages offered by the Transformer and a well-known and established neural network for impact detection, the Convolutional Neural Network (CNN). The comparison is undertaken on performance, scalability, and computational time. The inputs to the networks were created using a data transformation technique, which transforms the raw time series data collected from the network of piezoelectric sensors, installed on a composite panel, through the use of Fourier Transform. It is demonstrated that the Transformer method reduces the computational complexity of the impact detection significantly, while achieving excellent prediction results.

Structural Health Monitoring Cost Estimation of a Piezosensorized Aircraft Fuselage.

Guided waves-based SHM systems are of interest in the aeronautic sector due to their lightweight, long interrogation distances, and low power consumption. In this study, a bottom-up framework for the estimation of the initial investment cost (COTC) and the added weight (WAW) associated with the integration of a SHM system to an aircraft is presented. The framework provides a detailed breakdown of the activities and their costs for the sensorization of a structure using a fully wired approach or the adoption of the printed diagnostic film. Additionally, the framework considers the difference between configuring the system for Manual or Remote data acquisition. Based on the case study presented on the sensorization of a regional aircraft composite fuselage, there is a trade-off between COTC and WAW for the SHM options considered. The Wired-Manual case leads to the lowest COTC with the highest WAW, while the combination of diagnostic film with a Remote system leads to the highest COTC and the lowest WAW. These estimations capture the characteristics of each system and can be integrated into cost-benefit analyses for the final selection of a particular configuration.

Structural Integrity Assessment of Composites Plates with Embedded PZT Transducers for Structural Health Monitoring.

Active sensing using ultrasonic guided waves (UGW) is widely investigated for monitoring possible damages in composite structures. Recently, a novel diagnosed film based on a circuit-printed technique with piezoelectric lead zirconate titanate (PZT) transducers has been developed. The diagnostic film is a replacement for the traditional cable connection to PZT sensors and has been shown to significantly reduce the weight of the host structure. In this work, the diagnosed films were embedded into composite structures during manufacturing using a novel edge cut-out method during lay-up, which allowed for edge trimming after curing. In this paper, the effect of fatigue loading on the integrity of PZT transducers is initially investigated. The electro-mechanical impedance (EMI) properties at different fatigue loading cycles were used as the diagnostic measure for the performance of the sensors. At the same time, the behaviours of UGW were investigated at different fatigue loading cycles. It was found that the EMI properties and active sensing behaviours remained stable up to 1 million cycles for the force ranges of 0.5~5 kN and 1~10 kN. Next, the effect of embedding the diagnosed film on the mechanical properties of the host composite structure was investigated. Tensile and compressive tests were conducted and the elastic modulus of composite coupons with and without embedded PZT diagnosed films were compared. The elastic modulus of composite coupons with PZT diagnosed films embedded across the entire coupon reduced by as much as 20% for tensile tests and just over 10% for compressive tests compared to the coupons without embedded sensors. These reductions are considered the worst-case scenario, as in real structures the film would only be embedded in a relatively small area of the structure.

Correction: Seno, A.H.; Ferri Aliabadi, M.H. Impact Localisation in Composite Plates of Different Stiffness Impactors under Simulated Environmental and Operational Conditions. Sensors 2019, 19, 3659.

The author wishes to make the following correction to this paper [...].

A Multilevel Isolation Forrest and Convolutional Neural Network Algorithm for Impact Characterization on Composite Structures.

In this paper, a Deep Learning approach is proposed to classify impact data based on the type of impact (Hard or Soft Impacts), via obtaining voltage signals from Piezo-Electric sensors, mounted on a composite panel. The data is processed further to be classified based on their energy, location and material. Minimalistic and Automated feature extraction and selection is achieved via a deep learning algorithm. Convolutional Neural Networks (CNN) are employed to extract and select important features from the voltage data. Once features are selected the impacts, are classified based on either, Hard Impacts (simulated from steel impactors in a lab setting), Soft Impacts (simulated from silicon impactors in a lab setting) and their corresponding location and energy levels. Furthermore, in order to use the right data for training they are obtained from the signals as anomalies via Isolation Forests (IF) to speed up the process. Using this approach Hard and Soft Impacts, their corresponding locations and respective energies are identified with high accuracy.

Active Health Monitoring of Thick Composite Structures by Embedded and Surface-Mounted Piezo Diagnostic Layer.

An effective approach for an embedded piezo diagnostic layer into thick composite material is presented. The effectiveness of the approach is assessed in comparison to the surface-mounted layer. The proposed manufacturing alleviates difficulties associated with trimming edges of composites when embedding wires. The Electro-Mechanical Impedance technique is used to access the integrity of the piezoelectric sensors bonding process. Comparisons of ultrasonic guided waves are made between embedded and surface-mounted diagnostic layers and their penetration through and across the thickness of the composites. Temperature influences with the range from -40 °C up to 80 °C on embedded and surface-mounted guided waves are investigated. An investigation is carried out into the relationship between amplitude and time-of-flight with temperature at different excitation frequencies. The temperature has significant but different effects on amplitude and phase-shift of guided waves for the embedded layer compared to the surface-mounted layer. A Laser Doppler Vibrometer is used to identify the blue tack and impact damage. Both embedded and surface-mounted layers are shown to be an effective means of generating detectable wave scatter from damage.

Impact Localisation in Composite Plates of Different Stiffness Impactors under Simulated Environmental and Operational Conditions.

A parametric investigation of the effect of impactor stiffness as well as environmental and operational conditions on impact contact behaviour and the subsequently generated lamb waves in composite structures is presented. It is shown that differing impactor stiffness generates the most significant changes in contact area and lamb wave characteristics (waveform, frequency, and amplitude). A novel impact localisation method was developed based on the above observations that allows for variations due to differences in impactor stiffness based on modifications of the reference database method and the Akaike Information Criterion (AIC) time of arrival (ToA) picker. The proposed method was compared against a benchmark method based on artificial neural networks (ANNS) and the normalised smoothed envelope threshold (NSET) ToA extraction method. The results indicate that the proposed method had comparable accuracy to the benchmark method for hard impacts under various environmental and operational conditions when trained only using a single hard impact case. However, when tested with soft impacts, the benchmark method had very low accuracy, whilst the proposed method was able to maintain its accuracy at an acceptable level. Thus, the proposed method is capable of detecting the location of impacts of varying stiffness under various environmental and operational conditions using data from only a single impact case, which brings it closer to the application of data driven impact detection systems in real life structures.

An Innovative Diagnostic Film for Structural Health Monitoring of Metallic and Composite Structures.

A novel lightweight diagnostic film with sensors/actuators and a multiple-path wiring option using inkjet printing was developed. The diagnostic film allows for systematic, accurate, and repeatable sensor placement. Furthermore, the film is highly flexible and adaptable for placement on complex configurations. The film can be attached to the surface of the structure through a uniform secondary boundary procedure or embedded within the composite layup during curing. The surface-mounted film can simply be peeled off for repair or replacement without scratching or damaging the part. The film offers significant weight reduction compared to other available technologies. A set of extreme temperature, altitude, and vibration environment test profiles were carried out following the Radio Technical Commission for Aeronautics (RTCA) DO-160 document to assess the durability and performance of the diagnostic film for onboard application. The diagnostic film was shown to be durable and reliable in withstanding the variable operational and harsh environmental conditions of tests representing the conditions of regional aircraft.

Impact Damage Localisation with Piezoelectric Sensors under Operational and Environmental Conditions.

Guided-wave structural health monitoring (SHM) systems with piezoelectric sensors are investigated for localisation of barely visible impact damage in CFRP plates under vibration and different thermal conditions. A single baseline set is used in a delay-and-sum algorithm with temperature correction for damage localisation in a large temperature range. Damage localisation is also demonstrated under transient thermal conditions, with signals recorded while the temperature is rapidly decreased. Damage severity due to successive impact events is studied under constant temperature. Damage is also localised when the plate is subjected to random vibration.

A Multi-Level Decision Fusion Strategy for Condition Based Maintenance of Composite Structures.

In this work, a multi-level decision fusion strategy is proposed which weighs the Value of Information (VoI) against the intended functions of a Structural Health Monitoring (SHM) system. This paper presents a multi-level approach for three different maintenance strategies in which the performance of the SHM systems is evaluated against its intended functions. Level 1 diagnosis results in damage existence with minimum sensors covering a large area by finding the maximum energy difference for the guided waves propagating in pristine structure and the post-impact state; Level 2 diagnosis provides damage detection and approximate localization using an approach based on Electro-Mechanical Impedance (EMI) measures, while Level 3 characterizes damage (exact location and size) in addition to its detection by utilising a Weighted Energy Arrival Method (WEAM). The proposed multi-level strategy is verified and validated experimentally by detection of Barely Visible Impact Damage (BVID) on a curved composite fuselage panel.

Pharmacokinetic and pharmacodynamic modelling of marbofloxacin administered alone and in combination with tolfenamic acid in calves.

In a four-period, cross-over study, the fluoroquinolone antibacterial drug marbofloxacin (MB) was administered to calves, alone and in combination with the nonsteroidal anti-inflammatory drug tolfenamic acid (TA). Both drugs were administered intramuscularly (IM) at doses of 2 mg/kg. A tissue cage model of inflammation, based on the actions of the mild irritant carrageenan, was used to evaluate the pharmacokinetics (PK) of MB and MB in combination with TA. MB mean values of area under concentration-time curve (AUC) were 15.1 µg·h/mL for serum, 12.1 µg·h/mL for inflamed tissue cage fluid (exudate) and 9.6 µg·h/mL for noninflamed tissue cage fluid (transudate). Values of C(max) were 1.84, 0.35 and 0.31 µg/mL, respectively, for serum, exudate and transudate. Mean residence time (MRT) of 23.6 h (exudate) and 22.6 h (transudate) also differed significantly from serum MRT (8.6 h). Co-administration of TA did not affect the PK profile of MB. The pharmacodynamics of MB was investigated using a bovine strain of Mannheimia haemolytica. Time-kill curves were established ex vivo on serum, exudate and transudate samples. Modelling the ex vivo serum time-kill data to the sigmoid E(max) equation provided AUC(24 h) /MIC values required for bacteriostatic (18.3 h) and bactericidal actions (92 h) of MB and for virtual eradication of the organism was 139 h. Corresponding values for MB + TA were 20.1, 69 and 106 h. These data were used to predict once daily dosage schedules for a bactericidal action, assuming a MIC(90) value of 0.24 µg/mL, a dose of 2.6 mg/kg for MB and 2.19 mg/kg for MB + TA were determined, which are similar to the currently recommended dose of 2.0 mg/kg.

Determination of a dosage regimen of colistin by pharmacokinetic/pharmacodynamic integration and modeling for treatment of G.I.T. disease in pigs.

Colistin is an antimicrobial drug of the polymyxin group and COLIVET SOLUTION is an aqueous solution containing colistin sulphate (2 x 10(6) IU/mL), formulated for oral administration. The target species is the pig, particularly the suckling and post weaning animal. This investigation was undertaken to provide pharmacokinetic and pharmacodynamic data on which to base the selection of dosage rate and interval of the solution for the treatment of porcine colibacillosis. Colistin absorption from the gastrointestinal tract of young pigs, when administered at dosage rates of 25,000, 50,000 and 100,000 IU/kg, was slight or absent. The drug was therefore restricted almost entirely to the required site of action. The colistin concentration-time profile within the jejunum and ileum was established, and this enabled determination of the pharmacokinetic variables, maximum concentration (C(max)) and area under curve (AUC) and derivation of the surrogate indices of antibacterial activity, C(max)/minimum inhibitory concentration (MIC) and AUC/MIC through integration of in vivo data with the results of in vitro potency studies for four strains of Escherichia coli. In the in vitro bacterial growth inhibition studies colistin acted by a concentration-dependent killing mechanism. Numerical values for the surrogate parameter AUC/MIC producing bactericidal and eradication effects of colistin against four strains of E. coli were established by PK-PD modeling based on the sigmoidal E(max) equation. These data were used to predict a daily dosage regimen for colistin.

Movement of a spherical cell in capillaries using a boundary element method.

This study aims to investigate the translation and rotation of a spherical particle in capillaries and overcomes limitations in previous studies by using a boundary element method. The capillary, a straight cylindrical tube, is filled with a Newtonian viscous fluid. A spherical particle is arbitrarily positioned in the capillary either co-centrically or eccentrically and is free to translate and rotate. Flow in the capillary is first assumed to be caused solely by the movement of the sphere under the gravity. When a steady state is reached, the net force and torque on the sphere are zero. The translating velocity and rotation of the particle are calculated from equilibrium equations. For a co-centric sphere, our result agrees to Bohlin's analytical solution (Bohlin, 1960) and the difference is less than 1%. For an eccentrically positioned sphere in the tube, there are no analytical solutions unless the eccentricity is infinitesimal. Results by boundary element method (BEM) give an improved estimations on the velocity and rotation of the sphere than earlier results by a boundary singularity method (BSM), particularly when the clearance between the tube and the sphere becomes small. Movement of a spherical particle in a capillary driven by a pressure gradient is further investigated, which has closer relevance to movement of blood cells in capillaries. The current study using BEM enables investigation on cell movement in close proximities of the capillary wall.

Total hip arthroplasty wear simulation using the boundary element method.

In this paper an application of the boundary element method for simulating wear in total hip prosthesis is presented. Several examples including different update periods of the worn acetabular cup, various femoral head sizes and various materials for both the femoral head and the acetabular cup are simulated under the same variable loading conditions for up to 20 years of service. Moreover, two different femoral models are considered in order to investigate the influence of the femoral modelling. The analysis demonstrates that due to the boundary only modelling requirement, the computational time and storage remains low, allowing large service periods to be simulated. Generally, the results obtained are in good agreement with other researchers findings. Moreover, ignoring the bending of the femoral neck in the model, results in a small overestimation of the maximum wear depth, while the volumetric wear is slightly underestimated. However, these differences are trivial considering the reduction of the computational effort.

Bone remodelling using sensitivity analysis.

A boundary element formulation is presented for analysing the surface bone remodelling. The formulation is based on sensitivity analysis and utilises design parameters which are related to the shape of the bone. An application of the method to the modelling of bone ingrowth into a slot of an implant is presented.