scholarly journals Multigrid and Krylov Solvers for Large Scale Finite Element Groundwater Flow Simulations on Distributed Memory Parallel Platforms

1997 ◽  
Author(s):  
K Mahinthakumar
Keyword(s):  
Finite Element ◽  
Groundwater Flow ◽  
Large Scale ◽  
Distributed Memory ◽  
Flow Simulations

Multiscale-finite-element-based ensemble Kalman filter for large-scale groundwater flow

2012 ◽  
Vol 468-469 ◽  
pp. 22-34 ◽  
Author(s):  
Liangsheng Shi ◽  
Lingzao Zeng ◽  
Dongxiao Zhang ◽  
Jinzhong Yang
Keyword(s):  
Finite Element ◽  
Kalman Filter ◽  
Groundwater Flow ◽  
Ensemble Kalman Filter ◽  
Large Scale ◽  
Multiscale Finite Element

scholarly journals Stator Non-Uniform Radial Ventilation Design Methodology for a 15 MW Turbo-Synchronous Generator Based on Single Ventilation Duct Subsystem

Energies ◽  
2021 ◽  
Vol 14 (10) ◽  
pp. 2760
Author(s):  
Ruiye Li ◽  
Peng Cheng ◽  
Hai Lan ◽  
Weili Li ◽  
David Gerada ◽  
...  
Keyword(s):  
Finite Element ◽  
Temperature Distribution ◽  
Design Methodology ◽  
Large Scale ◽  
Synchronous Generator ◽  
Axial Direction ◽  
Scale Model ◽  
Element Analysis ◽  
Axial Temperature ◽  
Ventilation Duct

Within large turboalternators, the excessive local temperatures and spatially distributed temperature differences can accelerate the deterioration of electrical insulation as well as lead to deformation of components, which may cause major machine malfunctions. In order to homogenise the stator axial temperature distribution whilst reducing the maximum stator temperature, this paper presents a novel non-uniform radial ventilation ducts design methodology. To reduce the huge computational costs resulting from the large-scale model, the stator is decomposed into several single ventilation duct subsystems (SVDSs) along the axial direction, with each SVDS connected in series with the medium of the air gap flow rate. The calculation of electromagnetic and thermal performances within SVDS are completed by finite element method (FEM) and computational fluid dynamics (CFD), respectively. To improve the optimization efficiency, the radial basis function neural network (RBFNN) model is employed to approximate the finite element analysis, while the novel isometric sampling method (ISM) is designed to trade off the cost and accuracy of the process. It is found that the proposed methodology can provide optimal design schemes of SVDS with uniform axial temperature distribution, and the needed computation cost is markedly reduced. Finally, results based on a 15 MW turboalternator show that the peak temperature can be reduced by 7.3 ∘C (6.4%). The proposed methodology can be applied for the design and optimisation of electromagnetic-thermal coupling of other electrical machines with long axial dimensions.

Sign in / Sign up

Export Citation Format

Share Document