search

Exawind / nalu-wind

Solver for wind farm simulations targeting exascale computational platforms
https://exawind.github.io/nalu-wind/
Other
124 stars 85 forks source link

Failed running turbine_uniform_inflow & wind_farm/case3 example cases. #521

Closed yoonc5536 closed 4 years ago

yoonc5536 commented 4 years ago

Detailed description of the bug

Failed running turbine_uniform_inflow & wind_farm/case3 example cases. It is difficult to track which part of the code causes the issue.

Details to reproduce bug/build error

  1. SHA ID of the Nalu-Wind commit you are using d503c4e7fef16d0d7ec5ebfcb306d00e66f54e77

  2. Configure command [attach file] No errors

  3. CMake configure output [attach file] No errors

  4. Output of make step [attach file] No errors

  5. Details of your simulation (input file, mesh description, etc.) Used all inputs files as provided in the example directories.

Console output:

OpenFAST

Copyright (C) National Renewable Energy Laboratory Copyright (C) Envision Energy USA LTD

This program is licensed under Apache License Version 2.0 and comes with ABSOLUTELY NO WARRANTY. See the "LICENSE" file distributed with this software for details.

OpenFAST-v2.2.0 Compile Info:

Log file of simulation:

=============================================================================== Nalu-Wind
An incompressible, turbulent computational fluid dynamics solver
for wind turbine and wind farm simulations

Nalu-Wind Version: v1.2.0 Nalu-Wind GIT Commit SHA: d503c4e7fef16d0d7ec5ebfcb306d00e66f54e77 Trilinos Version: 12.17-gb42c5a1586

TPLs: Boost, HDF5, netCDF, STK, Trilinos, yaml-cpp and zlib
Optional TPLs enabled: OpenFAST, Hypre, TIOGA Copyright 2017 National Technology & Engineering Solutions of Sandia, LLC
(NTESS), National Renewable Energy Laboratory, University of Texas Austin,
Northwest Research Associates. Under the terms of Contract DE-NA0003525
with NTESS, the U.S. Government retains certain rights in this software. 

       This software is released under the BSD 3-clause license.

See LICENSE file at https://github.com/exawind/nalu-wind for more details.

Actuator::load

Realm Options Review: realm_1

Element-based scheme will be activated Warning: When using automatic_decomposition_type, one must have a serial file Nalu will deactivate aura ghosting

Turbulence Model Review:

Turbulence Model is: laminar 0

PPE review:

Realm::create_mesh(): Begin Realm::create_mesh() End

Boundary Condition Review:

Inflow BC name: bc_1 on west Open BC name: bc_2 on east Symmetry BC name: bc_3 on north Symmetry BC name: bc_4 on south Symmetry BC name: bc_5 on lower Symmetry BC name: bc_6 on upper

Initial Condition Review:

Initial Is Type constant constant IC: name: ic_1 , target[0] = Unspecified-2-HEX

Material Prop Review:

density is a constant property: 1 viscosity is a constant property: 1e-05

EqSys/options Review:

Edge projected nodal gradient for velocity: 0 Edge projected nodal gradient for pressure: 0

Realm memory Review: realm_1

Node count from meta data = 612977 Total memory estimate for Matrix solve (per core)= 0.061655 GB. Total memory estimate (per core) = 0.061655 GB.

Time Integrator Review:

StandardTimeIntegrator name= ti_1 second order = 0 maxTimeStepCount = 10000 fixed time step is active with time step: 0.125 StandardTimeIntegrator realm_name[0]= realm_1

Transfer Review:

Realm::initialize() Begin Created algorithm = Alg0000_elem_HEXAHEDRON_8_geometry Created algorithm = Alg0000_algorithm_momentum_nodal_grad Created the following interior elem alg: MomentumEQS_AssembleElemSolverAlg_HEXAHEDRON_8 Built Kernels for momentum are lumped_actuator',lumped_momentum_time_derivative', upw_advection_diffusion'. Created algorithm = Alg0000_algorithm_continuity_nodal_grad Created algorithm = Alg0000_algorithm_mdot_elem_interior Created the following interior elem alg: ContinuityEQS_AssembleElemSolverAlg_HEXAHEDRON_8 Built Kernels for continuity areadvection'. Created algorithm = Alg0001_face_QUADRILATERAL_4_geometry Created algorithm = Alg0002_face_QUADRILATERAL_4_momentum_nodal_grad Created algorithm = Alg0002_face_QUADRILATERAL_4_continuity_nodal_grad Created algorithm = Alg0002_face_QUADRILATERAL_4_mdot_inflow Created algorithm = Alg0005_face_QUADRILATERAL_4_momentum_nodal_grad Created algorithm = Alg0005_algorithm_continuity_nodal_grad Created algorithm = Alg0005_algorithm_mdot_open_elem Created algorithm = Alg0008_face_QUADRILATERAL_4_momentum_nodal_grad Created algorithm = Alg0008_face_QUADRILATERAL_4_continuity_nodalgrad Time step ratio dtNalu/dtFAST: 20 Realm::ioBroker->populatemesh() Begin Realm::ioBroker->populatemesh() End Realm::ioBroker->populate_fielddata() Begin Realm::ioBroker->populate_field_data() End Realm::create_output_mesh(): Begin Realm::create_output_mesh() End ActuatorLineFAST alg will ghost a number of entities: 977835 EquationSystems::initialize(): Begin EquationSystems::initialize(): End

Realm memory Review: realm_1

Total memory estimate for Matrix solve (per core)= 0.061655 GB. Total memory estimate for Fields (per core)= 0.00469933 GB. Total memory estimate (per core) = 0.0663543 GB. Realm::initialize() End

Simulation Shall Commence: number of processors = 60

Realm shall provide output files at : currentTime/timeStepCount: 0/0 (realm_1)

Time Step Count: 1 Current Time: 0.125 dtN: 0.125 dtNm1: 0.125 gammas: 1 -1 0 Max Courant: 0.25 Max Reynolds: 3.2e+06 (realm_1) Realm Nonlinear Iteration: 1/1

realm_1::advance_time_step() NLI Name Linear Iter Linear Res NLinear Res Scaled NLR

sayerhs commented 4 years ago

@yoonc5536 Thanks for reporting the issue we will look into this. A couple of requests:

yoonc5536 commented 4 years ago

Thank you for your feedback. I will try this these regression test cases and update you.

yoonc5536 commented 4 years ago

Quick question. Is there a way not to use OpenFAST for Nalu?

sayerhs commented 4 years ago

@yoonc5536 Currently, using actuator models (disk or line) requires integration with OpenFAST.

yoonc5536 commented 4 years ago

Confirmed the regression test cases, nrel5MWActuatorLine and nrel5MWActuatorLine properly. The example cases seem to be the trouble. Thank you!