Hybrid identification method of coupled viscoplastic-damage constitutive parameters based on BP neural network and genetic algorithm

Yao, Dan; Duan, Yongchuan; Li, Muyu; Guan, Yaping

doi:10.1016/j.engfracmech.2021.108027

Cited by 31 publications

(7 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Due to the high coupled nonlinear and differential of equations in the developed unified model, a GA-based algorithm is used to optimize the material constants via minimizing the residuals of experimental and calculated values. The parameters and corresponding procedure of the GA-based algorithm are described elsewhere [25,26,58].…”

Section: Determination Of Materials Constantsmentioning

confidence: 99%

Microstructure Evolution and a Unified Constitutive Model of Ti-55511 Alloy Compressed at Stepped Strain Rates

Zhong

Lin

et al. 2021

Materials

View full text Add to dashboard Cite

The flow behavior and microstructure change of the Ti-55511 alloy are investigated by thermal compression experiments with stepped strain rates. The phase transformation features, the dynamic recrystallization (DRX) behavior of the β matrix, the dynamic spheroidization mechanism of the lamellar α phase and the evolution of the β sub-grain size are quantitatively analyzed. A unified constitutive model is constructed to characterize the hot deformation features of the Ti-55511 alloy. In the established model, the work hardening effect is taken into account by involving the coupled effects of the equiaxed and lamellar α phases, as well as β substructures. The dynamic softening mechanisms including the dynamic recovery (DRV), DRX and dynamic spheroidization mechanisms are also considered. The material parameters are optimized by the multi-objective algorithm in the MATLAB toolbox. The consistency between the predicted and experimental data indicates that the developed unified model can accurately describe the flow features and microstructure evolution of the hot compressed Ti-55511 at stepped strain rates.

show abstract

Section: Determination Of Materials Constantsmentioning

confidence: 99%

Microstructure Evolution and a Unified Constitutive Model of Ti-55511 Alloy Compressed at Stepped Strain Rates

Zhong

Lin

et al. 2021

Materials

View full text Add to dashboard Cite

show abstract

“…The findings were validated through finite element simulation results. Yao et al [38] explored a rate-dependent model integrating plasticity and damage. The authors utilized a hybrid approach, merging neural networks with genetic algorithms for parameter identification, focusing on tensile experiments.…”

Section: Introductionmentioning

confidence: 99%

Recurrent Neural Networks and Three-Point Bending Test on the Identification of Material Hardening Parameters

Cruz,

Barbosa,

Santos

et al. 2024

Metals

View full text Add to dashboard Cite

The continuous evolution of metallic alloys in the automotive industry has led to the development of more advanced and flexible constitutive models that attempt to accurately describe the various fundamental properties and behavior of these materials. These models have become increasingly complex, incorporating a larger number of parameters that require an accurate calibration procedure to fit the constitutive parameters with experimental data. In this context, machine learning (ML) methodologies have the potential to advance material constitutive modeling, enhancing the efficiency of the material parameter calibration procedure. Recurrent neural networks (RNNs) stand out among various learning algorithms due to their ability to process sequential data and overcome limitations imposed by nonlinearities and multiple parameters involved in phenomenological models. This study explores the modeling capabilities of long short-term memory (LSTM) structures, a type of RNN, in predicting the hardening behavior of a sheet metal material using the results of a standardized experimental three-point bending test, with the aim of extending this methodology to other experimental tests and constitutive models. Additionally, a variable analysis is performed to select the most important variables for this experimental test and assess the influence of friction, material thickness, and elastic and plastic properties on the accuracy of predictions made by neural networks. The required data for designing and training the network solutions are collected from numerical simulations using finite element methodology (FEM), which are subsequently validated by experiments. The results demonstrate that the proposed LSTM-based approach outperforms traditional identification techniques in predicting the material hardening parameters. This suggests that the developed procedure can be effectively applied to efficiently characterize different materials, especially those extensively used in industrial applications, ranging from mild steels to advanced high-strength steels.

show abstract

“…It was found that the constitutive parameters A and B had a much greater impact on cutting force than parameters C, m, and n during simulation. According to hot tensile test, neural networks and genetic algorithm, Yao et al 20 obtained the constitutive parameters of AA6061 aluminum alloy via reverse identification. It was reported that the accuracy of hybrid identification method was better than single optimized algorithm, since the initial population of genetic algorithm was optimized by neural network.…”

Section: Introductionmentioning

confidence: 99%

Genetic algorithm-based parameter identification of Johnson- Cook constitutive model for 304 stainless steel

Jiang,

Ding,

Wang

et al. 2024

Preprint

View full text Add to dashboard Cite

Addressing the significant discrepancy between actual experimental cutting force and its predicted values derived from traditional constitutive model parameter identification methods, a reverse identification research of the Johnson-Cook (J-C) constitutive model for 304 stainless steel was conducted via genetic algorithm. Considering actual cutting zone and the influence of feed motion on the rake (flank) angle, an unequal division shear zone model was established to implement the theoretical calculation for shear zone stress. Through cutting experiments, the spindle speed was negatively correlated with the cutting force at first, and then became positively correlated; The empirical formula (EXP model) for turning force was corrected, revealing that the EXP model was unable to provide optimal predicted values for cutting force. The influence of the J-C constitutive parameter C on the cutting morphology was firstly investigated through simulation analysis, and determined an appropriate value for C, then obtained the precise values for the other four constitutive parameters by genetic algorithm. Moreover, the simulated values of cutting force in JC1 model (obtained from the Split Hopkinson Pressure Bar test) and JCM model (the improved model using genetic algorithm) were obtained by three-dimensional (3-D) simulation via FEM software. The results indicated that, the maximum error between actual experimental cutting force and its simulated values (by JCM model) was 14.8%, with an average error of 6.38%. These results outperformed the JC1 and EXP models, suggesting that the JCM model identified via genetic algorithm was more reliable.

show abstract

Hybrid identification method of coupled viscoplastic-damage constitutive parameters based on BP neural network and genetic algorithm

Cited by 31 publications

References 41 publications

Microstructure Evolution and a Unified Constitutive Model of Ti-55511 Alloy Compressed at Stepped Strain Rates

Microstructure Evolution and a Unified Constitutive Model of Ti-55511 Alloy Compressed at Stepped Strain Rates

Recurrent Neural Networks and Three-Point Bending Test on the Identification of Material Hardening Parameters

Genetic algorithm-based parameter identification of Johnson- Cook constitutive model for 304 stainless steel

Contact Info

Product

Resources

About