Physics informed neural networks for continuum micromechanics
Alexander Henkes, Henning Wessels, Rolf Mahnken
Abstract
Abstract The present work proposes a Physics Informed Neural Network (PINN) for solving boundary value problems in continuum micromechanics. The presented technique is therefore an alternative to the finite element method or Fourier transform based methods. In this context, a neural network is used to approximate the function solving the partial differential equation. The theory of PINN in the context of micromechanics is developed.
Topics & Concepts
MicromechanicsContext (archaeology)Artificial neural networkPhysicsBoundary value problemPartial differential equationStatistical physicsApplied mathematicsFinite element methodClassical mechanicsCalculus (dental)Mathematical analysisComputer scienceMathematicsArtificial intelligenceAlgorithmBiologyPaleontologyComposite numberDentistryMedicineThermodynamicsModel Reduction and Neural NetworksAdvanced machining processes and optimizationNumerical methods in engineering