000118788 001__ 118788
000118788 005__ 20230519145358.0
000118788 0247_ $$2doi$$a10.1016/j.cma.2021.113816
000118788 0248_ $$2sideral$$a125744
000118788 037__ $$aART-2021-125744
000118788 041__ $$aeng
000118788 100__ $$0(orcid)0000-0003-2564-6038$$aAyensa-Jimenez, J$$uUniversidad de Zaragoza
000118788 245__ $$aPrediction and identification of physical systems by means of physically-guided neural networks with meaningful internal layers
000118788 260__ $$c2021
000118788 5060_ $$aAccess copy available to the general public$$fUnrestricted
000118788 5203_ $$aSubstitution 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.
000118788 536__ $$9info:eu-repo/grantAgreement/ES/ISCIII/CIBER-BBN$$9info:eu-repo/grantAgreement/ES/MICINN-FEDER/PGC2018-097257-B-C31
000118788 540__ $$9info:eu-repo/semantics/openAccess$$aby-nc-nd$$uhttp://creativecommons.org/licenses/by-nc-nd/3.0/es/
000118788 590__ $$a6.588$$b2021
000118788 591__ $$aENGINEERING, MULTIDISCIPLINARY$$b8 / 92 = 0.087$$c2021$$dQ1$$eT1
000118788 591__ $$aMECHANICS$$b9 / 138 = 0.065$$c2021$$dQ1$$eT1
000118788 591__ $$aMATHEMATICS, INTERDISCIPLINARY APPLICATIONS$$b4 / 108 = 0.037$$c2021$$dQ1$$eT1
000118788 592__ $$a2.179$$b2021
000118788 593__ $$aComputational Mechanics$$c2021$$dQ1
000118788 593__ $$aPhysics and Astronomy (miscellaneous)$$c2021$$dQ1
000118788 593__ $$aMechanics of Materials$$c2021$$dQ1
000118788 593__ $$aComputer Science Applications$$c2021$$dQ1
000118788 594__ $$a10.3$$b2021
000118788 655_4 $$ainfo:eu-repo/semantics/article$$vinfo:eu-repo/semantics/acceptedVersion
000118788 700__ $$0(orcid)0000-0003-0088-7222$$aHamdy Doweidar, M$$uUniversidad de Zaragoza
000118788 700__ $$aSanz-Herrera, JA
000118788 700__ $$0(orcid)0000-0001-8741-6452$$aDoblare, M$$uUniversidad de Zaragoza
000118788 7102_ $$15004$$2605$$aUniversidad de Zaragoza$$bDpto. Ingeniería Mecánica$$cÁrea Mec.Med.Cont. y Teor.Est.
000118788 773__ $$g381 (2021), 113816 [33 pp]$$pComput. methods appl. mech. eng.$$tComputer Methods in Applied Mechanics and Engineering$$x0045-7825
000118788 8564_ $$s4916129$$uhttps://zaguan.unizar.es/record/118788/files/texto_completo.pdf$$yPostprint
000118788 8564_ $$s1056731$$uhttps://zaguan.unizar.es/record/118788/files/texto_completo.jpg?subformat=icon$$xicon$$yPostprint
000118788 909CO $$ooai:zaguan.unizar.es:118788$$particulos$$pdriver
000118788 951__ $$a2023-05-18-13:36:03
000118788 980__ $$aARTICLE