Deviatoric stress-strain curve construction with strain-softening account via the damage-modified Duncan-Chang, arctangent and informer models: a comparative analysis
Engineering construction in cold regions cannot be separated from permafrost research. This study aimed to determine the mechanical properties and changing laws of artificially frozen clay through triaxial tests. Two models have been established: a physical model based on the tradi-tional phenomenol...
Published in: | Frontiers in Earth Science |
---|---|
Main Authors: | , , , , , |
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
Frontiers Media S.A.
2023
|
Subjects: | |
Online Access: | https://doi.org/10.3389/feart.2023.1238286 https://doaj.org/article/4ac5b9f2cd1347fbbdbd50b79c3027ea |
id |
ftdoajarticles:oai:doaj.org/article:4ac5b9f2cd1347fbbdbd50b79c3027ea |
---|---|
record_format |
openpolar |
spelling |
ftdoajarticles:oai:doaj.org/article:4ac5b9f2cd1347fbbdbd50b79c3027ea 2023-08-27T04:11:31+02:00 Deviatoric stress-strain curve construction with strain-softening account via the damage-modified Duncan-Chang, arctangent and informer models: a comparative analysis Zhen-Chao Teng Ya-Dong Zhou Yun-Chao Teng Xiao-Yan Liu Jia-Lin Liu Bo Li 2023-08-01T00:00:00Z https://doi.org/10.3389/feart.2023.1238286 https://doaj.org/article/4ac5b9f2cd1347fbbdbd50b79c3027ea EN eng Frontiers Media S.A. https://www.frontiersin.org/articles/10.3389/feart.2023.1238286/full https://doaj.org/toc/2296-6463 2296-6463 doi:10.3389/feart.2023.1238286 https://doaj.org/article/4ac5b9f2cd1347fbbdbd50b79c3027ea Frontiers in Earth Science, Vol 11 (2023) triaxial test of frozen soil deep learning constitutive equation of frozen soil weibull distribution strain-softening Duncan-Chang model Science Q article 2023 ftdoajarticles https://doi.org/10.3389/feart.2023.1238286 2023-08-06T00:35:19Z Engineering construction in cold regions cannot be separated from permafrost research. This study aimed to determine the mechanical properties and changing laws of artificially frozen clay through triaxial tests. Two models have been established: a physical model based on the tradi-tional phenomenological constitutive theory and a deep learning model based on the data-driven constitutive theory, taking into account the softening phenomenon. The accuracy and applica-bility of the models were verified, followed by a comparative analysis. The results of the analysis are as follows. The Duncan-Chang model can describe the characteristics of the hardening-type deviatoric stress-strain curve, but it cannot describe the characteristics of the softening-type de-viatoric stress-strain curve. The Modified Duncan-Chang (MDC) model fails to accurately de-scribe the characteristics of a smooth deviatoric stress-strain curve. The Strain-Damage Modified Duncan-Chang (SD-MDC) model exhibits a good fit in both the ascending and descending seg-ments of the curve, but it lacks effectiveness in the convergence segment of the S-shaped sof-tening curve. For this reason, this paper has chosen the arctangent function to establish a Strain-Damage Modified arctangent constitutive model (SD-MAM). This model accurately re-flects the stress evolution process of different types of frozen soils. Additionally, the Informer time series prediction algorithm was utilized to develop the Informer permafrost deviatoric stress prediction model which achieved an R2 value above 99%. In comparison to the SD-MAM model, the Informer model demonstrates higher precision, does not rely on assumptions, is cost-effective, and has a wide range of applications. However, it lacks physical meaning, and interpretability, and requires further discussion regarding the reliability of the results. This study offers valuable insights into the development and application of constitutive models for frozen soils. Article in Journal/Newspaper permafrost Directory of Open Access Journals: DOAJ Articles Frontiers in Earth Science 11 |
institution |
Open Polar |
collection |
Directory of Open Access Journals: DOAJ Articles |
op_collection_id |
ftdoajarticles |
language |
English |
topic |
triaxial test of frozen soil deep learning constitutive equation of frozen soil weibull distribution strain-softening Duncan-Chang model Science Q |
spellingShingle |
triaxial test of frozen soil deep learning constitutive equation of frozen soil weibull distribution strain-softening Duncan-Chang model Science Q Zhen-Chao Teng Ya-Dong Zhou Yun-Chao Teng Xiao-Yan Liu Jia-Lin Liu Bo Li Deviatoric stress-strain curve construction with strain-softening account via the damage-modified Duncan-Chang, arctangent and informer models: a comparative analysis |
topic_facet |
triaxial test of frozen soil deep learning constitutive equation of frozen soil weibull distribution strain-softening Duncan-Chang model Science Q |
description |
Engineering construction in cold regions cannot be separated from permafrost research. This study aimed to determine the mechanical properties and changing laws of artificially frozen clay through triaxial tests. Two models have been established: a physical model based on the tradi-tional phenomenological constitutive theory and a deep learning model based on the data-driven constitutive theory, taking into account the softening phenomenon. The accuracy and applica-bility of the models were verified, followed by a comparative analysis. The results of the analysis are as follows. The Duncan-Chang model can describe the characteristics of the hardening-type deviatoric stress-strain curve, but it cannot describe the characteristics of the softening-type de-viatoric stress-strain curve. The Modified Duncan-Chang (MDC) model fails to accurately de-scribe the characteristics of a smooth deviatoric stress-strain curve. The Strain-Damage Modified Duncan-Chang (SD-MDC) model exhibits a good fit in both the ascending and descending seg-ments of the curve, but it lacks effectiveness in the convergence segment of the S-shaped sof-tening curve. For this reason, this paper has chosen the arctangent function to establish a Strain-Damage Modified arctangent constitutive model (SD-MAM). This model accurately re-flects the stress evolution process of different types of frozen soils. Additionally, the Informer time series prediction algorithm was utilized to develop the Informer permafrost deviatoric stress prediction model which achieved an R2 value above 99%. In comparison to the SD-MAM model, the Informer model demonstrates higher precision, does not rely on assumptions, is cost-effective, and has a wide range of applications. However, it lacks physical meaning, and interpretability, and requires further discussion regarding the reliability of the results. This study offers valuable insights into the development and application of constitutive models for frozen soils. |
format |
Article in Journal/Newspaper |
author |
Zhen-Chao Teng Ya-Dong Zhou Yun-Chao Teng Xiao-Yan Liu Jia-Lin Liu Bo Li |
author_facet |
Zhen-Chao Teng Ya-Dong Zhou Yun-Chao Teng Xiao-Yan Liu Jia-Lin Liu Bo Li |
author_sort |
Zhen-Chao Teng |
title |
Deviatoric stress-strain curve construction with strain-softening account via the damage-modified Duncan-Chang, arctangent and informer models: a comparative analysis |
title_short |
Deviatoric stress-strain curve construction with strain-softening account via the damage-modified Duncan-Chang, arctangent and informer models: a comparative analysis |
title_full |
Deviatoric stress-strain curve construction with strain-softening account via the damage-modified Duncan-Chang, arctangent and informer models: a comparative analysis |
title_fullStr |
Deviatoric stress-strain curve construction with strain-softening account via the damage-modified Duncan-Chang, arctangent and informer models: a comparative analysis |
title_full_unstemmed |
Deviatoric stress-strain curve construction with strain-softening account via the damage-modified Duncan-Chang, arctangent and informer models: a comparative analysis |
title_sort |
deviatoric stress-strain curve construction with strain-softening account via the damage-modified duncan-chang, arctangent and informer models: a comparative analysis |
publisher |
Frontiers Media S.A. |
publishDate |
2023 |
url |
https://doi.org/10.3389/feart.2023.1238286 https://doaj.org/article/4ac5b9f2cd1347fbbdbd50b79c3027ea |
genre |
permafrost |
genre_facet |
permafrost |
op_source |
Frontiers in Earth Science, Vol 11 (2023) |
op_relation |
https://www.frontiersin.org/articles/10.3389/feart.2023.1238286/full https://doaj.org/toc/2296-6463 2296-6463 doi:10.3389/feart.2023.1238286 https://doaj.org/article/4ac5b9f2cd1347fbbdbd50b79c3027ea |
op_doi |
https://doi.org/10.3389/feart.2023.1238286 |
container_title |
Frontiers in Earth Science |
container_volume |
11 |
_version_ |
1775354392556339200 |