118788 20230519145358.0 doi 10.1016/j.cma.2021.113816 sideral 125744 ART-2021-125744 eng Ayensa-Jimenez, J Universidad de Zaragoza (orcid)0000-0003-2564-6038 Prediction and identification of physical systems by means of physically-guided neural networks with meaningful internal layers 2021 Access copy available to the general public Unrestricted Substitution of well-grounded theoretical models by data-driven predictions is not as simple in engineering and sciences as it is in social and economic fields. Scientific problems suffer many times from paucity of data, while they may involve a large number of variables and parameters that interact in complex and non-stationary ways, obeying certain physical laws. Moreover, a physically-based model is not only useful for making predictions, but to gain knowledge by the interpretation of its structure, parameters, and mathematical properties. The solution to these shortcomings seems to be the seamless blending of the tremendous predictive power of the data-driven approach with the scientific consistency and interpretability of physically-based models. We use here the concept of Physically-Guided Neural Networks (PGNN) to predict the input-output relation in a physical system, while, at the same time, fulfilling the physical constraints. With this goal, the internal hidden state variables of the system are associated with a set of internal neuron layers, whose values are constrained by known physical relations, as well as any additional knowledge on the system. Furthermore, when having enough data, it is possible to infer knowledge about the internal structure of the system and, if parameterized, to predict the state parameters for a particular input-output relation. We show that this approach, besides getting physically-based predictions, accelerates the training process, reduces the amount of data required to get similar accuracy, partly filters the intrinsic noise in the experimental data and improves its extrapolation capacity. (C) 2021 ElsevierB.V. All rights reserved. info:eu-repo/grantAgreement/ES/ISCIII/CIBER-BBN info:eu-repo/grantAgreement/ES/MICINN-FEDER/PGC2018-097257-B-C31 info:eu-repo/semantics/openAccess by-nc-nd http://creativecommons.org/licenses/by-nc-nd/3.0/es/ 6.588 2021 ENGINEERING, MULTIDISCIPLINARY 8 / 92 = 0.087 2021 Q1 T1 MECHANICS 9 / 138 = 0.065 2021 Q1 T1 MATHEMATICS, INTERDISCIPLINARY APPLICATIONS 4 / 108 = 0.037 2021 Q1 T1 2.179 2021 Computational Mechanics 2021 Q1 Physics and Astronomy (miscellaneous) 2021 Q1 Mechanics of Materials 2021 Q1 Computer Science Applications 2021 Q1 10.3 2021 info:eu-repo/semantics/article info:eu-repo/semantics/acceptedVersion Hamdy Doweidar, M Universidad de Zaragoza (orcid)0000-0003-0088-7222 Sanz-Herrera, JA Doblare, M Universidad de Zaragoza (orcid)0000-0001-8741-6452 5004 605 Universidad de Zaragoza Dpto. Ingeniería Mecánica Área Mec.Med.Cont. y Teor.Est. 381 (2021), 113816 [33 pp] Comput. methods appl. mech. eng. Computer Methods in Applied Mechanics and Engineering 0045-7825 4916129 http://zaguan.unizar.es/record/118788/files/texto_completo.pdf Postprint 1056731 http://zaguan.unizar.es/record/118788/files/texto_completo.jpg?subformat=icon icon Postprint oai:zaguan.unizar.es:118788 articulos driver 2023-05-18-13:36:03 ARTICLE