Ferri M.H. Aliabadi

A Bayesian Approach for Sensor Optimisation in Impact Identification

This paper presents a Bayesian approach for optimizing the position of sensors aimed at impact identification in composite structures under operational conditions. The uncertainty in the sensor data has been represented by statistical distributions of the recorded signals. An optimisation strategy based on the genetic algorithm is proposed to find the best sensor combination aimed at locating impacts on composite structures. A Bayesian-based objective function is adopted in the optimisation procedure as an indicator of the performance of meta-models developed for different sensor combinations to locate various impact events. To represent a real structure under operational load and to increase the reliability of the Structural Health Monitoring (SHM) system, the probability of malfunctioning sensors is included in the optimisation. The reliability and the robustness of the procedure is tested with experimental and numerical examples. Finally, the proposed optimisation algorithm is applied to a composite stiffened panel for both the uniform and non-uniform probability of impact occurrence.

Damage Detectability Model of Pitch-Catch Configuration in Composite Plates

Detectability of damage using Lamb waves depends on many factors such as size and severity of damage, attenuation of the wave and distance to the transducers. This paper presents a detectability model for pitch-catch sensors configuration for structural health monitoring (SHM) applications. The proposed model considers the physical properties of lamb wave propagation and is independent of damage detection algorithm, which provides a generic solution for probability of detection. The applicability of the model in different environmental and operational conditions is also discussed.

Quasistatic Electro-Elastic Contact Modeling Using the Boundary Element Method

A three-dimensional boundary element methodology to study frictionless indentation response of piezoelectric (PE) materials is presented. The boundary element method (BEM) is used in order to compute the electro-elastic influence coeffcients of fully anisotropic piezoelectric solids. The proposed contact formulation is based on the augmented Lagrangian method presented in [33, 34, 35] and makes it possible to consider piezoelectric materials under different mechanical and electrical boundary conditions (i.e. insulating indenter and conducting indenter). The methodology is validated by comparison with theoretical solutions presented in the literature.

Effectiveness of RAPID and SSM Algorithms on Composite Scarf Repair

This work focuses on diagnostic methodologies for composite repair patch based on structural health monitoring (SHM) technology. Methodologies based on ultrasonic guided waves (GW) are developed and assessed for monitoring composite scarf repair with piezoelectric transducers. The effectiveness of the RAPID (reconstruction algorithm for probabilistic inspection of defects) algorithm was investigated for adhesively bonded composite patch repair. A composite scarf repair has been weakened by 4-point bending fatigue test and impacted after to generate a Barely Visible Damage (BVID). Both conventional RAPID technique, which requires baseline signals, and the Scaling Subtraction Method (SSM) were applied to detect damage in the bondline. The conventional method showed good performance for defect detection and localization whereas the SSM gives encouraging results for non-linear baseline-free RAPID.

Robust and Reliability-Based Design Optimization of a Composite Floor Beam

This study investigates the advantages and disadvantages of using probabilistic optimization methods in aircraft structural design. The necessity to achieve a design insensitive to systems variations (robust) and less likely to fail (reliable) is addressed, in order to reduce costs and risk of accidents. Because of the complex nature of composite structures, the need of surrogate models to predict structural responses arises. In order to build a meta-model, Design of Experiments (DOE) methods are used to determine the location of sampling points in the design space. Monte Carlo Simulations (MCS), creating random samples, are used to propagate uncertainties from the surrogate model inputs to variations in model outputs. Different optimization algorithms and surrogate models are compared, in order to speed up the optimization process and reduce modelling errors. The deterministic design resulted in a design which is neither robust nor reliable. Stochastic approaches accounting for uncertainties, on the other hand, resulted in enhanced robustness (Robust Design), enhanced reliability (Reliable Design) or a combination of both (Robust and Reliable Design).

Meshfree Continuum Damage Model for Twill Composites

This paper presents continuum damage mechanics analysis of twill woven composites. Element Free Galerkin formulation is utilized and enriched with mathematical representation of twill composite in a way that includes details of the wrap/weft/matrix without the requirement of a detailed mesh of individual components. Continuum damage mechanics formulation is developed within the meshfree context and applied to the twill composite.

A Boundary Element Model for Structural Health Monitoring Based on the S0 Lamb Wave Mode

The aim of this paper was to carry out numerical simulations of structural health monitoring applications for plate structures using the boundary element method (BEM). The fundamental symmetric Lamb mode (S0) is chosen for the SHM applications. The propagation, reflection and diffraction of the S0 mode Lamb wave are modelled using a boundary element formulation based on the plane stress theory. Piezoelectric (PZT) actuators are mounted on plate surfaces to excite the S0 mode wave. A semi-analytical method is adopted to couple the PZT actuators and the host plate. Numerical results show that BEM is a very efficient simulation method for the structural health monitoring of plates.

Statistical Inference of the Equivalent Initial Flaw Size Distribution Using the Boundary Element Method under Multiple Sources of Uncertainty

In this work, a method for determining the Equivalent Initial Flaw Size (EIFS) distribution using the Boundary Element Method (BEM) is proposed. Maximum Likelihood Estimation (MLE) is used to infer the EIFS distribution of a cracked stiffened panel under multiple sources of uncertainty, including uncertainty in the loading conditions, fatigue crack growth model parameters, and in the measurement of crack size found from routine inspections. Results suggest that MLE is an effective tool for estimating the parameters of an EIFS distribution when no prior knowledge is available regarding the EIFS distribution or its parameters.

An Effective Impact Detection Method for Composite Curved Panel

ANNs are only accurate for the scope of the given training data which is not suitable for real life impact localisation due to the large range of possible impact variation. Impact data was collected for a variation of impact cases (angle, mass and energy) on a sensorized curved composite panel. From observation of the obtained data, a robust signal Time of Arrival (TOA) extraction method is proposed using a Normalised Smooth Envelope Threshold (NSET) which is a modification of the currently known Normalised Threshold (NT) method. Two ANNs were trained using TOA extracted with the NT and NSET method from a baseline case and tested with TOA extracted from cases having added variation of impact condition. The results show that the proposed NSET method results in more accurate results for impact cases different to the training case and thus allows for only a single impact training case to accurately predict cases with multiple variation. This enhances the applicability of ANNs for impact localisation in real life conditions.

Structural Health Monitoring of Scarfed Repaired Composite Panels Using Inject-Printed Patterns

A novel strategy is proposed for monitoring of bonded composite repair patch is developed and tested. A specially designed pattern was inject-printed onto the step-sanded repair surface providing real time information about the bondline structural integrity. The obtained results indicated that the proposed methodology was able to detect damage induced by an impact event. Infrared thermography (IrT) and ultrasound inspection were also employed in order to validate the proposed methodology.