version:0.44

version:0.43

(28/04/2021)

• SDG/CDG methods are parallelized
• adding Hd in hysteresis data
• if restoring failed during relaxation, restart from t=0 instead of throwing an error
• change the option hystesis-ncycles to hysteresis-nhalfcycles
• bug correction on building Neumann Data for FEM displacement
• bug correction on CDG in swaping M at level l+1 and at level l
• bug correction on log for ODE solving
• debuging test added a module to save the M.H at each level at first iteration

version:0.40

(10/02/2021)

• thow an exception when
  • the size of the .vel .dis files are not equal to the number of the points of the mesh
  • the limit condition on points are necessary adapted to the number of the points of the mesh
• correction of a bug concerning the hysteresis restroing file when no data on hysteresis file
• correction of a bug concerning the periodic domain for magnetostriction operator
• correction on the updating of U at each equilibrium solver
• doc updated with UML diagrams

version:0.31

(29/01/2021)

• adding an equilibrium state for elastic equation (in dev FVM elasticity matrix is not symmetric)
• adding 2 solvers : CGS & biCGstab for non symmetric linear solver
• adding an equilibrium tests
• restore hysteresis even if no hysteresis file
• adding a mean of H at the hysteresis data and in the log of relaxation
• bug corection :
  • on core dump for elasticty-state=frozen (frozen|equilibrium|unsteady)

version:0.30

(16/12/2020)

• adding an ODE module for solving the landau lifschitz system with options: –relaxation-time-order=rk45 or rk45d for a different implementation –stepper=ode to set the adaptative stepper –stepper-evaluations=12: to set the maximum number of iterations for calculating a new stepper ie time step –stepper-tolerance=1.e-4: to set the tolerated error on the dM/dt to compute the stepper
• the time-... options have changed into stepper-... options: the time-... will be deprecated in future version

version:0.294

(09/12/2020)

• adding a time step computing method brent(as quadratic number which will be deprecated) and newton raphson (not pertinent for our minimuzation)
• add an inside box within the PUC geometry –puc-ibox-origin=[Px,Py,Pz] –puc-ibox-size=[Hx,Hy,Hz] : the size of the inside boxix Hx x Hy x Hz from the orgin point [Px,Py,Pz]. All real values are relative values in [0,1] with respect of the box size.
• add a non isotropy hyper elastic material
• for adimensionizing the elasticity equation the defaut parameter is . We can use instead . Set its value into the –elastic-tensor-adim option (if 0 the adimensionized parameter is the default one) and –magnetic-adim is either true to use Magnetic excitation field normalization or false to use elastic magnetic excitation field normalization.
• add a scale multiplied factor for U & V by the option –displacement-scale –velocity-scale (1.0 by default)
• adding a cubic anisotropy and static magnetostriction case test
• bugs correction:
  • for displacement with elastic steady state on : at t>0, return U & V at t=0 to avoid null
  • morse array bug correction with empty elements in begin(start,pack,iter) method
  • initial time step for iterative time method corrected

version:0.28

(20/11/04)

• add an iterative method based on optimal time : use the optimal time method even in out of its domain application. If the energy is deacreasing, the time step is the time step computed when considering all operators are linear. If not, the golden iterative method is used. (option –time=optimal-golden)
• add an option to use only the evolution of the displacement field when the magnetization field is fixed. The time step is constant (–use-elasticity-only=true| false option, false by default).
• a martensite interface added between grains:
  • martensite-width={0,1,2} width of the interface in cells dimension
  • martensite-matter=[name,Ms,...] is the matter description of the martensite matter (let it empty for no matter)
• add a neuman constraints on boundaries in displacement with 2 new options:
  • constraint-faces in [0,63] : indictaors of neumann faces
  • constraint-force=constraint.cst : file which describes the constraint on points or a constant vector [Cx,Cy,Cz]
• add a Finite Volume Method (FVM) based on Sigma.n = c on Neuman boundaries implemented in 4 methods for gradient rebuild :
  • FVM-VOGGR, for volume gradient rebuild based on Ostrogradsky Green integration
  • FVM-VTEGR, for volume gradient rebuild based on Taylor Expansion integration
  • FVM-SSGR, for tangential surface gradient rebuild based on Stokes integration
  • FVM-STEGR for tangential surface gradient rebuild based on Taylor Expansion integration
• add a constraint data for elastic operators
• bug corrections:
  • error in initialization of variables for computing E in iterative methods
  • modification of the time step iterative methods to ensure the decreasing of the energy. For the optimal direct method or regular time step mehod, the verification is not done.
  • velocity approximation bug
  • bad extension for VTI files
  • FDM Methods are removed because of bad approximation for not null neumann limit conditions.

version:0.27

(20/08/11)

• add a sheet waves with 3 numerical methods cross verification
• modifing the saving of the system : save the vti file, the magnetization in txt file, the relaxation state and the data operators
• change the option "vtkSteps" into "output-steps" for saving the data at output. It alawys saves teh vti files but also the relaxation file and the operators data
• bug corrections:
  • correction of the type of U/V/dV_dt in vti files for FEM displacement operator
  • reset to initial state of all data operators at each hysteresis step
• to do list:
  • magnetostriction data not optimized for holes.
  • add magnetostriction case test
  • add the finest periodical multi scale method
  • optimization of the symmetric tensor product & memory
  • static constant stress should be implemented for heterogenous matters if it becomes permanent

version:0.26

(20/08/05)

• add an option for maximum time step gap between two successive time step computing –time-max-gap option in [1,infinity[
• validation of the 1D periodical case for magnetostriction
• adding tDimension type for dimension of field tLimitConditionType for limit conditions and tCellFlag type for boolean value on faces of cell
• adding a backup file for epsilonU
• bug corrections:
  • disabled the memory check for multi-threadings version
  • correction in output Path for case tests
• to do list:
  • magnetostriction data not optimized for holes.
  • add magnetostriction case test
  • add the finest periodical multi scale method
  • optimization of the symmetric tensor product & memory
  • static constant stress should be implemented for heterogenous matters if it becomes permanent

version:0.25

(20/06/10)

• add a static magnetic stress operator
• FDM/FEM/C-FEM methods validated in 1D-Case
• bug corrections:
  • hysteresis log writing error for physical caracteritic data
  • mass matrix product error for product method in case of periodicity
• to do list:
  • magnetostriction data not optimized for holes.
  • add magnetostriction case test
  • add the finest periodical multi scale method
  • optimization of the symmetric tensor product & memory
  • static constant stress should be implemented for heterogenous matters if it becomes permanent

version:0.24

(20/06/10)

• add a steady state for displacement field of magnetostriction: the displacement field is constant within the relaxation process iteration. The corresponding option is –elasticity-steady-state=true | false (false by default)
• add a static magnetostriction operator
• add a condensed FEM mass matrix and rename AcceleratorMatrix to MassMatrix
• add a data-path & output-path options for data generator program
• modify the time order integration:
  • for the EMM_IterativeTime, the time order is 1 or 2 instead of GL1|GL4|TE2 (time-order=1|2).
  • The evolution of M,U with respect of t during the iterative process with a time step is as follow:
    • for integration of M over the time thanks to landau lisfchitz equation,
      • if order 1, it uses a GL1 integration for ggl method
      • if order 2, it uses a GL4 integration for ggl method
      • the evolution of M at each time step is given by relaxation-time-order=1|2 and relaxation-time-integration=te|gl|ggl
    • for integration of the displacement over the time thanks to the magnetostriction equation,
      • if order 1, it uses an integration of order 1 in time depending of the numerical method by displacement-time-integration=TE|P1|P2)
      • if order 2, it uses an integration of order 2 in time depending of the numerical method by displacement-time-order=1|2)
      • the evolution of U at each time step is given by displacement-time-order= 1|2 ;
• to do list:
  • magnetostriction data not optimized for holes.
  • validate FEM 1D
  • add magnetostriction case test
  • add the finest periodical multi scale method

version:0.232

(20/05/28)

• change the sign of the cubic anistropy energy and the magnetic excitation field
• bug corrections:
  • normalization of anisotropy directions done.
• to do list:
  • add a static magnetostriction
  • magnetostriction data not optimized for holes.
  • validate FEM 1D
  • add magnetostriction case test
  • add the finest periodical multi scale method

version:0.231

• add a through holes filter for ie –filter=THROUGH_HOLES[z,uniform,1000,0,327680,0.,1.,1.] for 1000 through uniform holes in the direction z with null seed and max of 327680 random numbers in [0,1] with the probability 1.
• bug corrections:
  • OpenMp array out of bounds in kappa arry in FDM energy computing
  • restore failing (bad key word)
  • 1D Wave FEM Elastic in progress

    version:0.23

• if the evaluations number (time-evaluation option) for the time step estimation is null then the time step is computed as the optimal time considering all operators as linear operators.
• add a taylor expansion approximation for displacement (time-order / integration-time-method options set to GL1/GL4/TE2)
• add a 4-order polynomial time step evaluation (time option set to optimal/quadratic/golden/polynomial)
• add a time-initial-reset option to initialize time step to optimal one when considering all operators as linear ones.
• add hysteresis-M-init option for hysteresis process to initialize the magnetization for Hext=H0. It may be
  • empty to initialize M in the same direction of H0
  • [Mx,My,Mz] to initialize M in the same direction as the 3D vector (Mx,My,Mz)
  • fileName.aim or fileName.txt to define M with a field description
• bug corrections:
  • FEM elastic strain error computing for gradient of base function integrals
  • hyperelastic matter bug in tensor buildings
  • bug correction in reading h-ext option value for floating point values

version:0.222

• bug corrections: -bad restore file read in hysteresis process -relaxation rate was not taken into account

version:0.221

• saving data in other hysteresis file .data
• bug corrections:
  • last hysteresis data flush added

version:0.22

• adding volumic mass parameter in matter
• adding volumic mass in displacement operator
• testing magnetostriction operator
• add an relaxation output file for hysteresis

version:0.21

• adding normal holes filter
• adding cleaner option in hysteresis process
• cases tests added:
  • normal holes cross tie geometries
• bug corrections:
  • adimensionization of the elastic and magnetic tensors for magnetostriction
• adding normal holes filter
• adding cleaner option in hysteresis process
• adding volumic mass parameter in matter
• cases tests added:
  • normal holes cross tie geometries
• bug corrections:
  • adimensionization of the elastic and magnetic tensors for magnetostriction

version:0.20

• adding magnetostriction operator and displacement operators for mixed Dirichlet and Neumann limit condition
• adding primary and elementary tests for magnetostriction problem
• adding iterative linear system solver package included conjugate gradient method for solving FEM method
• adding 4-tensor package for managing the data of magnetoscription process
• adding an order to time function evaluation for iterative time step computing
• adding an hyper elastic matter
• modification of the initialization of the initial time step when it is null for time iterative methods
• omp compilation
• adding a data files generator module containing a new geometry parser expression toolkit.
• bugs correction:
  • K2 reading in anisotropy operators

version:0.10

• bug corrections concerning the path for hysteresis i/o, material data readings,...
• adding magnetostriction/displacement operators
• splitting the operators package into 5 packages : one per operator (zeeman,exchange,anisotropy,demagnetized,magnetostriction)
• adding several implementations of multi grid scale for primary tests validation
• adding cubic anisotropy case test for a thin sheet

version:0.00

• this version contains the following operators in periodical domains :
  • zeeman
  • exchange with minimal and full storage
  • anisotropy (planar,uniaxial, cubic and mixed)
  • demagnetized with 4 implementations of Toeplitz Matrix . vector product
• It computes :
  • hysteresis cycles
• It contaoins the test cases with no periodical mesh:
  • cross tie
  • periodical cross tie
  • thin sheet difusion
  • double vortex bar
  • Y wires
  • holed cross tie
• The documentation has been updated.