Multi-Constitutive Neural Network for Large Deformation Poromechanics Problem

Qi Zhang; Yilin Chen; Ziyi Yang; Eric Darve

Multi-Constitutive Neural Network for Large Deformation Poromechanics Problem

Qi Zhang, Yilin Chen, Ziyi Yang, Eric Darve

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

Abstract

In this paper, we study the problem of large-strain consolidation in poromechanics with deep neural networks (DNN). Given different material properties and different loading conditions, the goal is to predict pore pressure and settlement. We propose a novel method *multi-constitutive neural network* (MCNN) such that one model can solve several different constitutive laws. We introduce a one-hot encoding vector as an additional input vector, which is used to label the constitutive law we wish to solve. Then we build a DNN which takes $(\hat{X}, \hat{t})$ as input along with a constitutive law label and outputs the corresponding solution. It is the first time, to our knowledge, that we can evaluate multi-constitutive laws through only one training process while still obtaining good accuracies. We found that MCNN trained to solve multiple PDEs outperforms individual neural network solvers trained with PDE in some cases.

Original language	English
Title of host publication	Multi-Constitutive Neural Network for Large Deformation Poromechanics Problem
Publisher	Proceedings of the Machine Learning and the Physical Sciences Workshop, 34th Conference on Neural Information Processing Systems (NeurIPS)
Number of pages	6
Publication status	Published - 11 Dec 2020
Externally published	Yes

More information

https://arxiv.org/abs/2010.15549

Cite this

Zhang, Q., Chen, Y., Yang, Z., & Darve, E. (2020). Multi-Constitutive Neural Network for Large Deformation Poromechanics Problem. In Multi-Constitutive Neural Network for Large Deformation Poromechanics Problem Proceedings of the Machine Learning and the Physical Sciences Workshop, 34th Conference on Neural Information Processing Systems (NeurIPS). https://arxiv.org/abs/2010.15549

@inproceedings{9d69563e8b3c49c2bfd1ede457538b72,

title = "Multi-Constitutive Neural Network for Large Deformation Poromechanics Problem",

abstract = "In this paper, we study the problem of large-strain consolidation in poromechanics with deep neural networks (DNN). Given different material properties and different loading conditions, the goal is to predict pore pressure and settlement. We propose a novel method *multi-constitutive neural network* (MCNN) such that one model can solve several different constitutive laws. We introduce a one-hot encoding vector as an additional input vector, which is used to label the constitutive law we wish to solve. Then we build a DNN which takes $(\hat{X}, \hat{t})$ as input along with a constitutive law label and outputs the corresponding solution. It is the first time, to our knowledge, that we can evaluate multi-constitutive laws through only one training process while still obtaining good accuracies. We found that MCNN trained to solve multiple PDEs outperforms individual neural network solvers trained with PDE in some cases.",

author = "Qi Zhang and Yilin Chen and Ziyi Yang and Eric Darve",

year = "2020",

month = dec,

day = "11",

language = "English",

booktitle = "Multi-Constitutive Neural Network for Large Deformation Poromechanics Problem",

publisher = "Proceedings of the Machine Learning and the Physical Sciences Workshop, 34th Conference on Neural Information Processing Systems (NeurIPS)",

}

Zhang, Q, Chen, Y, Yang, Z & Darve, E 2020, Multi-Constitutive Neural Network for Large Deformation Poromechanics Problem. in Multi-Constitutive Neural Network for Large Deformation Poromechanics Problem. Proceedings of the Machine Learning and the Physical Sciences Workshop, 34th Conference on Neural Information Processing Systems (NeurIPS). <https://arxiv.org/abs/2010.15549>

Multi-Constitutive Neural Network for Large Deformation Poromechanics Problem. / Zhang, Qi; Chen, Yilin; Yang, Ziyi et al.
Multi-Constitutive Neural Network for Large Deformation Poromechanics Problem. Proceedings of the Machine Learning and the Physical Sciences Workshop, 34th Conference on Neural Information Processing Systems (NeurIPS), 2020.

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

TY - GEN

T1 - Multi-Constitutive Neural Network for Large Deformation Poromechanics Problem

AU - Zhang, Qi

AU - Chen, Yilin

AU - Yang, Ziyi

AU - Darve, Eric

PY - 2020/12/11

Y1 - 2020/12/11

N2 - In this paper, we study the problem of large-strain consolidation in poromechanics with deep neural networks (DNN). Given different material properties and different loading conditions, the goal is to predict pore pressure and settlement. We propose a novel method *multi-constitutive neural network* (MCNN) such that one model can solve several different constitutive laws. We introduce a one-hot encoding vector as an additional input vector, which is used to label the constitutive law we wish to solve. Then we build a DNN which takes $(\hat{X}, \hat{t})$ as input along with a constitutive law label and outputs the corresponding solution. It is the first time, to our knowledge, that we can evaluate multi-constitutive laws through only one training process while still obtaining good accuracies. We found that MCNN trained to solve multiple PDEs outperforms individual neural network solvers trained with PDE in some cases.

AB - In this paper, we study the problem of large-strain consolidation in poromechanics with deep neural networks (DNN). Given different material properties and different loading conditions, the goal is to predict pore pressure and settlement. We propose a novel method *multi-constitutive neural network* (MCNN) such that one model can solve several different constitutive laws. We introduce a one-hot encoding vector as an additional input vector, which is used to label the constitutive law we wish to solve. Then we build a DNN which takes $(\hat{X}, \hat{t})$ as input along with a constitutive law label and outputs the corresponding solution. It is the first time, to our knowledge, that we can evaluate multi-constitutive laws through only one training process while still obtaining good accuracies. We found that MCNN trained to solve multiple PDEs outperforms individual neural network solvers trained with PDE in some cases.

UR - https://ml4physicalsciences.github.io/2020/

M3 - Conference article published in proceeding or book

BT - Multi-Constitutive Neural Network for Large Deformation Poromechanics Problem

PB - Proceedings of the Machine Learning and the Physical Sciences Workshop, 34th Conference on Neural Information Processing Systems (NeurIPS)

ER -

Multi-Constitutive Neural Network for Large Deformation Poromechanics Problem

Abstract

More information

Other files and links

Fingerprint

Cite this