Search results

The aim of this paper is to develop a suitable artificial neural network (ANN) model that fits best in predicting the experimental flow stress values to the closet proximity for mechanically alloyed Al6063/0.75Al2O3/0.75Y2O3 hybrid nanocomposite.

The ANN model is implemented on neural network toolbox of MATLAB^® using feed‐forward back propagation network and logsig functions. A set of 80 training data and 20 testing data were used in the ANN model. The layout of the network is arranged with three input parameters that include temperature, strain and strain rate, one hidden layer with 22 neurons and one output parameter consisting of flow stress. Flow stress was also predicted using Arrhenius constitutive model.

Based on the comparison of the predicted results using ANN model and Arrhenius constitutive model, it was observed that the ANN model has higher accuracy and could be used to estimate the flow stress values during hot deformation of Al6063/0.75Al2O3/0.75Y2O3 hybrid nanocomposite.

The ANN trained with feed forward back propagation algorithm developed, presents the excellent performance of flow stress prediction of Al6063/0.75Al2O3/0.75Y2O3 hybrid nanocomposite with minimum error rates.

Accurate prediction of residual stress requires precise knowledge of the constitutive behavior of as-quenched material. This study aims to model the flow stress behavior for as-quenched Al-Mg-Si alloy.

In the present work, the flow behavior of as-quenched Al-Mg-Si alloy is studied by the hot compression tests at various temperatures (573–723 K), strain rates (0.1–1 s⁻¹) and cooling rates (1–10 K/s). Flow stress behavior is then experimentally observed, and an Arrhenius model is used to predict the flow behavior. However, due to the fact that materials parameters and activation energy do not remain constant, the Arrhenius model has an unsatisfied prediction for the flow behavior. Considering the effects of temperatures, strain rates and cooling rates on constitutive behavior, a revised Arrhenius model is developed to describe the flow stress behavior.

The experimental results show that the flow stress increases by the increasing cooling rate, increasing strain state and decreasing temperature. In comparison to the experimental data, the revised Arrhenius model has an excellent prediction for as-quenched Al-Mg-Si alloy.

With the revised Arrhenius model, the flow behaviors at different quenching conditions can be obtained, which is an essential step to the residual stress prediction when the model is implemented in a finite element code, e.g. ABAQUS, in the future.

The purpose of this paper is to propose a progressive inverse identification algorithm to characterize flow stress of tubular materials from the material response, independent of choosing an a priori hardening constitutive model.

In contrast to the conventional forward flow stress identification methods, the flow stress is characterized by a multi-linear curve rather than a limited number of hardening model parameters. The proposed algorithm optimizes the slopes and lengths of the curve increments simultaneously. The objective of the optimization is that the finite element (FE) simulation response of the test estimates the material response within a predefined accuracy.

The authors employ the algorithm to identify flow stress of a 304 stainless steel tube in a tube bulge test as an example to illustrate application of the algorithm. Comparing response of the FE simulation using the obtained flow stress with the material response shows that the method can accurately determine the flow stress of the tube.

The obtained flow stress can be employed for more accurate FE simulation of the metal forming processes as the material behaviour can be characterized in a similar state of stress as the target metal forming process. Moreover, since there is no need for a priori choosing the hardening model, there is no risk for choosing an improper hardening model, which in turn facilitates solving the inverse problem.

The proposed algorithm is more efficient than the conventional inverse flow stress identification methods. In the latter, each attempt to select a more accurate hardening model, if it is available, result in constructing an entirely new inverse problem. However, this problem is avoided in the proposed algorithm.

The dynamic recrystallization (DRX) and flow stress of Nb‐bearing steels were investigated by means of isothermal single compression testing at temperatures of 850‐105° and at constant strain rate from 0.1 to 20s‐1 using a Gleeble 3800 thermo‐mechanical simulator in order to model the DRX processes and predict the flow stress during plate rolling. On the basis of the measured flow stress, a new model of DRX kinetics was proposed to calculate the volume fraction of dynamically recrystallized grains, which was a function of processing parameters such as deformation temperature, strain, strain rate, the initial austenite grain size and Nb content. The effect of deformation conditions was quantified by the Zener‐Hollomon parameter, in which the activation energy of deformation was expressed as a power function of Nb content. The critical strain was determined by using the method proposed by Jonas and co‐workers. It is shown that the ratio of the critical strain to the peak strain decreases with increasing Nb content, from which an empirical equation was developed. In addition, the influence of Nb content and deformation conditions on the steady state grain size was determined by fitting the experimental results to a linear relationship. Finally, the flow stress of Nb bearing steels was accurately predicted using a one‐internal‐variable evolution equation by taking Nb content as a parameter and including the influence of DRX. The comparison between the experimental and theoretical results confirmed that the modeling had a good accuracy to predict flow stresses during hot deformation.

The flow phenomenon of particle image velocimetry has revealed the transition process of the complex multi-scale vortex between the boundary layer and mainstream region. Nonetheless, present computational fluid dynamics methods inadequately distinguish the discernable flows in detail. A multi-physical field coupling model, which was applied in rotor-stator fluid machinery (Umavathi, 2015; Syawitri et al., 2020), was put forward to ensure the identification of multi-scale vortexes and the improvement of performance prediction in torque converter.

A newly-developed multi-physical field simulation framework that coupled the scale-resolving simulation method with a dynamic modified viscosity coefficient was proposed to comparatively investigate the influence of energy exchange on thermal and flow characteristics and the description of the flow field in detail.

Regardless of whether quantitative or qualitative, its description ability on turbulence statistics, pressure-streamline, vortex structure and eddy viscosity ratio were visually experimentally and numerically analyzed. The results revealed that the modification of transmission medium viscous can identify flows more exactly between the viscous sublayer and outer boundary layer. Compared with RANS and large eddy simulation, a stress-blended eddy simulation model with a dynamic modified viscosity coefficient, which was further used to achieve blending on the stress level, can effectively solve the calculating problem of the transition region between the near-wall boundary layer and mainstream region.

This indeed provides an excellent description of the transient flow field and vortex structure in different physical flow states. Furthermore, the experimental data has proven that the maximum error of the external performance prediction was less than 4%.

An improved model was applied to simulate and analyze the flow mechanism through the evolution of vortex structures in a working chamber, to deepen the designer with a fundamental understanding on how to reduce flow losses and flow non-uniformity in manufacturing.

With the discussion on the linear relationship of determined material parameters, this study aims to propose a new method to analyze the deformation mechanism.

A modified constitutive model based on the hyperbolic sine Arrhenius equation has been established, which is applied to describe the flow behavior of Ti-6Al-4V alloy during the superplastic forming (SPF).

The modified constitutive model in this work has a good ability to describe the flow behavior for Ti-6Al-4V in SPF. Besides, a deformation map of titanium material is obtained based on the parameters. As the supplement, finite element models of high-temperature tensile tests are carried out as the application of the constitutive model.

The relationship between constitutive model parameters and forming mechanism is established, which is a new angle in rheological behavior research and constitutive model analysis.

The purpose of this paper is to systematically investigate the fluid lag phenomena and its influence in the hydraulic fracturing process, including all stages of fluid-lag evolution, the transition between different stages and their coupling with dynamic fracture propagation under common conditions.

A plane 2D model is developed to simulate the complex evolution of fluid lag during the propagation of a hydraulic fracture driven by an impressible Newtonian fluid. Based on the finite element method, a fully implicit solution scheme is proposed to solve the strongly coupled rock deformation, fluid flow and fracture propagation. Using the proposed model, comprehensive parametric studies are performed to examine the evolution of fluid lag in various geological and operational conditions.

The numerical simulations predict that the lag ratio is around 5% or even lower at the beginning stage of hydraulic fracture under practical geological conditions. With the fracture propagation, the lag ratio keeps decreasing and can be ignored in the late stage of hydraulic fracturing for typical parameter combinations. On the numerical aspect, whether the fluid lag can be ignored depends not only on the lag ratio but also on the minimum mesh size used for fluid flow. In addition, an overall mixed-mode fracture propagation factor is proposed to describe the relationship between diverse parameters and fracture curvature.

In this study, relatively simple physical models such as linear elasticity for solid, Newtonian model for fluid and linear elasticity fracture mechanics for fracture are used. The current model does not account for such effects like leak off, poroelasticity and softening of rock formations, which may also visibly affect the fluid lag depending on specific reservoir conditions.

This study helps to understand the effect of fluid lag during hydraulic fracturing processes and provides numerical experience in dealing with the fluid lag with finite element simulation.

The knowledge about the heat transfer and flow field in the ribbed internal passage is particularly important in industrial and engineering applications. The purpose of this paper is to identify and analyze the performance of the constrained large-eddy simulation (CLES) method in predicting the fully developed turbulent flow and heat transfer in a stationary periodic square duct with two-side ribbed walls.

The rib height-to-duct hydraulic diameter ratio is 0.1 and the rib pitch-to-height ratio is 9. The bulk Reynolds number is set to 30,000, and the bulk Mach number of the flow is chosen as 0.1 in order to keep the flow almost incompressible. The CLES calculated results are thoroughly assessed in comparison with the detached-eddy simulation (DES) and traditional large-eddy simulation (LES) methods in the light of the experimentally measured data.

It is manifested that the CLES approach can predict both aerodynamic and thermodynamic quantities more accurately than the DES and traditional LES methods.

This is the first time for the CLES method to be applied to simulation of heat and fluid flow in this widely used geometry.

This paper aims to improve performance prediction and to acquire more detailed flow structures so as to analyze the turbulence in complex rotor-stator flow.

Hydraulic retarder as typical fluid machinery was numerically investigated by using hybrid Reynolds-averaged Navier–Stokes (RANS)/large eddy simulation (LES) models CIDDES Algebraic Wall-Modeled Large Eddy Simulation (LES) (WMLES) S-Ω and dynamic hybrid RANS/LES (DHRL). The prediction results were compared and analyzed with a RANS model shear stress transport (SST) k-omega which was a recommended choice in engineering.

The numerical results were verified by experiment and indicated that the predicted values for three hybrid turbulence models were more accurate. Then, the transient flow field was further analyzed visually in terms of turbulence statistics, Reynolds number, pressure-streamline, vortex structure and eddy viscosity ratio. The results indicated that HRL approaches could capture unsteady flow phenomena.

This study achieves both in performance prediction improvement and better flow mechanism understanding. The computational fluid dynamics (CFD) could be used instead of flow visualization to a certain extent. The improved CFD method, the fine computational grid and the reasonable simulation settings jointly enhance the application of CFD in the rotor-stator flow.

The improvement was quite encouraging compared with the reported literatures, contributing to the CFD playing a more important role in the flow machinery. DHRL provided the detailed explanation of flow transport between rotor and stator, which was not reported before. Through it, the flow mechanism can be better understood.

The purpose of this paper is to build a flow stress model and microstructure evolution models which can be used to fulfill the multi-physics prediction of hot forging process, in this way the process design can be virtually verified and optimized. This is especially crucial for micro-alloyed steel forging which microstructure determines the component properties, since the downstream quenching is usually not needed.

First, hot compression tests have been completed; second, experimental data are used to build the flow stress model and models for microstructure evolution; third, programming has been finished to integrate the proposed models into the commercial finite element method (FEM) code; fourth, case study is conducted to simulate multi-stage hot forging process of micro-alloyed steel F38MnV piston; and fifth, simulation results are validated by experiment.

First, simulation results in grain size and phase volume fraction are in well agreement with experimental ones; second, the austenite grain is dramatically refined by the dynamic recrystallization in pre-forging process and static recrystallization in the two intervals has no obvious change during the following final forging and cooling above the Ae3 temperature; third, during the cooling process below the Ae3 temperature, ferrite and pearlite transformation begin from the thin skirt to the thick skirt and piston bottom because of different cooling speeds at different areas.

First, flow stress model, dynamic recrystallization model, static recrystallization model, austenite grain growth model and phase transformation models are established for a micro-alloyed steel; second, the multi-physics FEM simulation of multi-stage hot forging of industrial piston has been conducted and verified by experiment, which show good agreement.