Search results

To develop a new method for estimation of damage configuration in composite laminate structure using acoustic wave propagation signal and a reduction‐prediction neural network to deal with high dimensional spectral data.

A reduction‐prediction network, which is a combination of an independent component analysis (ICA) and a multi‐layer perceptron (MLP) neural network, is proposed to quantify the damage state related to transverse matrix cracking in composite laminates using acoustic wave propagation model. Given the Fourier spectral response of the damaged structure under frequency band‐selective excitation, the problem is posed as a parameter estimation problem. The parameters are the stiffness degradation factors, location and approximate size of the stiffness‐degraded zone. A micro‐mechanics model based on damage evolution criteria is incorporated in a spectral finite element model (SFEM) for beam type structure to study the effect of transverse matrix crack density on the acoustic wave response. Spectral data generated by using this model is used in training and testing the network. The ICA network called as the reduction network, reduces the dimensionality of the broad‐band spectral data for training and testing and sends its output as input to the MLP network. The MLP network, in turn, predicts the damage parameters.

Numerical demonstration shows that the developed network can efficiently handle high dimensional spectral data and estimate the damage state, damage location and size accurately.

Only numerical validation based on a damage model is reported in absence of experimental data. Uncertainties during actual online health monitoring may produce errors in the network output. Fault‐tolerance issues are not attempted. The method needs to be tested using measured spectral data using multiple sensors and wide variety of damages.

The developed network and estimation methodology can be employed in practical structural monitoring system, such as for monitoring critical composite structure components in aircrafts, spacecrafts and marine vehicles.

A new method is reported in the paper, which employs the previous works of the authors on SFEM and neural network. The paper addresses the important problem of high data dimensionality, which is of significant importance from practical engineering application viewpoint.

The purpose of this research is to establish a robust numerical framework for the calibration of macroscopic constitutive parameters, based on the analysis of polycrystalline RVEs with computational homogenisation.

This framework is composed of four building-blocks: (1) the multi-scale model, consisting of polycrystalline RVEs, where the grains are modelled with anisotropic crystal plasticity, and computational homogenisation to link the scales, (2) a set of loading cases to generate the reference responses, (3) the von Mises elasto-plastic model to be calibrated, and (4) the optimisation algorithms to solve the inverse identification problem. Several optimisation algorithms are assessed through a reference identification problem. Thereafter, different calibration strategies are tested. The accuracy of the calibrated models is evaluated by comparing their results against an FE2 model and experimental data.

In the initial tests, the LIPO optimiser performs the best. Good results accuracy is obtained with the calibrated constitutive models. The computing time needed by the FE2 simulations is 5 orders of magnitude larger, compared to the standard macroscopic simulations, demonstrating how this framework is suitable to obtain efficient micro-mechanics-informed constitutive models.

This contribution proposes a numerical framework, based on FE2 and macro-scale single element simulations, where the calibration of constitutive laws is informed by multi-scale analysis. The most efficient combination of optimisation algorithm and definition of the objective function is studied, and the robustness of the proposed approach is demonstrated by validation with both numerical and experimental data.

Resin jetting with piezo print‐heads is in increasing use, and in the rapid prototyping industry, the merging quality between adjacent droplets will determine the mechanical properties and reliability of the products. Therefore, it is essential to find an experimental technique to ensure seamless inter‐droplet merging. Speckle interferometry with electron microscopy (SIEM) is a micro‐mechanics measurement technique that has a spatial resolution approaching a few nanometers. In this paper, SIEM is successfully applied to measure the ultimate tensile stress and tensile modulus of jetted, UV‐cured cationic resin microsamples. Results show that the microsamples exhibit similar properties to the bulk material properties and that jetting two layers on top of each other is not detrimental to the material properties.

The purpose of this paper is to present knowledge in estimating yield surfaces of heterogeneous media by use of homogenization, especially where the macroscopic behaviour is driven by weak interfaces between phase constituents.

A computational homogenization procedure is used to determine the yield surface of a Representative Volume Element (RVE) that contains a fully debonded inclusion embedded within ideally plastic matrix, whereby the interface is modelled by a Coulomb type friction law.

The macroscopic behaviour of the RVE is shown to coincide an RVE with a hole for expanding loads, whereas for compressive loads, it was shown to approach an RVE with a fully bonded inclusion.

The present paper builds on Gurson’s work in estimating macroscopic yield surfaces of heterogeneous materials. The work is novel in the sense that there had been no previous publications discussing influence of weak interfaces on yield surfaces.

This paper proposes a simplified two‐dimensional representation of the unit cell of the fabric that involves three bodies in contact.

The fabrics are not simple homogenous structures. They have a discrete structural character and this is essential for their complex mechanical behaviour. Low stress micro‐mechanics is mainly used for the prediction of the fabric hand. Modelling of the fabric microstructure is a powerful tool for the in‐depth study of their performance. Based on the geometrical models of the fabrics, finite element analysis (FEA) is a very useful method for the mechanical analysis of their complex shape structures. Especially FEA can be applied on a system of bodies in contact by taking into account the interactions between the individual bodies. The parametric FEA analysis of the unit cell of the fabric provides interesting results about its mechanical behaviour.

The present work states that the use of the finite element method is a friendly and convenient method for an in‐depth study of the contact phenomena, which are dominating on the total mechanical behaviour of the fabrics.

This paper provides a simplified two‐dimensional representation of a unit cell of a fabric that involves three bodies in contact. The parametric FEA analysis of the unit cell of the fabric provides interesting results.

The purpose of this paper is to study the integration between this technology and barrel finishing (BF) operation to improve part surface quality. Fused deposition modeling (FDM) processes have limitation in term of accuracy and surface finishing. Hence, post-processing operations are needed. A theoretical and experimental investigations have been carried out.

A geometrical model of the profile under the action of machining is proposed. The model takes into account FDM formulation and allows to predict the surface morphology achievable by BF. The MR needed in the model is obtained by a particular profilometer methodology, based on the alignment of Firestone–Abbot (F–A) curves. The experimental performed on a suitable geometry validated geometrical model. Profilometer and dimensional measurements have been used to assess the output of the coupled technologies in terms of surface roughness and accuracy.

The coupling of FDM and BF has been assessed and characterized in terms of obtained part surfaces and dimension evolution. Deposition angle strongly affects the BF removal speed and alters nominal dimensions of part. The geometric profile model gave interesting information about profile morphology and machining mechanism; moreover, the height prevision allows to estimate BF working time to accomplish part requirements.

The prediction of the geometric profile as a function of FDM fabrication parameters is a powerful tool which permits to investigate surface properties such as mechanical coupling or tribological aspects. The coupling of BF and FDM has been assessed and now optimization of this process can be performed just evaluating effects of parameters.

This research has been focused to an industrial application, and results can be used in a computer-aided manufacturing. The prevision of surface obtainable by this integration is a tool to find the part optimum orientation to accomplish the drawing requirements. Both the experimental findings and the model can guide operator toward a proper process improvement, thus reducing or eliminating expensive trial and error phase in the post-processing operation of FDM prototypes.

In this paper, a novel model has been presented. It allows to know in advance profile morphology achievable by a specific surface of a FDM part after a determined BF working time. A particular application of FA curves gives the MR values.

This paper aims to present a multi-scale stochastic damage model (SDM) for concrete and apply it to the stochastic response analysis of reinforced concrete shear wall structures.

The proposed SDM is constructed at two scales, i.e. the macro-scale and the micro-scale. The general framework of the SDM is established on the basis of the continuum damage mechanics (CDM) at the macro-scale, whereas the detailed damage evolution is determined through a parallel fiber buddle model at the micro-scale. The parallel buddle model is made up of micro-elements with stochastic fracture strains, and a one-dimensional random field is assumed for the fracture strain distribution. To represent the random field, a random functional method is adopted to quantify the stochastic damage evolution process with only two variables; thus, the numerical efficiency is greatly enhanced. Meanwhile, the probability density evolution method (PDEM) is introduced for the structural stochastic response analysis.

By combing the SDM and PDEM, the probabilistic analysis of a shear wall structure is performed. The mean value, standard deviation and the probability density function of the shear wall responses, e.g., shear capacity, accumulated energy consumption and damage evolution, are obtained.

It is noted that the proposed method can reflect the influences of randomness from material level to structural level, and is efficient for stochastic response determination of shear wall structures.

This paper computationally estimates the constitutive relationships of composite materials reinforced by single walled carbon nanotubes (SWNT).

A multiscale analysis is considered. At the nanoscale level, molecular dynamics (MD) are used to predict the stiffness for an equivalent beam. A BEM solver for the elasticity problems is extended to allow the presence of inclusions and hence is used to model a RVE for the composite matrix with the equivalent nanotube beams. A genetic algorithm (GA) is developed to generate an initial population of anisotropic materials based on FEM. The GA evolves the population of properties of anisotropic materials till a material is found whose mechanical response is the same as that of the nanocomposite.

The overall process is suitable for the constitutive relationships estimation according to the verification process outlined.

The present work is limited to 2D linear problems. However, extending it to 3D non‐linear applications is straight forward.

The present technique could be used to estimate properties of NCT composites, hence practical applications such as aeroplane structures or turbine blades could be analysed using commercial finite element software. The present methodology could be used to estimate non‐mechanical properties such as the thermal and electric properties.

The present computational technique has never been presented in the literature.