Journal Name:Journal of Photochemistry and Photobiology A: Chemistry
Journal ISSN:
IF:0
Journal Website:
Year of Origin:0
Publisher:
Number of Articles Per Year:0
Publishing Cycle:
OA or Not:Not
Journal of Photochemistry and Photobiology A: Chemistry ( IF 0 ) Pub Date: 2022-01-27 , DOI:
10.1007/s40192-021-00248-x
The development of a highly efficient methodology for establishing squeeze casting process parameters from past data is essential. However, designing squeeze casting process parameters based on past data is difficult when there are many missing values. Conventional missing data approaches are fraught with additional computational challenges when applied to high-dimensional multivariable missing data, especially material process data with correlation. As the relationship between material composition and process parameters has similar characteristics with that between users and information of interest, this paper proposes a method for missing data imputation based on a clustering-based collaborative filtering (ClubCF) algorithm to address this challenge. Data samples with and without missing values were divided into two groups. K-means clustering based on a canopy algorithm was applied to the data samples without missing values to obtain k subclass data, whose values were then selected to fill data samples with missing values via a collaborative filtering theory based on Pearson similarity user filling. The missing squeeze casting process parameters data of aluminum alloys were used to evaluate the method, and more comparative experiments were carried out to understand their performance and features. Two different indicators, including the mean absolute error and the standard deviation, were utilized to quantify the imputation performance, which was compared with those of three conventional methods (mean interpolation, regression interpolation, and the expectation maximization algorithm). The results indicate that the proposed approach is effective and outperforms conventional methods in processing high-dimensional correlated data.
Journal of Photochemistry and Photobiology A: Chemistry ( IF 0 ) Pub Date: 2022-09-28 , DOI:
10.1007/s40192-022-00277-0
This paper presents an uncertainty quantification (UQ) framework for the physics-based model prediction of material response with a large number of parameters. The application problem presented in this work is that of predicting creep in Grade 91 steel at 600 °C. The material response is defined with a physically based microstructural model with constitutive equations emulating several observed phenomena in Grade 91 and embodied into an explicit geometry mesoscale finite element model for prior austenite grains and grain boundaries. Creep within the grains and in grain boundaries are represented by crystal plasticity for dislocation motion and a physics-based model for cavity growth and nucleation, respectively. The creep behavior of this material is influenced by several parameters, some of which have a wide range of variation based on experimental data. UQ combined with microstructural modeling can discover the core microstructural causes of experimental variability, leading to improved materials with lower variability in critical long-term material properties. In this study, we investigate the model's uncertainty to identify material properties that may be modified during production to increase creep life and analyze different components of the crystal plasticity model for improvements. For this purpose, a quantity of interest is defined as time to minimum creep rate, which correlates well to the creep failure of the material. A deep neural network model was trained and validated to be used as a surrogate for the finite element model. Then, a variance-based sensitivity analysis is performed on the surrogate model to find the Sobol indices of the input parameters in respect to the output quantity of interest. The Sobol indices are used to reduce the dimensionality of the model. Generalized polynomial chaos expansion is used on the reduced basis models to propagate the uncertainty from the input parameters to the quantity of interest using the deep neural network surrogate model. These results are benchmarked against uncertainty propagation using Monte Carlo simulations. The UQ performed through the reduced basis model captures almost all the uncertainty in the model with significantly fewer simulations, making it possible to perform the UQ directly via simulations with the finite element model rather than surrogate machine-learned models.
Journal of Photochemistry and Photobiology A: Chemistry ( IF 0 ) Pub Date: 2022-01-14 , DOI:
10.1007/s40192-021-00246-z
Integrated Computational Materials Engineering (ICME)-based tools and techniques have been identified as the best path forward for distortion mitigation in thin-plate steel construction at shipyards. ICME tools require temperature-dependent material properties—including specific heat, thermal conductivity, coefficient of thermal expansion, elastic modulus, yield strength, flow stress, and microstructural evolution—to achieve accurate computational results for distortion and residual stress. However, the required temperature-dependent material property databases of US Navy-relevant steels are not available in the literature. Therefore, a comprehensive testing plan for some of the most common marine steels used in the construction of US Naval vessels was completed. This testing plan included DH36, HSLA-65, HSLA-80, HSLA-100, HY-80, and HY-100 steel with a nominal thickness of 4.76 mm (3/16-in.). This report is the second part of a seven-part series detailing the pedigreed steel data. The first six reports will report the material properties for each of the individual steel grades, whereas the final report will compare and contrast the measured steel properties across all six steels. This report will focus specifically on the data associated with HSLA-65 steel.
Journal of Photochemistry and Photobiology A: Chemistry ( IF 0 ) Pub Date: 2022-03-21 , DOI:
10.1007/s40192-022-00256-5
Fatigue indicator parameters (FIPs) serve as the measurement of the driving force for fatigue crack initiation (FCI). The extreme value distributions of FIPs for fatigue life scatter (i.e., fatigue crack initiated from surface or subsurface) in the high cycle fatigue (HCF) to very high cycle fatigue (VHCF) regime are computed of AA7075-T6. The FIPs are averaged in the volume of sub-band in the scale of sub-grain proposed based on the concept of slip system of AA7075-T6 where sub-bands are extracted from statistical volume elements (SVEs) using open-source called Dream-3D. SVEs are built via the experimental crystalline microstructure. The material response variables required in the computation of FIPs are obtained by embedding crystal plasticity theory to the crystal plasticity finite element model via open-source software PRISMS-Fatigue. To meet statistical requirements, 50 instantiations of SVEs are modeled for both traction-free and fully periodic boundary conditions, respectively, corresponding to the surface and subsurface cases under different strain amplitudes (i.e., 0.362%, 0.312%, 0.289%, and 0.254%). It is found that the FIPs are higher under high stress levels indicating that fatigue crack tends to initiate at high stress levels. In addition, the low driving force can be found in the simulations with fully periodic boundary conditions throughout the SVEs, whereas the higher FIPs under traction-free boundary conditions tend to form close to the free surface under cyclic loading conditions. Furthermore, the highest FIPs have a propensity to scatter along with the depth of SVEs with the traction-free surface under strain amplitude of 0.254%, which means the fatigue life in the VHCF regime has a higher scatter. The interactions between the high symmetry of FCC crystalline structure and free surface combining the loading amplitudes have a compound effect on the FIPs, especially at low stress amplitudes, which also has a greater impact on the ultra-high-cycle fatigue life. To this end, the VHCF and HCF experiments are performed under an ultrasonic fatigue experimental system with a frequency of 20 kHz with different stress levels corresponding to the strain amplitude in the simulations. Besides, the morphology fracture surface of presentative samples was observed via the scanning electron microscope (SEM). The comparison between the simulation and the experimental results indicates that the EVD of FIPs is efficient in preciously predicting the FCI and well explain the fatigue life scatter of the AA7075-T6 in the HCF to VHCF regime.
Journal of Photochemistry and Photobiology A: Chemistry ( IF 0 ) Pub Date: 2023-01-24 , DOI:
10.1007/s40192-022-00287-y
The research and development cycle of advanced manufacturing processes traditionally requires a large investment of time and resources. Experiments can be expensive and are hence conducted on relatively small scales. This poses problems for typically data-hungry machine learning tools which could otherwise expedite the development cycle. We build upon prior work by applying conditional generative adversarial networks (GANs) to scanning electron microscope (SEM) imagery from an emerging advanced manufacturing process, shear-assisted processing and extrusion (ShAPE). We generate realistic images conditioned on temper and either experimental parameters or material properties. In doing so, we are able to integrate machine learning into the development cycle, by allowing a user to immediately visualize the microstructure that would arise from particular process parameters or properties. This work forms a technical backbone for a fundamentally new approach for understanding manufacturing processes in the absence of first-principle models. By characterizing microstructure from a topological perspective, we are able to evaluate our models’ ability to capture the breadth and diversity of experimental scanning electron microscope (SEM) samples. Our method is successful in capturing the visual and general microstructural features arising from the considered process, with analysis highlighting directions to further improve the topological realism of our synthetic imagery.
Journal of Photochemistry and Photobiology A: Chemistry ( IF 0 ) Pub Date: 2022-10-19 , DOI:
10.1007/s40192-022-00278-z
Additive manufacturing (AM) is gaining attention because of its ability to design complex geometries. Direct energy deposition (DED), one of the AM processes, is widely used nowadays for its high deposition rate. When using DED process in manufacturing or repairing, it is important to know the melt pool dimensions as a function of processing parameters to obtain high deposition rate and avoid defects such as lack of fusion. In this study, we used the random forest (RF) algorithm to predict melt pool dimensions and compared the results against existing physics based lumped model by Doumanidis et al. [1]. The results show that RF model works well to predict the DED melt pool dimensions, where energy density and material volume deposited govern the dimensions. Further, we tested the ubiquitous semi-ellipsoidal shape assumption for DED cross section against the circular shape and found semi-ellipsoidal shape to be fair when deposition process is stable and free of defects. Overall, this study highlights the applicability of machine learning algorithms for small AM datasets.
Journal of Photochemistry and Photobiology A: Chemistry ( IF 0 ) Pub Date: 2022-10-17 , DOI:
10.1007/s40192-022-00279-y
In recent years, the design and development of materials are strongly interconnected with the development of digital technologies. In this respect, efficient data management is the building block of material digitization and, in the field of materials science and engineering (MSE), effective solutions for data standardization and sharing of different digital resources are needed. Therefore, ontologies are applied that represent a map of MSE concepts and relationships between them. Among different ontology development approaches, graphical editing based on standard conceptual modeling languages is increasingly used due to its intuitiveness and simplicity. This approach is also adopted by the Materials-open-Laboratory project (Mat-o-Lab), which aims to develop domain ontologies and method graphs in accordance with testing standards in the field of MSE. To suit the actual demands of domain experts in the project, Ontopanel was created as a plugin for the popular open-source graphical editor diagrams.net to enable graphical ontology editing. It includes a set of pipeline tools to foster ontology development in diagrams.net, comprising imports and reusage of ontologies, converting diagrams to Web Ontology Language (OWL), verifying diagrams using OWL rules, and mapping data. It reduces learning costs by eliminating the need for domain experts to switch between various tools. Brinell hardness testing is chosen in this study as a use case to demonstrate the utilization of Ontopanel.
Journal of Photochemistry and Photobiology A: Chemistry ( IF 0 ) Pub Date: 2022-05-13 , DOI:
10.1007/s40192-022-00262-7
Ex-situ estimation of strains from deformed micrographs is not possible as there are no persistent features which can be tracked. Two point spatial statistics enable the rigorous quantification of spatial patterns in heterogeneous media. In this paper, we propose a novel method for estimating strains directly from dissimilar micrographs using a continuum mechanics approach. Rather than operating directly on images from sequential frames, as is done in digital image correlation, we operate on different microstructure realizations. This is made possible by comparing the spatial autocorrelation maps of deformed and undeformed micrographs rather than direct comparison of images. A Bayesian framework is proposed for quantifying uncertainty. We first illustrate the efficacy of this method on speckle pattern images from digital image correlation experiments. Then, we demonstrate that the method is capable of operating on dissimilar micrographs using deformed synthetic binary microstructures. Finally, we present a case study on polycrystalline additively manufactured 316L deformed via tension. The proposed method works well and we discuss implications and limitations of the presented work.
Journal of Photochemistry and Photobiology A: Chemistry ( IF 0 ) Pub Date: 2021-11-29 , DOI:
10.1007/s40192-021-00242-3
Traditional, data-driven materials discovery involves screening chemical systems with machine learning algorithms and selecting candidates that excel in a target property. The number of screening candidates grows infinitely large as the fractional resolution of compositions and the number of included elements increase. The computational infeasibility and probability of overlooking a successful candidate grow likewise. Our approach takes inspiration from neural style transfer and shifts the optimization focus from model parameters to the fractions of each element in a composition. By leveraging a pretrained network with exceptional prediction accuracy (CrabNet) and writing a custom loss function to govern a vector consisting of element fractions, material compositions can be optimized such that a predicted property is maximized or minimized. It is expected that fractional optimization would excel in inverse design problems concerning the effects of dopants and those seeking a balance between competing properties.
Journal of Photochemistry and Photobiology A: Chemistry ( IF 0 ) Pub Date: 2022-09-28 , DOI:
10.1007/s40192-022-00275-2
The evolution of texture during different mechanical processes is an integral part of Integrated Computational Materials Engineering (ICME) of metals and alloys. In the present work, PRISMS-Plasticity TM, an open-source rapid texture evolution analysis pipeline, based on the Taylor model is integrated into the open-source crystal plasticity software, PRISMS-Plasticity. However, unlike PRISMS-Plasticity, the pipeline does not involve any finite element solver. The Taylor-type crystal plasticity model is used to capture the evolution of texture. The pipeline aims to significantly minimize the cost of texture evolution simulations, so one can use the framework along with data-driven ICME design methods and generate high-throughput data with low costs for general analysis or machine learning applications. All the rate-independent and rate-dependent crystal plasticity models available in the Materials library of PRISMS-Plasticity can be used in the pipeline. The models can capture extension twinning as an important deformation mechanism in alloys with hexagonal close-packed crystal structures. The PRISMS-Plasticity TM pipeline is efficiently integrated with experimental characterization techniques such as electron backscatter diffraction (EBSD), and these pipelines use other open-source software packages such as DREAM.3D where appropriate. In addition, the PRISMS-Plasticity TM results are integrated with the PRISMS Center Materials Commons information repository. In this investigation, three different examples of texture evolution in a polycrystalline OFHC copper, extruded Mg alloy ZK60A, and an aluminum alloy 7075-T6 are presented to show the capability of PRISMS-Plasticity TM. The pipeline can take advantage of the parallel performance of PRISMS-Plasticity and scales exceptionally well for large problems running on multiple processors. The computational efficiency of this tool can be used to build large databases for use in machine learning applications. Such a large dataset is established containing process–microstructure relationships of 7075-T6 Al. The extracted data is visualized in 2-D latent space using an unsupervised machine learning model called the variational autoencoder (VAE). The trained VAE model is used to identify economical process sequences which lead to desired properties.
Supplementary Information
Self Citation Rate | H-index | SCI Inclusion Status | PubMed Central (PML) |
---|---|---|---|
0 | Not |