Registered Data
Contents
- 1 [CT106]
- 1.1 [01135] Computational framework for design, optimization, and control of sintering process
- 1.2 [00493] Discontinuous Galerkin method for nonlinear time-fractional integro-partial differential equations
- 1.3 [00494] A two-grid discontinuous Galerkin method to nonlinear time-fractional diffusion equations
- 1.4 [00568] PRIMAL HYBRID METHOD FOR QUASILINEAR PARABOLIC PROBLEMS
- 1.5 [00627] Approximated Well-Balnced DG enhanced by deep learning
[CT106]
[01135] Computational framework for design, optimization, and control of sintering process
- Session Date & Time : 5D (Aug.25, 15:30-17:10)
- Type : Contributed Talk
- Abstract : Optimization and control of sintering, which is governed by macroscopic coupled electro-thermo-mechanical model, is critical for additive manufacturing. To reduce time in making process design decisions, we present a hybridized model-based and data-driven computational framework. In particular, we demonstrate an inverse estimation strategy combining offline-online efficient data assimilation and surrogate modeling. We test the performance of the proposed method in accelerating the estimation of sintering process parameters and hard-to-measure material properties like viscosity.
- Classification : 65M32, 65F55, 74H15, 60-08, 74D99
- Author(s) :
- Rahul Dhopeshwar (TU Eindhoven)
- Harshit Bansal (TU Eindhoven)
- Karen Veroy (TU Eindhoven)
[00493] Discontinuous Galerkin method for nonlinear time-fractional integro-partial differential equations
- Session Date & Time : 5D (Aug.25, 15:30-17:10)
- Type : Contributed Talk
- Abstract : A numerical method is proposed for semilinear time-fractional integro-partial differential equations. Applying the Newton’s quasilinearization, we obtain a sequence of linear problems. To discretize the continuous problem, we apply the L1-scheme for time-fractional derivative and the trapezoidal rule for integral term. Then, we apply the Nonsymmetric Interior Penalty Galerkin (NIPG) method for the spatial derivatives appearing in the semi-discrete problem. $L^2$-norm stability and error estimates are studied. Numerical experiments are presented to validate the result.
- Classification : 65M60, 65M15
- Author(s) :
- Natesan Srinivasan (Indian Institute of Technology Guwahati)
- Sandip Maji (Indian Institute of Technology Guwahati)
[00494] A two-grid discontinuous Galerkin method to nonlinear time-fractional diffusion equations
- Session Date & Time : 5D (Aug.25, 15:30-17:10)
- Type : Contributed Talk
- Abstract : A two-grid algorithm for discontinuous Galerkin method is proposed for nonlinear time-fractional diffusion initial-boundary-value problems. The numerical scheme consists of DG method for the spatial derivatives and L1-scheme for time stepping. Error estimate for the proposed scheme is obtained. The numerical experiments are presented to prove the efficiency of our algorithm.
- Classification : 65M60, 65M15
- Author(s) :
- SANDIP MAJI (Indian Institute of Technology Guwahati)
- Natesan Srinivasan (Indian Institute of Technology Guwahati)
[00568] PRIMAL HYBRID METHOD FOR QUASILINEAR PARABOLIC PROBLEMS
- Session Date & Time : 5D (Aug.25, 15:30-17:10)
- Type : Contributed Talk
- Abstract : A second-order quasi-linear parabolic initial-boundary value problem is approximated by using primal hybrid finite element method and Lagrange multipliers. Semidiscrete and backward Euler based fully discrete schemes are discussed and optimal order error estimates are established by applying modified elliptic projection. Optimal order error estimates in maximum norm are also derived. Earlier results on maximum-norm superconvergence of the gradient in piecewise linear finite-element approximations of elliptic and parabolic problems are now carried over to quasilinear case using primal hybrid method. Finally, the results on numerical experiments confirm our theoretical findings.
- Classification : 65M60
- Author(s) :
- Ajit Patel (The LNM Institute of Information Technology)
- Ravina Shokeen (The LNM Institute of Information Technology)
- Amiya Kumar Pani (Birla Institute of Technology & Science, Pilani)
[00627] Approximated Well-Balnced DG enhanced by deep learning
- Session Date & Time : 5D (Aug.25, 15:30-17:10)
- Type : Contributed Talk
- Abstract : We place ourselves in the framework of the solution of hyperbolic equations by discontinuous by discontinuous Galerkin methods. In order to improve the accuracy of the method we propose a new approach where we use as basis functions neural networks pre-trained on families of solutions. Theoretical elements will be given and the approach will be validated on classical hyperbolic equations with source term. The obtained schemes will be approximated well-balanced
- Classification : 65M60
- Author(s) :
- Emmanuel Franck (INRIA )
- emmanuel Franck (INRIA)