• Transactions of Materials Processing
• /
• v.31 no.4
• /
• pp.229-239
• /
• 2022

Convolution Neural Network(CNN) is a class of deep learning algorithms and can be used for image analysis. In particular, it has excellent performance in finding the pattern of images. Therefore, CNN is commonly applied for recognizing, learning and classifying images. In this study, the surface defect classification performance of Al 6061 extruded material using CNN-based algorithms were compared and evaluated. First, the data collection criteria were suggested and a total of 2,024 datasets were prepared. And they were randomly classified into 1,417 learning data and 607 evaluation data. After that, the size and quality of the training data set were improved using data augmentation techniques to increase the performance of deep learning. The CNN-based algorithms used in this study were VGGNet-16, VGGNet-19, ResNet-50 and DenseNet-121. The evaluation of the defect classification performance was made by comparing the accuracy, loss, and learning speed using verification data. The DenseNet-121 algorithm showed better performance than other algorithms with an accuracy of 99.13% and a loss value of 0.037. This was due to the structural characteristics of the DenseNet model, and the information loss was reduced by acquiring information from all previous layers for image identification in this algorithm. Based on the above results, the possibility of machine vision application of CNN-based model for the surface defect classification of Al extruded materials was also discussed.

• Transactions of Materials Processing
• /
• v.31 no.5
• /
• pp.296-301
• /
• 2022

In this study, influence of the process parameters of the roller leveling process on the straightness of the steel wire was analyzed using numerical analysis. To construct the numerical analysis model, cross-sectional and longitudinal element sizes, which affect the prediction accuracy of longitudinal stress caused by bending deformation of the steel wire, were optimized, and mass scaling that satisfies prediction accuracy while reducing computational time was confirmed. By using the constructed numerical analysis model, the influence of various process parameters such as input direction of the steel wire, initial diameter of the steel wire, back tension and intermesh on the straightness was confirmed. The simulation result shows that the 3^rd and 4^th roller of vertical straightener had a significant influence on vertical shape of the steel wire.

• Steel and Composite Structures
• /
• v.44 no.2
• /
• pp.183-198
• /
• 2022

Plastic deformation of link beams in eccentrically braced frames is the primary dissipating source of seismic energy. Despite the excellent compatibility with the architectural designs, previous researches indicate the deficiency of flexural yielding links compared to the shear yielding ones because of their localized plastic deformation. Previous investigations have shown that implementing web openings in beams could be an efficient method to improve the seismic performance of moment-resisting connections. Accordingly, this research investigates the use of flexural links with stiffened and un-stiffened web openings to eliminate localized plasticity at the ends of the link. For this purpose, the numerical models are generated in finite element software "Abaqus" and verified against experimental data gathered from other studies. Models are subjected to cyclic displacement history to evaluate their behavior. Failure of the numerical models under cyclic loading is simulated using a micromechanical based damage model known as Cyclic Void Growth Model (CVGM). The elastic stiffness and the strength-based and CVGM-based inelastic rotation capacity of the links are compared to evaluate the studied models' seismic response. The results of this investigation indicate that some of the flexural links with edge stiffened web openings show increased inelastic rotation capacity compared to an un-perforated link.

• Transactions of Materials Processing
• /
• v.32 no.3
• /
• pp.114-121
• /
• 2023

During rolling, rolling mill rolls endure wear when shaping metal billets into a desired form, such as bars, plates, and shapes. Such wear affects the lifespan of the rolls and product quality. Therefore, in addition to rigidity, wear performance is a key factor influencing the performance of rolling mill rolls. Conventional methods such as casting and forging have been used to manufacture rolling mill rolls. However, powder alloying methods are increasingly being adopted to enhance wear resistance. These powder manufacturing methods include atomization, canning to shape the powder, hot isostatic pressing to combine the powder alloy with conventional metals, and various wear performance tests on rolls prepared with powder alloys. In this study, numerical simulations and experimental tests were used to develop and elucidate the wear analysis mechanism of rolling mill rolls. The wear characteristics of the rolls under various rolling conditions were analyzed. In addition, experimental tests (wear and surface analysis tests) and wear theory (Archard wear model) were used to evaluate wear. These tests were performed on two different materials in various powder states to evaluate the different aspects of wear resistance. In particular, this study identifies the factors influencing the wear behavior of rolling mill rolls and proposes an analytical approach based on the actual production of products. The developed wear analysis mechanism can serve the future development of rolls with high wear resistance using new materials. Moreover, it can be applied in the mechanical and wear performance testing of new products.

• Advances in Computational Design
• /
• v.8 no.2
• /
• pp.97-112
• /
• 2023

The Finite Element (FE) modeling of Reinforced Concrete (RC) under seismic loading has a sensitive impact in terms of getting good contribution compared to experimental results. Several idealized model types for simulating the nonlinear response have been developed based on the plasticity distribution alone the model. The Continuum Models are the most used category of modeling, to understand the seismic behavior of structural elements in terms of their components, cracking patterns, hysteretic response, and failure mechanisms. However, the material modeling, contact and nonlinear analysis strategy are highly complex due to the joint operation of concrete and steel. This paper presents a numerical simulation of a chosen RC column under monotonic and cyclic loading using the FE Abaqus, to assessthe hysteretic response and failure mechanisms in the RC columns, where the perfect bonding option is used for the contact between concrete and steel. While results of the numerical study under cyclic loading compared to experimental tests might be unsuccessful due to the lack of bond-slip modeling. The monotonic loading shows a good estimation of the envelope response and deformation components. In addition, this work further demonstrates the advantage and efficiency of the damage distributions since the obtained damage distributions fit the expected results.

• Geomechanics and Engineering
• /
• v.33 no.6
• /
• pp.625-633
• /
• 2023

The volume changes associated with moisture or suction variation in expansive soils are of geotechnical and geoenvironmental design concern. These changes can impact the performance of infrastructure projects and lightweight structures. Assessment of unsaturated function for these materials leads to better interpretation and understanding, as well as providing accurate and economic design. In this study, expansive soils from different regions of Saudi Arabia were studied for their basic properties including gradation, plasticity and shrinkage, swelling, and consolidation characteristics. The unsaturated soil functions of saturated water content, air-entry values, and residual states were determined by conducting the tests for the entire soil water characteristic curves (SWCC) using different techniques. An attempt has been made to provide a prediction model for unsaturated properties based on the basic properties of these soils. Once the profile of SWCC has been predicted the time and cost for many tests can be saved. These predictions can be utilized in practice for the application of unsaturated soil mechanics on geotechnical and geoenvironmental projects.

• Transactions of Materials Processing
• /
• v.32 no.1
• /
• pp.28-34
• /
• 2023

In this research, a representative volume element (RVE)-based FE Model is presented to estimate the mechanical properties of additively manufactured continuous fiber-reinforced composites with different fiber orientations. To construct the model, an ABAQUS Python script has been implemented to produce matrix and fiber in the desired orientations at the RVE. A script has also been developed to apply the periodic boundary conditions to the RVE. Experimental tests were conducted to validate the numerical models. Tensile specimens with the fiber directions aligned in the 0, 45, and 90 degrees to the loading direction were manufactured using a continuous fiber 3D printer and tensile tests were performed in the three directions. Tensile tests were also simulated using the RVE models. The predicted Young's moduli compared well with the measurements: the Young's modulus prediction accuracy values were 83.73, 97.70, and 92.92 percent for the specimens in the 0, 45, and 90 degrees, respectively. The proposed method with periodic boundary conditions precisely evaluated the elastic properties of additively manufactured continuous fiber-reinforced composites with complex microstructures.

• Transactions of Materials Processing
• /
• v.32 no.6
• /
• pp.329-334
• /
• 2023

The process of heat treatment in cold forging is an essential role in enhancing mechanical properties. However, it relies heavily on the experience and skill of individuals. The aim of this study is to predict hardness using machine learning to optimize production efficiency in cold forging manufacturing. Random Forest (RF), Gradient Boosting Regressor (GBR), Extra Trees (ET), and ADAboosting (ADA) models were utilized. In the result, the RF, GBR, and ET models show the excellent performance. However, it was observed that GBR and ET models leaned significantly towards the influence of temperature, unlike the RF model. We suggest that RF model demonstrates greater reliability in predicting hardness due to its ability to consider various variables that occur during the cold forging process.

• Transactions of Materials Processing
• /
• v.33 no.1
• /
• pp.5-11
• /
• 2024

The development of a next-generation hydrogen compressor, a key component in the expansion of hydrogen charging infrastructure, is in progress. In order to improve compression efficiency and durability, it is important to optimize the precision forming and shearing processes of the diaphragm, which is the bellows unit cell, as well as the optimization of diaphragm shape itself. In this study, we aim to show that die and process design technology that can synchronize the inner and outer shearing points of the diaphragm for the precision forming of product can be constructed based on a numerical simulation. First, the damage model that can predict the fracture points will be determined using the shear load and shear zone measurements obtained by performing a blanking test of AISI-633 stainless steel. Next, we will explain the overall procedure based on numerical analysis model how to determine the shearing points according to the deformation pattern of urethane die for various shearing die design.

• Transactions of Materials Processing
• /
• v.33 no.2
• /
• pp.96-102
• /
• 2024

This study addresses evaluation of performance of hardening model for a titanium alloy (Ti6Al4V) based on the artificial neural network (ANN) regarding the strain rate and the temperature. Uniaxial compression tests were carried out at different strain rates from 0.001 /s to 10 /s and temperatures from 575 ℃ To 975 ℃. Using the experimental data, ANN models were trained and tested with different hyperparameters, such as size of hidden layer and optimizer. The input features were determined with the equivalent plastic strain, strain rate, and temperature while the output value was set to the equivalent stress. When the number of data is sufficient with a smooth tendency, both the Bayesian regulation (BR) and the Levenberg-Marquardt (LM) show good performance to predict the flow behavior. However, only BR algorithm shows a predictability when the number of data is insufficient. Furthermore, a proper size of the hidden layer must be confirmed to describe the behavior with the limited number of the data.