Prediction Of Mechanical Properties Research Articles

Machine learning prediction of the mechanical properties of injection-molded polypropylene through X-ray diffraction analysis

ABSTRACT Predicting the mechanical properties of polymer materials using machine learning is essential for the design of next-generation of polymers. However, the strong relationship between the higher-order structure of polymers and their mechanical properties hinders the mechanical property predictions based on their primary structures. To incorporate information on higher-order structures into the prediction model, X-ray diffraction (XRD) can be used. This study proposes a strategy to generate appropriate descriptors from the XRD analysis of the injection-molded polypropylene samples, which were prepared under almost the same injection molding conditions. To this end, first, Bayesian spectral deconvolution is used to automatically create high-dimensional descriptors. Second, informative descriptors are selected to achieve highly accurate predictions by implementing the black-box optimization method using Ising machine. This approach was applied to custom-built polymer datasets containing data on homo- polypropylene and derived composite polymers with the addition of elastomers. Results show that reasonable accuracy of predictions for seven mechanical properties can be achieved using only XRD.

Prediction of mechanical properties of bricks manufactured with recycled soil using artificial neural network

Prediction of mechanical properties of bricks manufactured with recycled soil using artificial neural network

Machine learning–assisted prediction of mechanical properties of high-entropy alloy/graphene nanocomposite

The addition of graphene (Gr) to high-entropy alloys (HEAs) has recently been demonstrated to endow the alloys with increased strength and plasticity in experiments and simulations. However, determining the effects of various factors on the mechanical properties of Gr during the experimental process is difficult due to its special structure and the complexity of its components. Investigation of the mechanical properties (Young's modulus, yield strength and toughness) of HEA/Gr under different simulation conditions through experiments and simulations is not only costly in terms of raw materials but also time-consuming. This study aimed to address the aforementioned problem and investigate the mechanical properties of HEA/Gr under different conditions. First, the dataset necessary for the machine learning model was constructed using molecular dynamics simulation and downscaled from 59 to 21 features using principal component analysis. Then, five machine learning models, including XGBoost, light gradient boosting (LGBoost) machine, random forest, AdaBoost, and decision tree, were employed. Finally, 10-fold cross-validation and grid search were utilized during the training process to search for the best parametric model, and the coefficient of determination R2 and mean square error MSE were used as metrics. Results indicate the superiority of the XGBoost and LGBoost models to other models in predicting HEA/Gr mechanical properties. Among the three properties studied, the models predicting Young’s modulus and toughness revealed an R2 value above 85%. Two models realized an R2 greater than 90% for Young’s modulus, while all three models achieved an R2 greater than 85% for toughness.

Machine learning accelerated the prediction of mechanical properties of copper modified by TMDs based on molecular dynamics simulation

Abstract In this study, we constructed a dataset of elastic modulus and ultimate stress for copper material enhanced by Transition Metal Dichalcogenides (TMDs) through Molecular Dynamics (MD) simulations. Subsequently, leveraging chemical insights, we selected appropriate descriptors and established machine learning prediction models for elastic modulus and ultimate stress, respectively. Finally, the performance of the machine learning models was evaluated using a test set. The results demonstrate excellent performance of the machine learning models in predicting material properties. This work presents a novel approach for efficient material screening, demonstrating the synergy between MD simulations and machine learning in advancing materials research and intelligent material selection platforms.

Prediction of Macroscopic Compressive Mechanical Properties for 2.5D Woven Composites Based on Artificial Neural Network

Prediction of Macroscopic Compressive Mechanical Properties for 2.5D Woven Composites Based on Artificial Neural Network

In Situ Prediction of Microstructure and Mechanical Properties in Laser-Remelted Al-Si Alloys: Towards Enhanced Additive Manufacturing.

Laser surface remelting of aluminum alloys has emerged as a promising technique to enhance mechanical properties through refined microstructures. This process involves rapid cooling rates ranging from 103 to 108 °C/s, which increase solid solubility within aluminum alloys, shifting their eutectic composition to a larger value of silicon content. Consequently, the resulting microstructure combines a strengthened aluminum matrix with silicon fibers. This study focuses on the laser scanning of Al-Si aluminum alloy to reduce the size of aluminum matrix spacings and transform fibrous silicon particles from micrometer to nanometer dimensions. Analysis revealed that the eutectic structure contained 17.55% silicon by weight, surpassing the equilibrium eutectic composition of 12.6% silicon. Microstructure dimensions within the molten zones, termed 'melt pools', were extensively examined using Scanning Electron Microscopy (SEM) at intervals of approximately 20 μm from the surface. A notable increase in hardness, exceeding 50% compared to the base plate, was observed in the melt pool regions. Thus, it is exemplified that laser surface remelting introduces a novel strengthening mechanism in the alloy. Moreover, this study develops an in situ method for predicting melt pool properties and dimensions. A predictive model is proposed, correlating energy density and spectral signals emitted during laser remelting with mechanical properties and melt pool dimensions. This method significantly reduces characterization time from days to seconds, offering a streamlined approach for future studies in additive manufacturing.

High-entropy carbide (Ti0.2Zr0.2Hf0.2Nb0.2Ta0.2)C mechanical properties prediction with the use of machine learning potential

High-entropy carbide (Ti0.2Zr0.2Hf0.2Nb0.2Ta0.2)C mechanical properties prediction with the use of machine learning potential

Prediction of Microstructure and Mechanical Properties of Ultrasonically Treated PLA Materials Using Convolutional Neural Networks

Prediction of Microstructure and Mechanical Properties of Ultrasonically Treated PLA Materials Using Convolutional Neural Networks

Optimization Strategy for Process Design in Rubber Injection Molding: A Simulation-Based Approach Allowing for the Prediction of Mechanical Properties of Vulcanizates.

Selecting the optimal settings for the production of rubber goods can be a very time-consuming and resource-intensive process. A promising method for optimizing rubber processing in a short period of time is the use of simulation routines. However, process simulations have only recently enabled meaningful predictions of not only the part's state of cure but also its mechanical characteristics. As a first approach, second-order polynomials were considered suitable for describing the properties of compression-molded parts. However, more precision is required for injection molding due to the narrower distribution of mechanical characteristics of parts produced at different vulcanization temperatures. This became evident when the approximation of mechanical data with second order models partly revealed significant failures of part behavior prediction. To tackle this issue, a combined approach for approximation is proposed in this contribution by means of logistic growth function in addition to second order polynomials. To feed the model, an experimental plan was designed for producing injection-molded parts from an SBR compound at various temperatures and to different degrees of cure. The parts obtained were then characterized mechanically, and the results were opposed to varying degrees of cure and extents of reaction to calculate the model coefficients. Once available, a simulation-based calculation of the mechanical part quality is possible. The comparison of test results from the simulation and the real process has shown a reliable prediction, as simulation results were found within the natural deviation of the real measurements.

Prediction and analysis of mechanical properties of hot-rolled strip steel based on an interpretable machine learning

Establishing a low-cost, high-precision predictive model for mechanical properties is crucial for enhancing the properties of steel. Due to the complexity of the steel production process, traditional mechanistic models struggle to efficiently and accurately describe the relationships among alloy elements, processes, and properties. In response to this, the paper introduces a high-precision framework for predicting and optimizing mechanical properties through feature engineering and machine learning models. Firstly, an effective dimension reduction of features is achieved through a weighted fusion heterogeneous feature selection strategy, thereby identifying the optimal model input features. Subsequently, an improved seagull optimization algorithm is utilized to optimize the hyperparameters of Extreme Gradient Boosting, further enhancing the predictive accuracy of the model. Moreover, based on the Shapley Additive Explanation method, a quantitative analysis of the model's predictive outcomes is conducted, elucidating the impacts of alloy elements and processes on mechanical properties, consistent with established principles of physical metallurgy. The proposed framework not only enables precise prediction of mechanical properties in steel but also provides theoretical guidance and technical support for process optimization design and the development of new steel grades.

Theoretical model for elastic modulus prediction on basis of atomic structure information: Beginning from amorphous carbon to covalent bonding materials

The mechanical property is one of the most important properties of a material and is determined by the atomic-scale structure. It is thus crucial to establish the theoretical model for the prediction of materials' mechanical properties, benefiting the material design. In our previous work, we proposed an atomic-scale structure-based model for predicting the Young's modulus of hydrogenated amorphous carbon; however, the application range of this model is too narrow to be used practically for the materials development. Therefore, in this work, an extended prediction model of Young's modulus of materials with wide applicability is successfully constructed, based on the two important inherent properties of materials: one is effective coordination number (CNeff) evaluating how densely atoms are structured and the another one is effective bond stiffness (Keff) that is first proposed here and indicates how bonding types contribute the elastic properties. Through the high-throughput molecular dynamics simulations, the predictive model of Young's modulus (E) is determined as E = 3.37Keff (CNeff - 2.0)1.5. Then we demonstrate that this model is valid for a large variety of materials including both amorphous and crystalline structures, and the accuracy is proven by comparing with other work. Overall, this fundamental work may benefit the preparation, development, and utilization of new materials.

Cr content dependent lattice distortion and solid solution strengthening in additively manufactured CoFeNiCrx complex concentrated alloys – a first principles approach

In the previous study, the authors reported that multicomponent CoFeNiCrx (x = 0, 12, 18, 24 at%) complex concentrated alloys fabricated through laser directed energy deposition experienced a substantial reduction in the saturation magnetization (Ms) and the Curie transition temperature (Tc) with the increase in Cr content. The present investigation additionally explores the effect of Cr concentration on the structural and strengthening behavior in laser directed energy deposition processed CoFeNiCrx alloy system. The novel approach adopted in the present work included the integration of experimental and computational efforts involving laser direct energy deposition (L-DED) based synthesis/fabrication of a series of new CoFeNiCrx alloys through in-situ compositional tuning which is often difficult using conventional techniques and prediction of mechanical properties of these compositionally tuned alloys using a computational methodology based on the density functional theory (DFT). X-ray diffraction and nanoindentation studies revealed that while there was a slight reduction in lattice parameters, the elastic modulus increased by approximately 30 % and the yield strength increased by 25 % with increasing Cr content. The density functional theory framework was employed to interpret the influence of Cr atoms on local lattice distortions and observed yield strength. The calculated local lattice distortion, in terms of first nearest neighbors, significantly increased with Cr concentration, which is attributed to the enlargement of Cr atomic radius in the alloyed multielement environment in comparison to its radius in the elemental form. This distortion at the local level caused solid solution strengthening. The theoretically predicted solid solution strengthening values using the Varvenne model are consistent with experimental results.

Study of the effect of heat temperature on the chemical changes and hygroscopicity of eucalyptus wood by FT-IR and prediction of mechanical properties by the MLR regression method

This study describes the effect of heat treatment on some physical, chemical, and mechanical properties of Eucalyptus Camaldulensis (EC) wood at different temperatures and treatment times (200 °C–260 °C for 5, 60, and 90 min). The evaluation of hygroscopic properties was determined by relative humidity, mass loss, dimensional stability tests, and density. The results showed that the heat treatment leads to an increase in mass loss of 5.2 %–11.9 % at 200 °C. The density changed significantly for this studied species as well as the dimensional stabilization. Chemical changes in wood structure were assessed by Fourier Transform Infrared Spectroscopy.To verify the validity of the superposition “Mass loss-Density-water absorption” on the mechanical properties (modulus of elasticity (MOE) and modulus of rupture (MOR)) during heat treatment, we have developed a mathematical model based on Multiple Linear Regression (MLR), in order to establish a relationship between the independent parameters and the dependent parameters (MOE and MOR). The evaluation of the quality of the models developed was based on several statistical tools, namely R = 0.99, R2 = 0.99, R2adj = 0.98, and F = 132.33. The results demonstrated that elaborate models of mechanical properties have a high predictive capacity (MOR and MOE). The wood’s carbohydrates (particularly hemicelluloses) are then degraded during the heat treatment. The % of carbon increases from 47.8 to 49.8 %, which is proportional to mass loss, while the % of oxygen decreases by 46.1 %, which is inversely proportional to mass loss. Furthermore, FTIR analysis revealed that the effect of heat-treated wood chemical changes was related to the hydroxyl OH function of cellulose, functional groups, and aromatic system of lignin. In conclusion, the results demonstrated that at 200 °C, heat treatment caused a 5.2–11.9 % increase in mass loss; dimensional stability and density underwent considerable changes. FTIR spectroscopy confirmed the chemical changes in the wood structure during heat treatment. Furthermore, the “MLR” mathematical model showed that density contributed to the increase in MOR and MOE properties, while water absorption and mass loss contributed to the decrease in MOR and MOE properties. Finally, the % of oxygen decreased by 46.1 %, which is inversely proportional to the loss of mass, and the % of carbon increased from 47.8 % to 49.8 %, which is proportional to the loss of mass.

Prediction of mechanical properties of high‐performance concrete and ultrahigh‐performance concrete using soft computing techniques: A critical review

AbstractA cement‐based material that meets the general goals of mechanical properties, workability, and durability as well as the ever‐increasing demands of environmental sustainability is produced by varying the type and quantity of individual constituents in high‐performance concrete (HPC) and ultrahigh‐performance concrete (UHPC). Expensive and time‐consuming laboratory experiments can be used to estimate the properties of concrete mixtures and elements. As an alternative, these attributes can be approximated by means of predictive models created through the application of artificial intelligence (AI) methodologies. AI approaches are among the most effective ways to solve engineering problems due to their capacity for pattern recognition and knowledge processing. Machine learning (ML) and deep learning (DL) are a subfield of AI that is gaining popularity across many scientific domains as a result of its many benefits over statistical and experimental models. These include, but are not limited to, better accuracy, faster performance, greater responsiveness in complex environments, and lower economic costs. In order to assess the critical features of the literature, a comprehensive review of ML and DL applications for HPC and UHPC was conducted in this study. This paper offers a thorough explanation of the fundamental terms and ideas of ML and DL algorithms that are frequently used to predict mechanical properties of HPC and UHPC. Engineers and researchers working with construction materials will find this paper useful in helping them choose accurate and appropriate methods for their needs.

Theoretical prediction of phase transition, mechanical and electronic properties of manganese diboride under pressure

Theoretical prediction of phase transition, mechanical and electronic properties of manganese diboride under pressure

Spatial reconstruction, microstructure-based modeling of compressive deformation behavior, and prediction of mechanical properties in lightweight Al-based entropy alloys

Lightweight Al-based entropy alloys are newly emerged materials with promising potential for structural applications; however, their modeling and strengthening prediction require further exploration. In this study, the spatial distributions of various phases in three Al-based entropy alloys were investigated. A closely interconnected intermetallic compound (IC) network comprising Al2Cu and Al45Cr7 phases was identified. Finite element (FE) modeling was conducted based on the reconstructed three-dimensional microstructure to simulate compressive deformation behavior. The IC network served as the main stress bearer during deformation. Thin sections in the Al2Cu network were the weak sites where stress concentration and damage first occurred. However, the breakage of this limited region contributes to relatively coordinated deformation and extended plasticity. The breakage of large particles accounts for the final alloy failure. The results of the FE model were compared with the experimentally measured stress–strain behavior and mechanical properties, showing very good agreement. Given the non-uniform distribution of strain and stress during deformation, a strengthening model, merging the Voigt and Reuss models, was also developed to predict the mechanical properties of Al-based entropy alloys with the aim of facilitating Al-based entropy alloy development.

Prediction of mechanical properties of LPBF built part based on process monitoring and Gaussian process regression

As a highly promising technology in additive manufacturing, the laser powder bed fusion has only limited application due to its low reproducibility. In this study, the image information of the 316L specimen after laser scanning and powder paving of each layer was acquired by a complementary metal–oxide–semiconductor industrial camera. The important features were selected, extracted and quantificated by analyzing the tensile test results. Finally, combined with the laser power, the quantified features were as input of a Gaussian process regression model based on optimization algorithm of grid search to predict the tensile strength of 316L specimen. The results show that the quantized image features have a significant improvement on the regression effect, and the coefficient of determination (R 2) is improved from 63% to 90.57% compared to using only the laser power as input.

Predicting altered bone biomechanics in juvenile mice: insights from microgravity simulation, loading interventions, and Raman Spectroscopy

BackgroundMicrogravity, a condition experienced in a spatial environment, poses unique challenges to the skeletal system, particularly in juvenile organisms. This study aimed to investigate alterations in bone biomechanics of juvenile mice due to unloading – that simulates microgravity in the laboratory—and the effects of a bone-loading intervention. We compared bone compositional and mechanical properties between 21-six-week-old C57Bl/6 from a control group (wild type) and a group that underwent a tail-suspension unloading protocol to mimic microgravity (MG). The second group (MG) experienced additional in vivo loading protocol (MG + LDG) on the right hind leg, where dynamic compressive loading was applied to the right knee using a custom-built loading device.ResultsOur results show that after two weeks, we successfully induced bone alterations by (i) decreasing the energy dissipated before fracture and (ii) decreasing the yield and maximum stress. In addition, we showed that Mineral to matrix component [ν1PO4/Amide I], Carbonate to Amide [CO3/Amide I], and Crystallinity [1/FWHM(ν1PO4)] are strongly linked in physiological bone but not in microgravity even after loading intervention. While Crystallinity is very sensitive to bone deformation (strain) alterations coming from simulated microgravity, we show that Carbonate to Amide [CO3/Amide I] – a common marker of turnover rate/remodeling activity—is a specific predictor of bone deformation for bone after simulated microgravity. Our results also invalidate the current parameters of the loading intervention to prevent bone alterations entirely in juvenile mice.ConclusionsOur study successfully induced bone alterations in juvenile mice by using an unloading protocol to simulate microgravity, and we provided a new Raman Spectroscopy (RS) dataset of juvenile mice that contributes to the prediction of cortical bone mechanical properties, where the degree of interrelationship for RS data for physiological bone is improved compared to the most recent evidence.

A model predicting mechanical properties of asphalt mixtures considering void ratio and loading conditions

The void ratio is an essential parameter in asphalt mixture design, which significantly affects asphalt mixtures' mechanical properties and road performance. The prediction model for mechanical properties will be established based on the Computed tomography (CT) scanning, strength, and fatigue test results in this research to reveal the relationship between the asphalt mixtures' mechanical properties and void ratio. Firstly, the voids' transverse and longitudinal distribution characteristics in asphalt mixtures of different gradations were investigated based on CT scanning. Then, the influences of void ratio, loading rates on the strength, and the impacts of void ratio and stress levels on the fatigue life were investigated by conducting indirect tensile tests. Finally, the mechanical properties prediction model was established for void ratio and loading conditions. Further, the feasibility of the model is verified in conjunction with existing research results. The results show that the void ratio calculated based on CT scanning is about 1.15 times the measured void ratio in the water weight method. The number of voids was uniformly distributed in the range of 15 mm∼50 mm in height and 10 mm∼40 mm in radius of the specimen. The asphalt mixtures' strength decreases with increasing void ratio and increases with increasing loading rate. The asphalt mixtures' fatigue life declines with increasing void ratio and stress level. The correlation coefficients of the established strength and fatigue prediction models were 0.963 and 0.976, respectively. The relative error between the measured strength, fatigue life, and predicted results is about 10.75% and 11.27%, respectively. This research provides a basis for designing high-performance asphalt mixtures by establishing the relationship between asphalt mixtures' microscopic and mechanical characteristics.

An integrated computation framework for predicting mechanical performance of single-phase alloys manufactured using laser powder bed fusion: A case study of CoCrFeMnNi high-entropy alloy

Establishing a process-structure-property (PSP) relationship is crucial for understanding and optimizing the mechanical properties of components via laser powder bed fusion (L-PBF) additive manufacturing (AM). However, the parameters pertinent to the PSP relationship are encapsulated in a “black box” with a high dimension, resulting in relying on trial-and-error approaches to elucidate hidden associations becoming impracticable, while the acquisition of microscopic data proves financially burdensome. Thus this paper presents an integrated computation framework based on machine learning (ML) for predicting the tensile yield strengths and post-yield hardening rates of single-phase alloys fabricated through L-PBF, for efficient process optimization purposes. Research herein focuses on single-phase alloys to be built via L-PBF, taking CoCrFeMnNi high-entropy alloy as a study case, and provide an efficient integrated computation solution. The methodology proposed in this paper generates as-built microstructures associated with the varying L-PBF process parameters via cellular automata (CA). Subsequently, the mechanical responses of these microstructures are estimated by employing crystal plasticity finite element (CPFE) simulations. In addition, the approach raised in the paper leverages particle-level microstructure descriptors instead of average feature descriptors for the entire representative volume element (RVE) structure in an ML framework to predict the characteristics of the newly randomly growing grains. This research demonstrate the practicality and instructiveness of the methodology for modeling realistic grains. Then, trained ML model exhibits promising accuracy for the prediction of mechanical properties. This work showcases an integrated computation framework combining advanced materials science and ML techniques for enhancing understanding of the L-PBF process and offering valuable insights into process optimization for L-PBF-built diverse materials.