Stochastic deep collocation method based on neural architecture search and transfer learning for heterogeneous porous media

authored by

HW Guo, XY Zhuang, Pengwan Chen, N Alajlan, T Rabczuk

Abstract

We present a stochastic deep collocation method (DCM) based on neural architecture search (NAS) and transfer learning for heterogeneous porous media. We first carry out a sensitivity analysis to determine the key hyper-parameters of the network to reduce the search space and subsequently employ hyper-parameter optimization to finally obtain the parameter values. The presented NAS based DCM also saves the weights and biases of the most favorable architectures, which is then used in the fine-tuning process. We also employ transfer learning techniques to drastically reduce the computational cost. The presented DCM is then applied to the stochastic analysis of heterogeneous porous material. Therefore, a three dimensional stochastic flow model is built providing a benchmark to the simulation of groundwater flow in highly heterogeneous aquifers. The performance of the presented NAS based DCM is verified in different dimensions using the method of manufactured solutions. We show that it significantly outperforms finite difference methods in both accuracy and computational cost.

Organisation(s)

Institute of Photonics

External Organisation(s)

Tongji University
Beijing Institute of Technology
King Saud University

Type

Article

Journal

Engineering with computers

Volume

38

Pages

5173-5198

No. of pages

26

ISSN

0177-0667

Publication date

12.2022

Publication status

Published

Peer reviewed

Yes

ASJC Scopus subject areas

Software, Engineering(all), Computer Science Applications, Modelling and Simulation

Electronic version(s)

https://doi.org/10.1007/s00366-021-01586-2 (Access: Open)