Complicated Composite Structures Research Articles

Machine learning-based inverse design methods considering data characteristics and design space size in materials design and manufacturing: a review.

In the last few decades, the influence of machine learning has permeated many areas of science and technology, including the field of materials science. This toolkit of data driven methods accelerated the discovery and production of new materials by accurately predicting the complicated physical processes and mechanisms that are not fully described by existing materials theories. However, the availability of a growing number of increasingly complex machine learning models confronts us with the question of "which machine learning algorithm to employ". In this review, we provide a comprehensive review of common machine learning algorithms used for materials design, as well as a guideline for selecting the most appropriate model considering the nature of the design problem. To this end, we classify the material design problems into four categories of: (i) the training data set being sufficiently large to capture the trend of design space (interpolation problem), (ii) a vast design space that cannot be explored thoroughly with the initial training data set alone (extrapolation problem), (iii) multi-fidelity datasets (small accurate dataset and large approximate dataset), and (iv) only a small dataset available. The most successful machine learning-based surrogate models and design approaches will be discussed for each case along with pertinent literature. This review focuses mostly on the use of ML algorithms for the inverse design of complicated composite structures, a topic that has received a lot of attention recently with the rise of additive manufacturing.

Lamb wave-based damage localization and quantification algorithms for CFRP composite structures

Accurately and quantitatively monitoring the damages in complicated composite structures is very challenging, but essential for assessing their integrity. In this paper, an innovative damage monitoring method based on Lamb wave is proposed for large-scale complicated composite laminates. A combination of a resource-constraint Levenberg-Marquardt algorithm associated with a quantum-inspired gravitational search algorithm is introduced to identify the damage location. A damage contour algorithm using convex envelope of damage reflection points and maximum inscribed n-polygon is developed to determine the quantitative damage size. Meanwhile, a singular loci removal scheme based on an automatically updated boundary region constraint and bivariate normal distribution is utilized to separate the correctly positioned points on the damage periphery from those associated with noise. Experiments on two large-scale composite panels with multiple damage cases are conducted to substantiate the proposed algorithms. Sensor layers with embedded piezoelectric transducer (PZT) sensor networks mounted on the structures are employed to excite and receive Lamb wave signals because of their superiority of flexibility, electrical stability and apt for complex structures. The results demonstrate that the proposed technique is capable of pinpointing both location and size of the damage in the complicated composite structures.

Identification and imaging of multi-defects on a complicated composite structure by ultrasonic guided wave

Combination of probabilistic damage imaging (PDI) algorithm and the fusion ultrasonic guided wave damage index can be used to detect the damage in a complicated composite structure with acceptable accuracy. However, the damage signatures associated with PDI algorithm cannot accurately identify multi-defects features such as the number, position and dimension information, which greatly restrict the inspection effect in real practice. In this paper, the damage shape factor βM of each damage-impaired path is optimized to enhance the resolution of PDI algorithm. Then, a modified probabilistic damage imaging (MPDI) algorithm with the βM is proposed to identify the multi-defects existed in the curved carbon/epoxy composite structures. The validity of the algorithm is firstly checked in the curved composite structures with the single defect. The MPDI with the βM is applied to identify the depth of defect. The obtained results show that the precision of different defects identification by the MPDI algorithm is much higher than that used by the PDI algorithm. The number, position and dimension of multi-defects existed in the curved composite structure are clearly imaged by the method. This study also verifies that the majorization of damage shape factor in the MPDI algorithm is practicable and necessary for better identification of multi-defects in composite structure.

Solid Truss to Shell Numerical Homogenization of Prefabricated Composite Slabs.

The need for quick and easy deflection calculations of various prefabricated slabs causes simplified procedures and numerical tools to be used more often. Modelling of full 3D finite element (FE) geometry of such plates is not only uneconomical but often requires the use of complex software and advanced numerical knowledge. Therefore, numerical homogenization is an excellent tool, which can be easily employed to simplify a model, especially when accurate modelling is not necessary. Homogenization allows for simplifying a computational model and replacing a complicated composite structure with a homogeneous plate. Here, a numerical homogenization method based on strain energy equivalence is derived. Based on the method proposed, the structure of the prefabricated concrete slabs reinforced with steel spatial trusses is homogenized to a single plate element with an effective stiffness. There is a complete equivalence between the full 3D FE model built with solid elements combined with truss structural elements and the simplified homogenized plate FE model. The method allows for the correct homogenization of any complex composite structures made of both solid and structural elements, without the need to perform advanced numerical analyses. The only requirement is a correctly formulated stiffness matrix of a representative volume element (RVE) and appropriate formulation of the transformation between kinematic constrains on the RVE boundary and generalized strains.

A novel aspect of composite sandwich fairing structure optimization of a two-stage launch vehicle (Safir) using multidisciplinary design optimization independent subspace approach

In this article, a novel composite sandwich structure analysis of launch vehicle fairing is considered by a new multidisciplinary design optimization methodology. Among the most important roles of this method, in addition to the convergence of optimization process, is tackling with complicated composite structure discipline. The bidirectional coupling existing between composite structure and trajectory disciplines is one of the complex problems of this multidisciplinary design optimization method. Accordingly, multidisciplinary design optimization based on independent subspaces is employed using the fixed point iteration method to achieve the best convergence at system level and segregate the disciplines. Therefore, the two proposed subspaces overcome the difficulties of common mentioned multidisciplinary design optimization of launch vehicles as the main novelty of this study. The first subspace is a multidisciplinary design optimization which includes propulsion, aerodynamics, weight and trajectory disciplines. The second one includes the novel composite fairing structure optimization as the other single discipline optimization which is analytically and numerically considered as a compact problem and is regarded as the other novelty of this work. In a case study, by applying the proposed architecture on Safir launch vehicle and considering propulsion, trajectory and also composite sandwich fairing structure design as the variables and then performing an optimization process, the Safir fairing mass is reduced from 100 kg to 57.8 kg. This causes the launch vehicle gross mass to decrease from 26 tons to 25.2 tons due to the payload nature of fairing. This 3% mass decrease of an operational launch vehicle, despite the preservation of mission performance, can be called an industrial novelty which leads to the cost reduction of space transportation. The proposed system engineering demonstrates the great importance of using multidisciplinary design optimization in complicated designs using independent subspaces to be employed for the design of future launch vehicles. It can also be a road map for future designers of space vehicles, especially those who want to consider structure optimization in the design loop of launch vehicles.

Magnetostatic interaction in multi-shell Ni80Fe20/Cu composite wires

Magnetostatic interaction in multilayered structure materials has long been a key to optimize the magnetic devices due to the modulation of their magnetic properties. In this study, the multi-shell Ni80Fe20(tout)/Ni80Fe20(tin)/Cu composite wires were prepared by magnetron sputtering. The results show that the hysteresis loops of the multi-shell composite wires do not present significant changes compared with the single shell composite wire. A four-peak structure GMI profile may appear when both the inner and outer Ni80Fe20 shell have a preferred circular anisotropy, which indicated that the strong dipole interactions existed between the shells. In such multi-shell system, the response of dipole field Hdi to the driving frequency is against the compensation field Hcomp. If the inner shell has axial anisotropy structure, there are almost no dipole interaction observed according to disappearance of the interaction energy. All the results suggest a facile way to achieve a higher GMI response, and a sensitive method, especially in the low magnetic field region, to investigate the magnetic interactions in complicated composite structure materials.

Effective dielectric constants and spectral density analysis of plasmonic nanocomposites

Cermet or ceramic-metal composite coatings promise great potentials in light harvesting, but the complicated composite structure at the nanoscale induces a design challenge to predict their optical properties. We find that the effective dielectric constants of nanocomposites predicted by finite-difference-time-domain (FDTD) simulation results match those of different classical effective medium theories in their respective validity range. However, a precise prediction of the fabricated nanocomposite properties for different filling factors is very challenging. In this work, we extract the spectral density functions in the Bergman representation from the analytical models, numerical simulations, and experimental data of plasmonic nanocomposites. The spectral density functions, which only depend on geometry of the nanocomposite material, provide a unique measure on the contribution of individual and percolated particles inside the nanocomposite. According to the spectral density analysis of measured dielectr...

광 적외선열화상을 이용한 풍력 블레이드의 결함 크기 정량화 연구

Abstract A wind turbine blade is an important component in wind-power generation, and is generally exposed toharsh environmental conditions. Ultrasonic inspection is mainly used to inspect such blades, but it has beendifficult to quantify defect sizes in complicated composite structures. Recently, active infrared thermography hasbeen widely studied for inspecting composite structures, in which thermal energy is applied to an object, and aninfrared camera detects the energy emitted from it. In this paper, a calibration method for active optical lock-inthermography is proposed to quantify the size. Inclusion, debonding and wrinkle defects, created in a wind bladefor 100 kW wind power generation, were all successfully detected using this method. In particular, a 50.0φ mmdebonding defect was sized with 98.0% accuracy. Keywords: Wind Turbine Blade, Optical Infrared Thermography, Lock-in Method, Quantitative Defects Detection[Received: December 18, 2014, Revised: February 3, 2015, Accepted: February 17, 2015] *과 학기술연합 대원교(UST) ** ,항공기시스템공학 한국표준과학연구원 안전측정센터 Corresponding Author: Safety MeasurementCenter, Korea Research Institute of Standards and Science, Daejeon 305-340, Korea (E-mail: mychoi@kriss.re.kr)

Mechanical Performance and Failure Mechanism of Thick-walled Composite Connecting Rods Fabricated by Resin Transfer Molding Technique

A resin transfer molding technique was used to fabricate thick-walled composite connecting rods, and then the mechanical performance of the connecting rod was studied experimentally, at the same time the stress and failure index distributions were simulated numerically. The experimental results show that under a tensile load, the connecting rod first cracks near the vertex of the triangle areas at the two ends, and then the damage propagates along the interface between the main bearing beam and the triangle area as well as along the round angle of the triangle area. Whereas under a compressive load, the delamination primarily occurs at the corner of the U-shaped flange, and the final destruction is caused by the fracture of fibers in the main bearing beam. The simulated results reveal that the tensile failure is originated from the delamination at the round angle transition areas of the T-joints, and the failure strength is determined by the interlaminar strength. Whereas the compressive failure is caused by the fracture of fibers in the main bearing beam, and the failure strength of the structure is determined by the longitudinal compressive strength of the composite material. The simulated results are basically consistent with the experimental results. Hence the mechanical performance and failure mechanism of the complicated composite structure are revealed in great detail through the coupling of the two kinds of research methods, which is helpful for the optimal design of composite structures.

Augmented Cohesive Elements for Efficient Delamination Analyses of Composite Laminates

In this paper, a new type of cohesive element that employs multiple subdomain integration (MSDI) for improved cohesive stress integration accuracy of bonded plate/shell elements has been formulated. Within each subdomain, stress integration is compatible with existing schemes such as Gaussian integration (GI), Newton–Cotes integration, or the mixed Gaussian and subdomain integration (mixed GI&SDI). The numerical accuracy, efficiency, and robustness of this element when employing three integration methods for MSD cohesive stress integration have been evaluated and compared through a benchmark mode-I fracture problem of bonded double-cantilever plates. The MSDI offers at least 50% improvement of numerical accuracy as compared to the best integration method in literature and has the best numerical robustness. This significant improvement pushes the structural mesh size restriction from limiting size of 1/3–1/5 cohesive zone length to 1.5–2 times the cohesive zone length. The formulation is very easy to be implemented into any finite element programs including commercial packages. Furthermore, this formulation enables the use of dual-mesh for delamination analyses of bonded structural shells/plates, which is of practical importance because it greatly reduces the burden of mesh generation for complicated composite structures. It has also been demonstrated that using high-order shell/plate elements can improve the numerical accuracy in general because the nonlinear deformation profile permitted by this type of elements can better describe the nonlinear deformation in the crack-tip element (partially bonded elements).

Cross-section TEM investigation of quasicrystalline catalysts prepared by aqueous NaOH leaching

The Al–Cu–Fe quasicrystal (QC) has been shown to reveal high catalytic activity and thermal stability for steam reforming of methanol (SRM: CH3OH + H2O → 3H2 + CO2), after leaching with NaOH aqueous solution. To investigate the origin of the high catalytic performance, cross-sectional transmission electron microscopy (TEM) observations of the catalyst particles were examined. By leaching with NaOH aq., a homogeneous composite structure, composed of Cu and Fe and/or their oxides or hydroxides, is formed at the outer layer of catalyst particles and the QC phase remains in the inner area. The homogeneous dispersion of Fe and/or the oxide/hydroxide in the composite suppresses sintering of Cu particles. In contrast, for the catalyst prepared from crystalline ω-phase (Al70Cu20Fe10), the original phase completely collapses and, instead, a complicated composite structure is formed. The inhomogeneity results in the formation of Cu-enriched regions at the rim of the catalyst particles. We suggest that the high catalytic performance of the Al–Cu–Fe QC catalyst prepared by leaching with NaOH aq. is due to the formation of a homogeneous composite region at the outer layer of the catalyst particles.

Calculation of the Thermal Conductivity of Randomly Dispersed Composites using a Finite Element Modeling Method

A new method to generate 3 dimensional finite element method (FEM) models of composites with random materials arrangement has been proposed. Some basic models are used to represent the structure of the composites, in each basic model, the dispersions are assumed to have the same geometry, and the structure of a basic model can be described by some structural parameters such as the geometry, number and volume fraction of the dispersion and so on. A program has been developed to automatically generate the geometric model and FEM mesh of the basic models according to given structure parameters. More complicated composite structure can be composed by combination of the basic models. The effective thermal conductivity of composite has been calculated based on the models generated with this method. Optimization of modeling parameters such as model scale and mesh refinement has been discussed with consideration of both calculation accuracy and computation efficiency. The calculated thermal conductivity has been compared to the values obtained by analytical method.

Synthesis and Characterization of Short Carbon Fibre-Reinforced PMR-Type Polyimide Composites

Short carbon fibre-reinforced polyimide composites were prepared by wet impregnating chopped short carbon fibres with PMR-type polyimide matrix resin solution followed by evaporating the solvent with strong stirring. The processing properties of the prepolymer moulding powders were investigated. The mechanical properties and thermo-oxidative stability of the fabricated composite were also studied. Experimental results demonstrate that the short fibre filled moulding powders exhibit good melt resin flowing characteristics and can be processed by a hot press moulding technique to produce small-size, complicated composite structures which are difficult to produce using continuous fibre-reinforced composite prepregs by an autoclave technique. The composites exhibit outstanding mechanical properties and excellent thermal and thermo-oxidative stabilities at temperatures as high as 320 °C.

Use of homogenization theory to build a beam element with thermo-mechanical microscale properties

The homogenization method is used to develop a beam element in space for thermo-mechanical analysis of unidirectional composites. Local stress and temperature field in the microscale are described using the function of homogenization. The global (macroscopic) behaviour of the structure is supposed to be that of a beam. Beam-type kinematical hypotheses (including independent shear rotations) are hence applied and superposed on the microdescription. A macroscopic stiffness matrix for such a beam element is then developed which contains the microscale properties of the single cell of periodicity. The presented model enables us to analyse without too much computational effort complicated composite structures such as e.g. toroidal coils of a fusion reactor. We need only a FE mesh sufficiently fine for a correct description of the local geometry of a single cell and a few of the newly developed elements for the description of the global behaviour. An unsmearing procedure gives the stress and temperature field in the different materials of a single cell.

Sensitivity and optimization of composite structures in MSC/NASTRAN

The main purpose of this paper is to present the considerations and the resultant approach used to implement design sensitivity capability for composites into a large-scale, general-purpose finite element system ( msc/nastran). The design variables for composites can be lamina thicknesses, orientation angles, material properties, or a combination of all three. As a secondary goal, the sensitivity analysis has been coupled with a general-purpose optimizer to validate sensitivity analysis results and also demonstrate how easy it is couple to an optimizer once a general senssitivity capability is available. This preliminary version of the optimizer is capable of dealing with minimum weight structural design with a rather general design variable linking capability at the element level or the system level. Only sizing type of design variables (i.e., lamina thicknesses) can be handled by the optimizer. Test cases have been run and validated by comparison witj independent finite element packages. The linking of a design sensitivity capability for composites in msc/ nastran with an optimizer would give designsrs a powerful, automated tool to carry out practical optimization design of real-life, complicated composite structures.