OOFEM
2.4
OOFEM.org - Object Oriented Finite Element Solver
|
This class implements NonLinear (- may be changed) solution of dynamic problems using Direct Explicit Integration scheme - Central Difference Method. More...
#include <nldeidynamic.h>
Public Member Functions | |
NlDEIDynamic (int i, EngngModel *_master=NULL) | |
virtual | ~NlDEIDynamic () |
virtual void | solveYourself () |
Starts solution process. More... | |
virtual void | solveYourselfAt (TimeStep *tStep) |
Solves problem for given time step. More... | |
virtual void | updateYourself (TimeStep *tStep) |
Updates internal state after finishing time step. More... | |
virtual double | giveUnknownComponent (ValueModeType type, TimeStep *tStep, Domain *d, Dof *dof) |
Returns requested unknown. More... | |
virtual IRResultType | initializeFrom (InputRecord *ir) |
Initializes receiver according to object description in input reader. More... | |
virtual TimeStep * | giveNextStep () |
Returns next time step (next to current step) of receiver. More... | |
virtual NumericalMethod * | giveNumericalMethod (MetaStep *mStep) |
Returns reference to receiver's numerical method. More... | |
virtual contextIOResultType | saveContext (DataStream &stream, ContextMode mode) |
Stores the state of model to output stream. More... | |
virtual contextIOResultType | restoreContext (DataStream &stream, ContextMode mode) |
Restores the state of model from output stream. More... | |
virtual void | printOutputAt (FILE *file, TimeStep *tStep) |
Prints output of receiver to output domain stream, for given time step. More... | |
virtual void | printDofOutputAt (FILE *stream, Dof *iDof, TimeStep *tStep) |
DOF printing routine. More... | |
virtual const char * | giveInputRecordName () const |
virtual const char * | giveClassName () const |
Returns class name of the receiver. More... | |
virtual fMode | giveFormulation () |
Indicates type of non linear computation (total or updated formulation). More... | |
virtual int | giveNumberOfFirstStep (bool force=false) |
Returns number of first time step used by receiver. More... | |
virtual int | estimateMaxPackSize (IntArray &commMap, DataStream &buff, int packUnpackType) |
Determines the space necessary for send/receive buffer. More... | |
Public Member Functions inherited from oofem::StructuralEngngModel | |
StructuralEngngModel (int i, EngngModel *_master=NULL) | |
Creates new StructuralEngngModel with number i, associated to domain d. More... | |
virtual | ~StructuralEngngModel () |
Destructor. More... | |
virtual int | checkConsistency () |
Allows programmer to test some receiver's internal data, before computation begins. More... | |
void | computeReaction (FloatArray &answer, TimeStep *tStep, int di) |
Computes reaction forces. More... | |
virtual void | terminate (TimeStep *tStep) |
Terminates the solution of time step. More... | |
void | buildReactionTable (IntArray &restrDofMans, IntArray &restrDofs, IntArray &eqn, TimeStep *tStep, int di) |
Builds the reaction force table. More... | |
virtual void | showSparseMtrxStructure (int type, oofegGraphicContext &gc, TimeStep *tStep) |
Shows the sparse structure of required matrix, type == 1 stiffness. More... | |
Public Member Functions inherited from oofem::EngngModel | |
EngngModel (int i, EngngModel *_master=NULL) | |
Constructor. More... | |
virtual | ~EngngModel () |
Destructor. More... | |
EngngModel (const EngngModel &)=delete | |
EngngModel & | operator= (const EngngModel &)=delete |
Domain * | giveDomain (int n) |
Service for accessing particular problem domain. More... | |
void | setDomain (int i, Domain *ptr, bool iDeallocateOld=true) |
Sets i-th domain of receiver. More... | |
int | giveNumberOfDomains () |
Returns number of domains in problem. More... | |
const std::string & | giveDescription () const |
const time_t & | giveStartTime () |
bool | giveSuppressOutput () const |
virtual ErrorEstimator * | giveDomainErrorEstimator (int n) |
Service for accessing ErrorEstimator corresponding to particular domain. More... | |
virtual MaterialInterface * | giveMaterialInterface (int n) |
Returns material interface representation for given domain. More... | |
void | setNumberOfEquations (int id, int neq) |
FILE * | giveOutputStream () |
Returns file descriptor of output file. More... | |
std::string | giveOutputBaseFileName () |
Returns base output file name to which extensions, like .out .vtu .osf should be added. More... | |
std::string | giveReferenceFileName () |
Returns reference file name. More... | |
void | letOutputBaseFileNameBe (const std::string &src) |
Sets the base output file name. More... | |
ContextOutputMode | giveContextOutputMode () |
Returns domain context output mode. More... | |
int | giveContextOutputStep () |
Returns domain context output step. More... | |
void | setContextOutputMode (ContextOutputMode contextMode) |
Sets context output mode of receiver. More... | |
void | setUDContextOutputMode (int cStep) |
Sets user defined context output mode (it sets contextOutputMode to contextOutputMode), setting contextOutputStep to given value. More... | |
void | setProblemMode (problemMode pmode) |
Sets domain mode to given mode. More... | |
void | setParallelMode (bool newParallelFlag) |
Sets the problem to run in parallel (or not). More... | |
problemMode | giveProblemMode () |
Returns domain mode. More... | |
void | setProblemScale (problemScale pscale) |
Sets scale in multiscale simulation. More... | |
problemScale | giveProblemScale () |
Returns scale in multiscale simulation. More... | |
virtual void | setRenumberFlag () |
Sets the renumber flag to true. More... | |
virtual void | resetRenumberFlag () |
Sets the renumber flag to false. More... | |
double | giveSolutionStepTime () |
Returns the user time of the current simulation step in seconds. More... | |
void | giveAnalysisTime (int &rhrs, int &rmin, int &rsec, int &uhrs, int &umin, int &usec) |
Returns the real and user time for the analysis. More... | |
void | terminateAnalysis () |
Performs analysis termination after finishing analysis. More... | |
virtual void | doStepOutput (TimeStep *tStep) |
Prints the ouput of the solution step (using virtual this->printOutputAtservice) to the stream detemined using this->giveOutputStream() method and calls exportModuleManager to do output. More... | |
void | saveStepContext (TimeStep *tStep, ContextMode mode) |
Saves context of given solution step, if required (determined using this->giveContextOutputMode() method). More... | |
virtual void | initializeYourself (TimeStep *tStep) |
Provides the opportunity to initialize state variables stored in element integration points according to initial conditions using function initializeYourself() on element level. More... | |
virtual int | initializeAdaptive (int tStepNumber) |
Initializes the newly generated discretization state according to previous solution. More... | |
virtual int | giveNumberOfDomainEquations (int di, const UnknownNumberingScheme &num) |
Returns number of equations for given domain in active (current time step) time step. More... | |
virtual FieldPtr | giveField (FieldType key, TimeStep *) |
Returns the smart pointer to requested field, Null otherwise. More... | |
EngngModel * | giveMasterEngngModel () |
Returns the master engnmodel. More... | |
virtual double | giveLoadLevel () |
Returns the current load level. More... | |
virtual double | giveEigenValue (int eigNum) |
Only relevant for eigen value analysis. Otherwise returns zero. More... | |
virtual void | setActiveVector (int i) |
Only relevant for eigen value analysis. Otherwise does noting. More... | |
int | updateSharedDofManagers (FloatArray &answer, const UnknownNumberingScheme &s, int ExchangeTag) |
Exchanges necessary remote DofManagers data. More... | |
int | exchangeRemoteElementData (int ExchangeTag) |
Exchanges necessary remote element data with remote partitions. More... | |
virtual int | giveCurrentNumberOfIterations () |
Returns number of iterations that was required to reach equilibrium - used for adaptive step length in staggered problem. More... | |
MPI_Comm | giveParallelComm () |
Returns the communication object of reciever. More... | |
int | packRemoteElementData (ProcessCommunicator &processComm) |
Packs data of local element to be received by their remote counterpart on remote partitions. More... | |
int | unpackRemoteElementData (ProcessCommunicator &processComm) |
Unpacks data for remote elements (which are mirrors of remote partition's local elements). More... | |
int | packDofManagers (ArrayWithNumbering *src, ProcessCommunicator &processComm) |
Packing function for vector values of DofManagers. More... | |
int | unpackDofManagers (ArrayWithNumbering *dest, ProcessCommunicator &processComm) |
Unpacking function for vector values of DofManagers . More... | |
ProblemCommunicator * | giveProblemCommunicator (EngngModelCommType t) |
void | initializeCommMaps (bool forceInit=false) |
virtual int | instanciateYourself (DataReader &dr, InputRecord *ir, const char *outFileName, const char *desc) |
Initializes whole problem according to its description stored in inputStream. More... | |
void | Instanciate_init () |
Initialization of the receiver state (opening the default output stream, empty domain creation, initialization of parallel context, etc) before Initialization form DataReader. More... | |
int | instanciateDomains (DataReader &dr) |
Instanciate problem domains by calling their instanciateYourself() service. More... | |
int | instanciateMetaSteps (DataReader &dr) |
Instanciate problem meta steps by calling their instanciateYourself() service. More... | |
virtual int | instanciateDefaultMetaStep (InputRecord *ir) |
Instanciate default metastep, if nmsteps is zero. More... | |
virtual void | updateAttributes (MetaStep *mStep) |
Update receiver attributes according to step metaStep attributes. More... | |
void | initMetaStepAttributes (MetaStep *mStep) |
Update e-model attributes attributes according to step metaStep attributes. More... | |
virtual void | updateDomainLinks () |
Updates domain links after the domains of receiver have changed. More... | |
MetaStep * | giveCurrentMetaStep () |
Returns current meta step. More... | |
virtual TimeStep * | giveCurrentStep (bool force=false) |
Returns current time step. More... | |
virtual TimeStep * | givePreviousStep (bool force=false) |
Returns previous time step. More... | |
TimeStep * | generateNextStep () |
Generate new time step (and associate metastep). More... | |
virtual void | preInitializeNextStep () |
Does a pre-initialization of the next time step (implement if necessarry) More... | |
virtual TimeStep * | giveSolutionStepWhenIcApply (bool force=false) |
Returns the solution step when Initial Conditions (IC) apply. More... | |
int | giveNumberOfMetaSteps () |
Return number of meta steps. More... | |
MetaStep * | giveMetaStep (int i) |
Returns the i-th meta step. More... | |
int | giveNumberOfSteps (bool force=false) |
Returns total number of steps. More... | |
virtual double | giveEndOfTimeOfInterest () |
Returns end of time interest (time corresponding to end of time integration). More... | |
int | giveNumberOfTimeStepWhenIcApply () |
Returns the time step number, when initial conditions should apply. More... | |
ExportModuleManager * | giveExportModuleManager () |
Returns receiver's export module manager. More... | |
EngngModelTimer * | giveTimer () |
Returns reference to receiver timer (EngngModelTimer). More... | |
virtual int | giveNewEquationNumber (int domain, DofIDItem) |
Increases number of equations of receiver's domain and returns newly created equation number. More... | |
virtual int | giveNewPrescribedEquationNumber (int domain, DofIDItem) |
Increases number of prescribed equations of receiver's domain and returns newly created equation number. More... | |
std::string | giveContextFileName (int tStepNumber, int stepVersion) const |
Returns the filename for the context file for the given step and version. More... | |
std::string | giveDomainFileName (int domainNum, int domainSerNum) const |
Returns the filename for the given domain (used by adaptivity and restore) More... | |
virtual void | updateComponent (TimeStep *tStep, NumericalCmpn cmpn, Domain *d) |
Updates components mapped to numerical method if necessary during solution process. More... | |
virtual void | initStepIncrements () |
Initializes solution of new time step. More... | |
virtual int | forceEquationNumbering (int i) |
Forces equation renumbering on given domain. More... | |
virtual int | forceEquationNumbering () |
Forces equation renumbering on all domains associated to engng model. More... | |
virtual int | requiresUnknownsDictionaryUpdate () |
Indicates if EngngModel requires Dofs dictionaries to be updated. More... | |
virtual bool | requiresEquationRenumbering (TimeStep *tStep) |
Returns true if equation renumbering is required for given solution step. More... | |
virtual void | updateDofUnknownsDictionary (DofManager *, TimeStep *) |
Updates necessary values in Dofs unknown dictionaries. More... | |
virtual int | giveUnknownDictHashIndx (ValueModeType mode, TimeStep *tStep) |
This method is responsible for computing unique dictionary id (ie hash value) from given valueModeType and time step. More... | |
virtual ParallelContext * | giveParallelContext (int n) |
Returns the parallel context corresponding to given domain (n) and unknown type Default implementation returns i-th context from parallelContextList. More... | |
virtual void | initParallelContexts () |
Creates parallel contexts. More... | |
virtual void | assemble (SparseMtrx &answer, TimeStep *tStep, const MatrixAssembler &ma, const UnknownNumberingScheme &s, Domain *domain) |
Assembles characteristic matrix of required type into given sparse matrix. More... | |
virtual void | assemble (SparseMtrx &answer, TimeStep *tStep, const MatrixAssembler &ma, const UnknownNumberingScheme &r_s, const UnknownNumberingScheme &c_s, Domain *domain) |
Assembles characteristic matrix of required type into given sparse matrix. More... | |
void | assembleVector (FloatArray &answer, TimeStep *tStep, const VectorAssembler &va, ValueModeType mode, const UnknownNumberingScheme &s, Domain *domain, FloatArray *eNorms=NULL) |
Assembles characteristic vector of required type from dofManagers, element, and active boundary conditions, into given vector. More... | |
void | assembleVectorFromDofManagers (FloatArray &answer, TimeStep *tStep, const VectorAssembler &va, ValueModeType mode, const UnknownNumberingScheme &s, Domain *domain, FloatArray *eNorms=NULL) |
Assembles characteristic vector of required type from dofManagers into given vector. More... | |
void | assembleVectorFromElements (FloatArray &answer, TimeStep *tStep, const VectorAssembler &va, ValueModeType mode, const UnknownNumberingScheme &s, Domain *domain, FloatArray *eNorms=NULL) |
Assembles characteristic vector of required type from elements into given vector. More... | |
void | assembleVectorFromBC (FloatArray &answer, TimeStep *tStep, const VectorAssembler &va, ValueModeType mode, const UnknownNumberingScheme &s, Domain *domain, FloatArray *eNorms=NULL) |
Assembles characteristic vector of required type from boundary conditions. More... | |
void | assembleExtrapolatedForces (FloatArray &answer, TimeStep *tStep, CharType type, Domain *domain) |
Assembles the extrapolated internal forces vector, useful for obtaining a good initial guess in nonlinear analysis with Dirichlet boundary conditions. More... | |
void | assemblePrescribedExtrapolatedForces (FloatArray &answer, TimeStep *tStep, CharType type, Domain *domain) |
void | assembleVectorFromContacts (FloatArray &answer, TimeStep *tStep, CharType type, ValueModeType mode, const UnknownNumberingScheme &s, Domain *domain, FloatArray *eNorms=NULL) |
virtual int | checkProblemConsistency () |
Allows programmer to test problem its internal data, before computation begins. More... | |
virtual void | init () |
Initializes the receiver state. More... | |
virtual void | postInitialize () |
Performs post-initialization for all the problem contents (which is called after initializeFrom). More... | |
virtual void | printOutputAt (FILE *file, TimeStep *tStep, const IntArray &nodeSets, const IntArray &elementSets) |
void | outputNodes (FILE *file, Domain &domain, TimeStep *tStep, int setNum) |
Outputs all nodes in the given set. More... | |
void | outputElements (FILE *file, Domain &domain, TimeStep *tStep, int setNum) |
Outputs all elements in the given set. More... | |
void | printYourself () |
Prints state of receiver. Useful for debugging. More... | |
virtual int | useNonlocalStiffnessOption () |
Returns nonzero if nonlocal stiffness option activated. More... | |
bool | isParallel () const |
Returns true if receiver in parallel mode. More... | |
int | giveRank () const |
Returns domain rank in a group of collaborating processes (0..groupSize-1) More... | |
int | giveNumberOfProcesses () const |
Returns the number of collaborating processes. More... | |
EngngModelContext * | giveContext () |
Context requesting service. More... | |
virtual int | giveNumberOfSlaveProblems () |
Returns number of slave problems. More... | |
virtual EngngModel * | giveSlaveProblem (int i) |
Returns i-th slave problem. More... | |
virtual bool | giveEquationScalingFlag () |
Returns the Equation scaling flag, which is used to indicate that governing equation(s) are scaled, or non-dimensionalized. More... | |
virtual double | giveVariableScale (VarScaleType varId) |
Returns the scale factor for given variable type. More... | |
virtual void | balanceLoad (TimeStep *tStep) |
Recovers the load balance between processors, if needed. More... | |
virtual LoadBalancer * | giveLoadBalancer () |
Returns reference to receiver's load balancer. More... | |
virtual LoadBalancerMonitor * | giveLoadBalancerMonitor () |
Returns reference to receiver's load balancer monitor. More... | |
void | initParallel () |
Request domain rank and problem size. More... | |
EngngModel * | giveEngngModel () |
Returns reference to itself -> required by communicator.h. More... | |
virtual bool | isElementActivated (int elemNum) |
virtual bool | isElementActivated (Element *e) |
virtual void | drawYourself (oofegGraphicContext &gc) |
virtual void | drawElements (oofegGraphicContext &gc) |
virtual void | drawNodes (oofegGraphicContext &gc) |
std::string | errorInfo (const char *func) const |
Returns string for prepending output (used by error reporting macros). More... | |
Protected Member Functions | |
void | computeLoadVector (FloatArray &answer, ValueModeType mode, TimeStep *tStep) |
Assembles the load vector. More... | |
void | computeMassMtrx (FloatArray &mass, double &maxOm, TimeStep *tStep) |
Assembles the diagonal mass matrix of receiver. More... | |
void | computeMassMtrx2 (FloatMatrix &mass, double &maxOm, TimeStep *tStep) |
Protected Member Functions inherited from oofem::StructuralEngngModel | |
void | printReactionForces (TimeStep *tStep, int id, FILE *out) |
Computes and prints reaction forces, computed from nodal internal forces. More... | |
virtual void | computeExternalLoadReactionContribution (FloatArray &reactions, TimeStep *tStep, int di) |
Computes the contribution external loading to reaction forces in given domain. More... | |
virtual void | giveInternalForces (FloatArray &answer, bool normFlag, int di, TimeStep *tStep) |
Evaluates the nodal representation of internal forces by assembling contributions from individual elements. More... | |
void | updateInternalState (TimeStep *tStep) |
Updates nodal values (calls also this->updateDofUnknownsDictionary for updating dofs unknowns dictionaries if model supports changes of static system). More... | |
Protected Member Functions inherited from oofem::EngngModel | |
virtual void | packMigratingData (TimeStep *tStep) |
Packs receiver data when rebalancing load. More... | |
virtual void | unpackMigratingData (TimeStep *tStep) |
Unpacks receiver data when rebalancing load. More... | |
Protected Attributes | |
FloatArray | massMatrix |
Mass matrix. More... | |
FloatArray | loadVector |
Load vector. More... | |
FloatArray | previousIncrementOfDisplacementVector |
Vector storing displacement increments. More... | |
FloatArray | displacementVector |
Displacement, velocity and acceleration vectors. More... | |
FloatArray | velocityVector |
FloatArray | accelerationVector |
FloatArray | internalForces |
Vector of real nodal forces. More... | |
double | dumpingCoef |
Dumping coefficient (C = dumpingCoef * MassMtrx). More... | |
double | deltaT |
Time step. More... | |
int | initFlag |
Flag indicating the need for initialization. More... | |
int | drFlag |
Flag indicating whether dynamic relaxation takes place. More... | |
FloatArray | loadRefVector |
Reference load vector. More... | |
double | c |
Parameter determining rate of the loading process. More... | |
double | pt |
Load level. More... | |
double | Tau |
End of time interval. More... | |
double | pyEstimate |
Estimate of loadRefVector^T*displacementVector(Tau). More... | |
double | pMp |
Product of p^tM^(-1)p; where p is reference load vector. More... | |
SparseMtrx * | massMatrixConsistent |
LinSystSolverType | solverType |
SparseMtrxType | sparseMtrxType |
SparseLinearSystemNM * | nMethod |
Protected Attributes inherited from oofem::StructuralEngngModel | |
StateCounterType | internalVarUpdateStamp |
Contains last time stamp of internal variable update. More... | |
FloatArray | internalForcesEBENorm |
Norm of nodal internal forces evaluated on element by element basis (squared) More... | |
Protected Attributes inherited from oofem::EngngModel | |
int | ndomains |
Number of receiver domains. More... | |
std::vector< std::unique_ptr< Domain > > | domainList |
List of problem domains. More... | |
int | numberOfSteps |
Total number of time steps. More... | |
int | numberOfEquations |
Total number of equation in current time step. More... | |
int | numberOfPrescribedEquations |
Total number or prescribed equations in current time step. More... | |
IntArray | domainNeqs |
Number of equations per domain. More... | |
IntArray | domainPrescribedNeqs |
Number of prescribed equations per domain. More... | |
bool | renumberFlag |
Renumbering flag (renumbers equations after each step, necessary if Dirichlet BCs change). More... | |
bool | profileOpt |
Profile optimized numbering flag (using Sloan's algorithm). More... | |
int | equationNumberingCompleted |
Equation numbering completed flag. More... | |
int | nMetaSteps |
Number of meta steps. More... | |
std::vector< MetaStep > | metaStepList |
List of problem metasteps. More... | |
std::unique_ptr< TimeStep > | stepWhenIcApply |
Solution step when IC (initial conditions) apply. More... | |
std::unique_ptr< TimeStep > | currentStep |
Current time step. More... | |
std::unique_ptr< TimeStep > | previousStep |
Previous time step. More... | |
int | number |
Receivers id. More... | |
std::string | dataOutputFileName |
Path to output stream. More... | |
std::string | coreOutputFileName |
String with core output file name. More... | |
FILE * | outputStream |
Output stream. More... | |
std::string | referenceFileName |
String with reference file name. More... | |
ContextOutputMode | contextOutputMode |
Domain context output mode. More... | |
int | contextOutputStep |
ExportModuleManager * | exportModuleManager |
Export module manager. More... | |
InitModuleManager * | initModuleManager |
Initialization module manager. More... | |
problemMode | pMode |
Domain mode. More... | |
problemScale | pScale |
Multiscale mode. More... | |
time_t | startTime |
Solution start time. More... | |
EngngModel * | master |
Master e-model; if defined receiver is in maintained (slave) mode. More... | |
EngngModelContext * | context |
Context. More... | |
EngngModelTimer | timer |
E-model timer. More... | |
int | parallelFlag |
Flag indicating that the receiver runs in parallel. More... | |
enum fMode | nonLinFormulation |
Type of non linear formulation (total or updated formulation). More... | |
ErrorEstimator * | defaultErrEstimator |
Error estimator. Useful for adaptivity, or simply printing errors output. More... | |
int | rank |
Domain rank in a group of collaborating processes (0..groupSize-1). More... | |
int | numProcs |
Total number of collaborating processes. More... | |
int | nonlocalExt |
Flag indicating if nonlocal extension active, which will cause data to be sent between shared elements before computing the internal forces. More... | |
char | processor_name [PROCESSOR_NAME_LENGTH] |
Processor name. More... | |
MPI_Comm | comm |
Communication object for this engineering model. More... | |
CommunicatorBuff * | commBuff |
Common Communicator buffer. More... | |
ProblemCommunicator * | communicator |
Communicator. More... | |
ProblemCommunicator * | nonlocCommunicator |
NonLocal Communicator. Necessary when nonlocal constitutive models are used. More... | |
std::vector< ParallelContext > | parallelContextList |
List where parallel contexts are stored. More... | |
bool | suppressOutput |
Flag for suppressing output to file. More... | |
std::string | simulationDescription |
LoadBalancer * | lb |
Load Balancer. More... | |
LoadBalancerMonitor * | lbm |
bool | loadBalancingFlag |
If set to true, load balancing is active. More... | |
bool | force_load_rebalance_in_first_step |
Debug flag forcing load balancing after first step. More... | |
Additional Inherited Members | |
Public Types inherited from oofem::EngngModel | |
enum | EngngModel_UpdateMode { EngngModel_Unknown_Mode, EngngModel_SUMM_Mode, EngngModel_SET_Mode } |
enum | EngngModelCommType { PC_default, PC_nonlocal } |
enum | InitialGuess { IG_None = 0, IG_Tangent = 1 } |
Means to choose methods for finding a good initial guess. More... | |
Protected Types inherited from oofem::EngngModel | |
enum | { InternalForcesExchangeTag, MassExchangeTag, LoadExchangeTag, ReactionExchangeTag, RemoteElementExchangeTag } |
Message tags. More... | |
This class implements NonLinear (- may be changed) solution of dynamic problems using Direct Explicit Integration scheme - Central Difference Method.
For efficiency reasons it uses diagonal mass matrix. It is formulated in increments of displacements rather than in total variables.
Solution of this problem is series of loading cases, maintained as sequence of time-steps.
Analysis starts assembling governing equations at time step 0 ( 0 given by boundary and initial cond.) they result in response at time step 1. For time step 0 we need special start code. Because this method is explicit, when solving equations for step t, we obtain solution in step t+dt. But printing is performed for step t. So, when analyst specifies initial conditions, then he/she specifies them in time step 0.
Current implementation supports parallel processing. Both node- and element cut strategies can be used.
Definition at line 88 of file nldeidynamic.h.
oofem::NlDEIDynamic::NlDEIDynamic | ( | int | i, |
EngngModel * | _master = NULL |
||
) |
Definition at line 59 of file nldeidynamic.C.
References initFlag, and oofem::EngngModel::ndomains.
|
virtual |
Definition at line 69 of file nldeidynamic.C.
|
protected |
Assembles the load vector.
If in parallel mode, the loads of shared/remote nodes are exchanged and remote contributions are taken into account.
answer | Load vector. |
mode | Value type mode of load vector. |
tStep | Solution step. |
Definition at line 487 of file nldeidynamic.C.
References oofem::EngngModel::assembleVector(), oofem::EngngModel::giveDomain(), oofem::EngngModel::giveNumberOfDomainEquations(), oofem::EngngModel::LoadExchangeTag, oofem::FloatArray::resize(), oofem::EngngModel::updateSharedDofManagers(), and oofem::FloatArray::zero().
Referenced by giveNumberOfFirstStep(), and solveYourselfAt().
|
protected |
Assembles the diagonal mass matrix of receiver.
Local or Global variant of zero mass elements replacement is performed. If runs in parallel, the masses of shared nodes are exchanged and remote contributions are added accordingly.
mass | Assembled mass matrix. |
maxOm | Estimate of eigenfrequency. |
tStep | Time step. |
Definition at line 506 of file nldeidynamic.C.
References oofem::IntArray::at(), oofem::FloatArray::at(), oofem::FloatMatrix::at(), oofem::Element_remote, oofem::Element::giveCharacteristicMatrix(), oofem::EngngModel::giveDomain(), oofem::Domain::giveElement(), oofem::Element::giveLocationArray(), oofem::EngngModel::giveNumberOfDomainEquations(), oofem::Domain::giveNumberOfElements(), oofem::FloatMatrix::giveNumberOfRows(), oofem::Element::giveParallelMode(), oofem::EngngModel::giveRank(), oofem::Element::giveRotationMatrix(), oofem::IntArray::giveSize(), oofem::FloatMatrix::isNotEmpty(), oofem::EngngModel::MassExchangeTag, oofem::max(), OOFEM_ERROR, OOFEM_WARNING, oofem::FloatArray::resize(), oofem::FloatMatrix::rotatedWith(), oofem::EngngModel::updateSharedDofManagers(), VERBOSEPARALLEL_PRINT, oofem::FloatArray::zero(), and ZERO_REL_MASS.
Referenced by giveNumberOfFirstStep(), and solveYourselfAt().
|
protected |
Referenced by giveNumberOfFirstStep().
|
virtual |
Determines the space necessary for send/receive buffer.
It uses related communication map pattern to determine the maximum size needed.
commMap | Communication map used to send/receive messages. |
buff | Communication buffer. |
packUnpackType | Determines the type of packed quantity, used by receiver to estimate the size of pack/unpack buffer accordingly. |
Reimplemented from oofem::EngngModel.
Definition at line 683 of file nldeidynamic.C.
References oofem::Element::estimatePackSize(), oofem::Domain::giveDofManager(), oofem::EngngModel::giveDomain(), oofem::Domain::giveElement(), oofem::DataStream::givePackSizeOfDouble(), and oofem::max().
Referenced by giveNumberOfFirstStep().
|
inlinevirtual |
Returns class name of the receiver.
Implements oofem::EngngModel.
Definition at line 153 of file nldeidynamic.h.
|
inlinevirtual |
Indicates type of non linear computation (total or updated formulation).
This is used for example on Nodal level to update coordinates if updated formulation is done, or on element level, when non linear contributions are computed.
Reimplemented from oofem::EngngModel.
Definition at line 154 of file nldeidynamic.h.
References oofem::TL.
|
inlinevirtual |
Definition at line 152 of file nldeidynamic.h.
References _IFT_NlDEIDynamic_Name.
|
virtual |
Returns next time step (next to current step) of receiver.
Reimplemented from oofem::EngngModel.
Definition at line 156 of file nldeidynamic.C.
References oofem::EngngModel::currentStep, deltaT, and oofem::EngngModel::previousStep.
|
inlinevirtual |
Returns number of first time step used by receiver.
force | when set to true then receiver reply is returned instead of master (default) |
Reimplemented from oofem::EngngModel.
Definition at line 156 of file nldeidynamic.h.
References computeLoadVector(), computeMassMtrx(), computeMassMtrx2(), and estimateMaxPackSize().
|
virtual |
Returns reference to receiver's numerical method.
Reimplemented from oofem::EngngModel.
Definition at line 72 of file nldeidynamic.C.
References oofem::classFactory, oofem::ClassFactory::createSparseLinSolver(), oofem::EngngModel::giveDomain(), nMethod, and solverType.
|
virtual |
Returns requested unknown.
Unknown at give time step is characterized by its type and mode and by its equation number. This function is used by Dofs, when they are requested for their associated unknowns.
Reimplemented from oofem::EngngModel.
Definition at line 120 of file nldeidynamic.C.
References oofem::Dof::__giveEquationNumber(), accelerationVector, oofem::FloatArray::at(), displacementVector, oofem::EngngModel::giveCurrentStep(), OOFEM_ERROR, previousIncrementOfDisplacementVector, and velocityVector.
|
virtual |
Initializes receiver according to object description in input reader.
InitString can be imagined as data record in component database belonging to receiver. Receiver may use value-name extracting functions to extract particular field from record.
Reimplemented from oofem::EngngModel.
Definition at line 84 of file nldeidynamic.C.
References _IFT_NlDEIDynamic_deltat, _IFT_NlDEIDynamic_drflag, _IFT_NlDEIDynamic_dumpcoef, _IFT_NlDEIDynamic_nonlocalext, _IFT_NlDEIDynamic_py, _IFT_NlDEIDynamic_tau, oofem::EngngModel::commBuff, oofem::EngngModel::communicator, deltaT, drFlag, dumpingCoef, oofem::EngngModel::giveNumberOfProcesses(), oofem::EngngModel::giveRank(), oofem::InputRecord::hasField(), oofem::EngngModel::initializeFrom(), IR_GIVE_FIELD, IR_GIVE_OPTIONAL_FIELD, oofem::IRRT_OK, oofem::EngngModel::nonlocalExt, oofem::EngngModel::nonlocCommunicator, pyEstimate, and Tau.
DOF printing routine.
Called by DofManagers to print Dof specific part. Dof class provides component printing routines, but emodel is responsible for what will be printed at DOF level.
stream | output stream |
iDof | dof to be processed |
tStep | solution step |
Reimplemented from oofem::EngngModel.
Definition at line 777 of file nldeidynamic.C.
References oofem::Dof::printMultipleOutputAt().
|
virtual |
Prints output of receiver to output domain stream, for given time step.
Corresponding function for element gauss points is invoked (gaussPoint::printOutputAt).
Reimplemented from oofem::StructuralEngngModel.
Definition at line 789 of file nldeidynamic.C.
References oofem::OutputManager::doDofManOutput(), oofem::OutputManager::doElementOutput(), drFlag, oofem::EngngModel::giveDomain(), oofem::TimeStep::giveNumber(), oofem::Domain::giveOutputManager(), oofem::TimeStep::giveTargetTime(), oofem::StructuralEngngModel::printReactionForces(), and pt.
|
virtual |
Restores the state of model from output stream.
Restores not only the receiver state, but also same function is invoked for all DofManagers and Elements in associated domain. Note that by restoring element context also contexts of all associated integration points (and material statuses) are restored. Each context is associated with unique time step. Only one context per time step is allowed. Restore context function will restore such context, which is related (through its step number) to time step number and version given in obj parameter. Restoring context will change current time step in order to correspond to newly restored context.
stream | Context file. |
mode | Determines amount of info in stream. |
ContextIOERR | exception if error encountered. |
Reimplemented from oofem::EngngModel.
Definition at line 744 of file nldeidynamic.C.
References accelerationVector, oofem::CIO_IOERR, oofem::CIO_OK, deltaT, displacementVector, previousIncrementOfDisplacementVector, oofem::DataStream::read(), oofem::EngngModel::restoreContext(), oofem::FloatArray::restoreYourself(), THROW_CIOERR, and velocityVector.
|
virtual |
Stores the state of model to output stream.
Stores not only the receiver state, but also same function is invoked for all DofManagers and Elements in associated domain. Note that by storing element context also contexts of all associated integration points (and material statuses) are stored.
stream | Context stream. |
mode | Determines amount of info in stream. |
ContextIOERR | If error encountered. |
Reimplemented from oofem::EngngModel.
Definition at line 712 of file nldeidynamic.C.
References accelerationVector, oofem::CIO_IOERR, oofem::CIO_OK, deltaT, displacementVector, previousIncrementOfDisplacementVector, oofem::EngngModel::saveContext(), oofem::FloatArray::storeYourself(), THROW_CIOERR, velocityVector, and oofem::DataStream::write().
|
virtual |
Starts solution process.
Implementation should invoke for each time step solveYourselfAt function with time step as parameter. Time steps are created using giveNextStep function (this will set current time step to newly created, and updates previous step).
Reimplemented from oofem::EngngModel.
Definition at line 176 of file nldeidynamic.C.
References oofem::EngngModel::giveNumberOfDomainEquations(), oofem::EngngModel::giveRank(), oofem::EngngModel::initializeCommMaps(), oofem::EngngModel::isParallel(), OOFEM_LOG_INFO, and oofem::EngngModel::solveYourself().
|
virtual |
Solves problem for given time step.
Should assemble characteristic matrices and vectors if necessary and solve problem using appropriate numerical method. After finishing solution, this->updateYourself function for updating solution state and then this->terminate function (for updating nodal and element values) should be called.
Reimplemented from oofem::EngngModel.
Definition at line 191 of file nldeidynamic.C.
References accelerationVector, oofem::FloatArray::add(), oofem::FloatArray::at(), c, computeLoadVector(), computeMassMtrx(), deltaT, displacementVector, oofem::DofManager_null, oofem::DofManager_remote, oofem::DofManager_shared, drFlag, dumpingCoef, oofem::Domain::giveDofManager(), oofem::EngngModel::giveDomain(), oofem::StructuralEngngModel::giveInternalForces(), oofem::TimeStep::giveNumber(), oofem::Domain::giveNumberOfDofManagers(), oofem::EngngModel::giveNumberOfDomainEquations(), oofem::DofManager::giveParallelMode(), oofem::DofManager::givePartitionsConnectivitySize(), oofem::TimeStep::giveTargetTime(), oofem::TimeStep::incrementStateCounter(), initFlag, internalForces, oofem::TimeStep::isTheFirstStep(), loadRefVector, loadVector, M_PI, massMatrix, OOFEM_LOG_DEBUG, OOFEM_LOG_RELEVANT, pMp, previousIncrementOfDisplacementVector, pt, pyEstimate, oofem::FloatArray::resize(), oofem::TimeStep::setTimeIncrement(), oofem::FloatArray::subtract(), Tau, velocityVector, and oofem::FloatArray::zero().
|
virtual |
Updates internal state after finishing time step.
(for example total values may be updated according to previously solved increments). Then element values are also updated (together with related integration points and material statuses).
Reimplemented from oofem::StructuralEngngModel.
Definition at line 475 of file nldeidynamic.C.
References oofem::StructuralEngngModel::updateYourself().
|
protected |
Definition at line 98 of file nldeidynamic.h.
Referenced by giveUnknownComponent(), restoreContext(), saveContext(), and solveYourselfAt().
|
protected |
Parameter determining rate of the loading process.
Definition at line 114 of file nldeidynamic.h.
Referenced by solveYourselfAt().
|
protected |
Time step.
Definition at line 104 of file nldeidynamic.h.
Referenced by giveNextStep(), initializeFrom(), restoreContext(), saveContext(), and solveYourselfAt().
|
protected |
Displacement, velocity and acceleration vectors.
Definition at line 98 of file nldeidynamic.h.
Referenced by giveUnknownComponent(), restoreContext(), saveContext(), and solveYourselfAt().
|
protected |
Flag indicating whether dynamic relaxation takes place.
Definition at line 110 of file nldeidynamic.h.
Referenced by initializeFrom(), printOutputAt(), and solveYourselfAt().
|
protected |
Dumping coefficient (C = dumpingCoef * MassMtrx).
Definition at line 102 of file nldeidynamic.h.
Referenced by initializeFrom(), and solveYourselfAt().
|
protected |
Flag indicating the need for initialization.
Definition at line 106 of file nldeidynamic.h.
Referenced by NlDEIDynamic(), and solveYourselfAt().
|
protected |
Vector of real nodal forces.
Definition at line 100 of file nldeidynamic.h.
Referenced by solveYourselfAt().
|
protected |
Reference load vector.
Definition at line 112 of file nldeidynamic.h.
Referenced by solveYourselfAt().
|
protected |
|
protected |
|
protected |
Definition at line 124 of file nldeidynamic.h.
|
protected |
Definition at line 127 of file nldeidynamic.h.
Referenced by giveNumericalMethod().
|
protected |
Product of p^tM^(-1)p; where p is reference load vector.
Definition at line 122 of file nldeidynamic.h.
Referenced by solveYourselfAt().
|
protected |
Vector storing displacement increments.
Definition at line 96 of file nldeidynamic.h.
Referenced by giveUnknownComponent(), restoreContext(), saveContext(), and solveYourselfAt().
|
protected |
Load level.
Definition at line 116 of file nldeidynamic.h.
Referenced by printOutputAt(), and solveYourselfAt().
|
protected |
Estimate of loadRefVector^T*displacementVector(Tau).
Definition at line 120 of file nldeidynamic.h.
Referenced by initializeFrom(), and solveYourselfAt().
|
protected |
Definition at line 125 of file nldeidynamic.h.
Referenced by giveNumericalMethod().
|
protected |
Definition at line 126 of file nldeidynamic.h.
|
protected |
End of time interval.
Definition at line 118 of file nldeidynamic.h.
Referenced by initializeFrom(), and solveYourselfAt().
|
protected |
Definition at line 98 of file nldeidynamic.h.
Referenced by giveUnknownComponent(), restoreContext(), saveContext(), and solveYourselfAt().