@article{GuoZhuangChenetal., author = {Guo, Hongwei and Zhuang, Xiaoying and Chen, Pengwan and Alajlan, Naif and Rabczuk, Timon}, title = {Analysis of three-dimensional potential problems in non-homogeneous media with physics-informed deep collocation method using material transfer learning and sensitivity analysis}, series = {Engineering with Computers}, volume = {2022}, journal = {Engineering with Computers}, doi = {10.1007/s00366-022-01633-6}, url = {http://nbn-resolving.de/urn:nbn:de:gbv:wim2-20220811-46764}, pages = {1 -- 22}, abstract = {In this work, we present a deep collocation method (DCM) for three-dimensional potential problems in non-homogeneous media. This approach utilizes a physics-informed neural network with material transfer learning reducing the solution of the non-homogeneous partial differential equations to an optimization problem. We tested different configurations of the physics-informed neural network including smooth activation functions, sampling methods for collocation points generation and combined optimizers. A material transfer learning technique is utilized for non-homogeneous media with different material gradations and parameters, which enhance the generality and robustness of the proposed method. In order to identify the most influential parameters of the network configuration, we carried out a global sensitivity analysis. Finally, we provide a convergence proof of our DCM. The approach is validated through several benchmark problems, also testing different material variations.}, subject = {Deep learning}, language = {en} } @article{ChakrabortyAnitescuZhuangetal., author = {Chakraborty, Ayan and Anitescu, Cosmin and Zhuang, Xiaoying and Rabczuk, Timon}, title = {Domain adaptation based transfer learning approach for solving PDEs on complex geometries}, series = {Engineering with Computers}, volume = {2022}, journal = {Engineering with Computers}, doi = {10.1007/s00366-022-01661-2}, url = {http://nbn-resolving.de/urn:nbn:de:gbv:wim2-20220811-46776}, pages = {1 -- 20}, abstract = {In machine learning, if the training data is independently and identically distributed as the test data then a trained model can make an accurate predictions for new samples of data. Conventional machine learning has a strong dependence on massive amounts of training data which are domain specific to understand their latent patterns. In contrast, Domain adaptation and Transfer learning methods are sub-fields within machine learning that are concerned with solving the inescapable problem of insufficient training data by relaxing the domain dependence hypothesis. In this contribution, this issue has been addressed and by making a novel combination of both the methods we develop a computationally efficient and practical algorithm to solve boundary value problems based on nonlinear partial differential equations. We adopt a meshfree analysis framework to integrate the prevailing geometric modelling techniques based on NURBS and present an enhanced deep collocation approach that also plays an important role in the accuracy of solutions. We start with a brief introduction on how these methods expand upon this framework. We observe an excellent agreement between these methods and have shown that how fine-tuning a pre-trained network to a specialized domain may lead to an outstanding performance compare to the existing ones. As proof of concept, we illustrate the performance of our proposed model on several benchmark problems.}, subject = {Maschinelles Lernen}, language = {en} } @article{RabczukGuoZhuangetal., author = {Rabczuk, Timon and Guo, Hongwei and Zhuang, Xiaoying and Chen, Pengwan and Alajlan, Naif}, title = {Stochastic deep collocation method based on neural architecture search and transfer learning for heterogeneous porous media}, series = {Engineering with Computers}, volume = {2022}, journal = {Engineering with Computers}, publisher = {Springer}, address = {London}, doi = {10.1007/s00366-021-01586-2}, url = {http://nbn-resolving.de/urn:nbn:de:gbv:wim2-20220209-45835}, pages = {1 -- 26}, 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.}, subject = {Maschinelles Lernen}, language = {en} } @article{RenZhuangOterkusetal., author = {Ren, Huilong and Zhuang, Xiaoying and Oterkus, Erkan and Zhu, Hehua and Rabczuk, Timon}, title = {Nonlocal strong forms of thin plate, gradient elasticity, magneto-electro-elasticity and phase-field fracture by nonlocal operator method}, series = {Engineering with Computers}, volume = {2021}, journal = {Engineering with Computers}, doi = {10.1007/s00366-021-01502-8}, url = {http://nbn-resolving.de/urn:nbn:de:gbv:wim2-20211207-45388}, pages = {1 -- 22}, abstract = {The derivation of nonlocal strong forms for many physical problems remains cumbersome in traditional methods. In this paper, we apply the variational principle/weighted residual method based on nonlocal operator method for the derivation of nonlocal forms for elasticity, thin plate, gradient elasticity, electro-magneto-elasticity and phase-field fracture method. The nonlocal governing equations are expressed as an integral form on support and dual-support. The first example shows that the nonlocal elasticity has the same form as dual-horizon non-ordinary state-based peridynamics. The derivation is simple and general and it can convert efficiently many local physical models into their corresponding nonlocal forms. In addition, a criterion based on the instability of the nonlocal gradient is proposed for the fracture modelling in linear elasticity. Several numerical examples are presented to validate nonlocal elasticity and the nonlocal thin plate.}, subject = {Bruchmechanik}, language = {en} } @article{NooriMortazaviKeshtkarietal., author = {Noori, Hamidreza and Mortazavi, Bohayra and Keshtkari, Leila and Zhuang, Xiaoying and Rabczuk, Timon}, title = {Nanopore creation in MoS2 and graphene monolayers by nanoparticles impact: a reactive molecular dynamics study}, series = {Applied Physics A}, volume = {2021}, journal = {Applied Physics A}, number = {volume 127, article 541}, publisher = {Springer}, address = {Heidelberg}, doi = {10.1007/s00339-021-04693-5}, url = {http://nbn-resolving.de/urn:nbn:de:gbv:wim2-20210804-44756}, pages = {1 -- 13}, abstract = {In this work, extensive reactive molecular dynamics simulations are conducted to analyze the nanopore creation by nano-particles impact over single-layer molybdenum disulfide (MoS2) with 1T and 2H phases. We also compare the results with graphene monolayer. In our simulations, nanosheets are exposed to a spherical rigid carbon projectile with high initial velocities ranging from 2 to 23 km/s. Results for three different structures are compared to examine the most critical factors in the perforation and resistance force during the impact. To analyze the perforation and impact resistance, kinetic energy and displacement time history of the projectile as well as perforation resistance force of the projectile are investigated. Interestingly, although the elasticity module and tensile strength of the graphene are by almost five times higher than those of MoS2, the results demonstrate that 1T and 2H-MoS2 phases are more resistive to the impact loading and perforation than graphene. For the MoS2nanosheets, we realize that the 2H phase is more resistant to impact loading than the 1T counterpart. Our reactive molecular dynamics results highlight that in addition to the strength and toughness, atomic structure is another crucial factor that can contribute substantially to impact resistance of 2D materials. The obtained results can be useful to guide the experimental setups for the nanopore creation in MoS2or other 2D lattices.}, subject = {Nanomechanik}, language = {en} } @article{VuBacRafieeZhuangetal., author = {Vu-Bac, N. and Rafiee, Roham and Zhuang, Xiaoying and Lahmer, Tom and Rabczuk, Timon}, title = {Uncertainty quantification for multiscale modeling of polymer nanocomposites with correlated parameters}, series = {Composites Part B: Engineering}, journal = {Composites Part B: Engineering}, pages = {446 -- 464}, abstract = {Uncertainty quantification for multiscale modeling of polymer nanocomposites with correlated parameters}, subject = {Angewandte Mathematik}, language = {en} } @article{VuBacLahmerZhuangetal., author = {Vu-Bac, N. and Lahmer, Tom and Zhuang, Xiaoying and Nguyen-Thoi, T. and Rabczuk, Timon}, title = {A software framework for probabilistic sensitivity analysis for computationally expensive models}, series = {Advances in Engineering Software}, journal = {Advances in Engineering Software}, pages = {19 -- 31}, abstract = {A software framework for probabilistic sensitivity analysis for computationally expensive models}, subject = {Angewandte Mathematik}, language = {en} } @article{NanthakumarLahmerZhuangetal., author = {Nanthakumar, S.S. and Lahmer, Tom and Zhuang, Xiaoying and Park, Harold S. and Rabczuk, Timon}, title = {Topology optimization of piezoelectric nanostructures}, series = {Journal of the Mechanics and Physics of Solids}, journal = {Journal of the Mechanics and Physics of Solids}, pages = {316 -- 335}, abstract = {Topology optimization of piezoelectric nanostructures}, subject = {Angewandte Mathematik}, language = {en} } @article{NanthakumarLahmerZhuangetal., author = {Nanthakumar, S.S. and Lahmer, Tom and Zhuang, Xiaoying and Zi, Goangseup and Rabczuk, Timon}, title = {Detection of material interfaces using a regularized level set method in piezoelectric structures}, series = {Inverse Problems in Science and Engineering}, journal = {Inverse Problems in Science and Engineering}, pages = {153 -- 176}, abstract = {Detection of material interfaces using a regularized level set method in piezoelectric structures}, subject = {Angewandte Mathematik}, language = {en} } @article{ZhuangHuangRabczuketal., author = {Zhuang, Xiaoying and Huang, Runqiu and Rabczuk, Timon and Liang, C.}, title = {A coupled thermo-hydro-mechanical model of jointed hard rock for compressed air energy storage}, series = {Mathematical Problems in Engineering}, journal = {Mathematical Problems in Engineering}, abstract = {A coupled thermo-hydro-mechanical model of jointed hard rock for compressed air energy storage}, subject = {Angewandte Mathematik}, language = {en} }