/* Code for timestepping with implicit generalized-\alpha method for second order systems. */ #include /*I "petscts.h" I*/ static PetscBool cited = PETSC_FALSE; static const char citation[] = "@article{Chung1993,\n" " title = {A Time Integration Algorithm for Structural Dynamics with Improved Numerical Dissipation: The Generalized-$\\alpha$ Method},\n" " author = {J. Chung, G. M. Hubert},\n" " journal = {ASME Journal of Applied Mechanics},\n" " volume = {60},\n" " number = {2},\n" " pages = {371--375},\n" " year = {1993},\n" " issn = {0021-8936},\n" " doi = {http://dx.doi.org/10.1115/1.2900803}\n}\n"; typedef struct { PetscReal stage_time; PetscReal shift_V; PetscReal shift_A; PetscReal scale_F; Vec X0, Xa, X1; Vec V0, Va, V1; Vec A0, Aa, A1; Vec vec_dot; PetscReal Alpha_m; PetscReal Alpha_f; PetscReal Gamma; PetscReal Beta; PetscInt order; Vec vec_sol_prev; Vec vec_dot_prev; Vec vec_lte_work[2]; TSStepStatus status; TSAlpha2Predictor predictor; void *predictor_ctx; } TS_Alpha; /*@C TSAlpha2SetPredictor - sets the callback for computing a predictor (i.e., initial guess for the nonlinear solver). Input Parameters: + ts - timestepping context . predictor - callback to set the predictor in each step - ctx - the application context, which may be set to `NULL` if not used Level: intermediate Notes: If this function is never called, a same-state-vector predictor will be used, i.e., the initial guess will be the converged solution from the previous time step, without regard for the previous velocity or acceleration. .seealso: [](ch_ts), `TS`, `TSALPHA2`, `TSAlpha2Predictor` @*/ PetscErrorCode TSAlpha2SetPredictor(TS ts, TSAlpha2Predictor predictor, void *ctx) { TS_Alpha *th = (TS_Alpha *)(ts->data); PetscFunctionBegin; th->predictor = predictor; th->predictor_ctx = ctx; PetscFunctionReturn(PETSC_SUCCESS); } static PetscErrorCode TSAlpha_ApplyPredictor(TS ts, Vec X1) { /* Apply a custom predictor if set, or default to same-displacement. */ TS_Alpha *th = (TS_Alpha *)(ts->data); PetscFunctionBegin; if (th->predictor) PetscCall(th->predictor(ts, th->X0, th->V0, th->A0, X1, th->predictor_ctx)); else PetscCall(VecCopy(th->X0, X1)); PetscFunctionReturn(PETSC_SUCCESS); } static PetscErrorCode TSAlpha_StageTime(TS ts) { TS_Alpha *th = (TS_Alpha *)ts->data; PetscReal t = ts->ptime; PetscReal dt = ts->time_step; PetscReal Alpha_m = th->Alpha_m; PetscReal Alpha_f = th->Alpha_f; PetscReal Gamma = th->Gamma; PetscReal Beta = th->Beta; PetscFunctionBegin; th->stage_time = t + Alpha_f * dt; th->shift_V = Gamma / (dt * Beta); th->shift_A = Alpha_m / (Alpha_f * dt * dt * Beta); th->scale_F = 1 / Alpha_f; PetscFunctionReturn(PETSC_SUCCESS); } static PetscErrorCode TSAlpha_StageVecs(TS ts, Vec X) { TS_Alpha *th = (TS_Alpha *)ts->data; Vec X1 = X, V1 = th->V1, A1 = th->A1; Vec Xa = th->Xa, Va = th->Va, Aa = th->Aa; Vec X0 = th->X0, V0 = th->V0, A0 = th->A0; PetscReal dt = ts->time_step; PetscReal Alpha_m = th->Alpha_m; PetscReal Alpha_f = th->Alpha_f; PetscReal Gamma = th->Gamma; PetscReal Beta = th->Beta; PetscFunctionBegin; /* A1 = ... */ PetscCall(VecWAXPY(A1, -1.0, X0, X1)); PetscCall(VecAXPY(A1, -dt, V0)); PetscCall(VecAXPBY(A1, -(1 - 2 * Beta) / (2 * Beta), 1 / (dt * dt * Beta), A0)); /* V1 = ... */ PetscCall(VecWAXPY(V1, (1.0 - Gamma) / Gamma, A0, A1)); PetscCall(VecAYPX(V1, dt * Gamma, V0)); /* Xa = X0 + Alpha_f*(X1-X0) */ PetscCall(VecWAXPY(Xa, -1.0, X0, X1)); PetscCall(VecAYPX(Xa, Alpha_f, X0)); /* Va = V0 + Alpha_f*(V1-V0) */ PetscCall(VecWAXPY(Va, -1.0, V0, V1)); PetscCall(VecAYPX(Va, Alpha_f, V0)); /* Aa = A0 + Alpha_m*(A1-A0) */ PetscCall(VecWAXPY(Aa, -1.0, A0, A1)); PetscCall(VecAYPX(Aa, Alpha_m, A0)); PetscFunctionReturn(PETSC_SUCCESS); } static PetscErrorCode TSAlpha_SNESSolve(TS ts, Vec b, Vec x) { PetscInt nits, lits; PetscFunctionBegin; PetscCall(SNESSolve(ts->snes, b, x)); PetscCall(SNESGetIterationNumber(ts->snes, &nits)); PetscCall(SNESGetLinearSolveIterations(ts->snes, &lits)); ts->snes_its += nits; ts->ksp_its += lits; PetscFunctionReturn(PETSC_SUCCESS); } /* Compute a consistent initial state for the generalized-alpha method. - Solve two successive backward Euler steps with halved time step. - Compute the initial second time derivative using backward differences. - If using adaptivity, estimate the LTE of the initial step. */ static PetscErrorCode TSAlpha_Restart(TS ts, PetscBool *initok) { TS_Alpha *th = (TS_Alpha *)ts->data; PetscReal time_step; PetscReal alpha_m, alpha_f, gamma, beta; Vec X0 = ts->vec_sol, X1, X2 = th->X1; Vec V0 = ts->vec_dot, V1, V2 = th->V1; PetscBool stageok; PetscFunctionBegin; PetscCall(VecDuplicate(X0, &X1)); PetscCall(VecDuplicate(V0, &V1)); /* Setup backward Euler with halved time step */ PetscCall(TSAlpha2GetParams(ts, &alpha_m, &alpha_f, &gamma, &beta)); PetscCall(TSAlpha2SetParams(ts, 1, 1, 1, 0.5)); PetscCall(TSGetTimeStep(ts, &time_step)); ts->time_step = time_step / 2; PetscCall(TSAlpha_StageTime(ts)); th->stage_time = ts->ptime; PetscCall(VecZeroEntries(th->A0)); /* First BE step, (t0,X0,V0) -> (t1,X1,V1) */ th->stage_time += ts->time_step; PetscCall(VecCopy(X0, th->X0)); PetscCall(VecCopy(V0, th->V0)); PetscCall(TSPreStage(ts, th->stage_time)); PetscCall(TSAlpha_ApplyPredictor(ts, X1)); PetscCall(TSAlpha_SNESSolve(ts, NULL, X1)); PetscCall(VecCopy(th->V1, V1)); PetscCall(TSPostStage(ts, th->stage_time, 0, &X1)); PetscCall(TSAdaptCheckStage(ts->adapt, ts, th->stage_time, X1, &stageok)); if (!stageok) goto finally; /* Second BE step, (t1,X1,V1) -> (t2,X2,V2) */ th->stage_time += ts->time_step; PetscCall(VecCopy(X1, th->X0)); PetscCall(VecCopy(V1, th->V0)); PetscCall(TSPreStage(ts, th->stage_time)); PetscCall(TSAlpha_ApplyPredictor(ts, X2)); PetscCall(TSAlpha_SNESSolve(ts, NULL, X2)); PetscCall(VecCopy(th->V1, V2)); PetscCall(TSPostStage(ts, th->stage_time, 0, &X2)); PetscCall(TSAdaptCheckStage(ts->adapt, ts, th->stage_time, X1, &stageok)); if (!stageok) goto finally; /* Compute A0 ~ dV/dt at t0 with backward differences */ PetscCall(VecZeroEntries(th->A0)); PetscCall(VecAXPY(th->A0, -3 / ts->time_step, V0)); PetscCall(VecAXPY(th->A0, +4 / ts->time_step, V1)); PetscCall(VecAXPY(th->A0, -1 / ts->time_step, V2)); /* Rough, lower-order estimate LTE of the initial step */ if (th->vec_lte_work[0]) { PetscCall(VecZeroEntries(th->vec_lte_work[0])); PetscCall(VecAXPY(th->vec_lte_work[0], +2, X2)); PetscCall(VecAXPY(th->vec_lte_work[0], -4, X1)); PetscCall(VecAXPY(th->vec_lte_work[0], +2, X0)); } if (th->vec_lte_work[1]) { PetscCall(VecZeroEntries(th->vec_lte_work[1])); PetscCall(VecAXPY(th->vec_lte_work[1], +2, V2)); PetscCall(VecAXPY(th->vec_lte_work[1], -4, V1)); PetscCall(VecAXPY(th->vec_lte_work[1], +2, V0)); } finally: /* Revert TSAlpha to the initial state (t0,X0,V0), but retain potential time step reduction by factor after failed solve. */ if (initok) *initok = stageok; PetscCall(TSSetTimeStep(ts, 2 * ts->time_step)); PetscCall(TSAlpha2SetParams(ts, alpha_m, alpha_f, gamma, beta)); PetscCall(VecCopy(ts->vec_sol, th->X0)); PetscCall(VecCopy(ts->vec_dot, th->V0)); PetscCall(VecDestroy(&X1)); PetscCall(VecDestroy(&V1)); PetscFunctionReturn(PETSC_SUCCESS); } static PetscErrorCode TSStep_Alpha(TS ts) { TS_Alpha *th = (TS_Alpha *)ts->data; PetscInt rejections = 0; PetscBool stageok, accept = PETSC_TRUE; PetscReal next_time_step = ts->time_step; PetscFunctionBegin; PetscCall(PetscCitationsRegister(citation, &cited)); if (!ts->steprollback) { if (th->vec_sol_prev) PetscCall(VecCopy(th->X0, th->vec_sol_prev)); if (th->vec_dot_prev) PetscCall(VecCopy(th->V0, th->vec_dot_prev)); PetscCall(VecCopy(ts->vec_sol, th->X0)); PetscCall(VecCopy(ts->vec_dot, th->V0)); PetscCall(VecCopy(th->A1, th->A0)); } th->status = TS_STEP_INCOMPLETE; while (!ts->reason && th->status != TS_STEP_COMPLETE) { if (ts->steprestart) { PetscCall(TSAlpha_Restart(ts, &stageok)); if (!stageok) goto reject_step; } PetscCall(TSAlpha_StageTime(ts)); PetscCall(TSAlpha_ApplyPredictor(ts, th->X1)); PetscCall(TSPreStage(ts, th->stage_time)); PetscCall(TSAlpha_SNESSolve(ts, NULL, th->X1)); PetscCall(TSPostStage(ts, th->stage_time, 0, &th->Xa)); PetscCall(TSAdaptCheckStage(ts->adapt, ts, th->stage_time, th->Xa, &stageok)); if (!stageok) goto reject_step; th->status = TS_STEP_PENDING; PetscCall(VecCopy(th->X1, ts->vec_sol)); PetscCall(VecCopy(th->V1, ts->vec_dot)); PetscCall(TSAdaptChoose(ts->adapt, ts, ts->time_step, NULL, &next_time_step, &accept)); th->status = accept ? TS_STEP_COMPLETE : TS_STEP_INCOMPLETE; if (!accept) { PetscCall(VecCopy(th->X0, ts->vec_sol)); PetscCall(VecCopy(th->V0, ts->vec_dot)); ts->time_step = next_time_step; goto reject_step; } ts->ptime += ts->time_step; ts->time_step = next_time_step; break; reject_step: ts->reject++; accept = PETSC_FALSE; if (!ts->reason && ++rejections > ts->max_reject && ts->max_reject >= 0) { ts->reason = TS_DIVERGED_STEP_REJECTED; PetscCall(PetscInfo(ts, "Step=%" PetscInt_FMT ", step rejections %" PetscInt_FMT " greater than current TS allowed, stopping solve\n", ts->steps, rejections)); } } PetscFunctionReturn(PETSC_SUCCESS); } static PetscErrorCode TSEvaluateWLTE_Alpha(TS ts, NormType wnormtype, PetscInt *order, PetscReal *wlte) { TS_Alpha *th = (TS_Alpha *)ts->data; Vec X = th->X1; /* X = solution */ Vec V = th->V1; /* V = solution */ Vec Y = th->vec_lte_work[0]; /* Y = X + LTE */ Vec Z = th->vec_lte_work[1]; /* Z = V + LTE */ PetscReal enormX, enormV, enormXa, enormVa, enormXr, enormVr; PetscFunctionBegin; if (!th->vec_sol_prev) { *wlte = -1; PetscFunctionReturn(PETSC_SUCCESS); } if (!th->vec_dot_prev) { *wlte = -1; PetscFunctionReturn(PETSC_SUCCESS); } if (!th->vec_lte_work[0]) { *wlte = -1; PetscFunctionReturn(PETSC_SUCCESS); } if (!th->vec_lte_work[1]) { *wlte = -1; PetscFunctionReturn(PETSC_SUCCESS); } if (ts->steprestart) { /* th->vec_lte_prev is set to the LTE in TSAlpha_Restart() */ PetscCall(VecAXPY(Y, 1, X)); PetscCall(VecAXPY(Z, 1, V)); } else { /* Compute LTE using backward differences with non-constant time step */ PetscReal h = ts->time_step, h_prev = ts->ptime - ts->ptime_prev; PetscReal a = 1 + h_prev / h; PetscScalar scal[3]; Vec vecX[3], vecV[3]; scal[0] = +1 / a; scal[1] = -1 / (a - 1); scal[2] = +1 / (a * (a - 1)); vecX[0] = th->X1; vecX[1] = th->X0; vecX[2] = th->vec_sol_prev; vecV[0] = th->V1; vecV[1] = th->V0; vecV[2] = th->vec_dot_prev; PetscCall(VecCopy(X, Y)); PetscCall(VecMAXPY(Y, 3, scal, vecX)); PetscCall(VecCopy(V, Z)); PetscCall(VecMAXPY(Z, 3, scal, vecV)); } /* XXX ts->atol and ts->vatol are not appropriate for computing enormV */ PetscCall(TSErrorWeightedNorm(ts, X, Y, wnormtype, &enormX, &enormXa, &enormXr)); PetscCall(TSErrorWeightedNorm(ts, V, Z, wnormtype, &enormV, &enormVa, &enormVr)); if (wnormtype == NORM_2) *wlte = PetscSqrtReal(PetscSqr(enormX) / 2 + PetscSqr(enormV) / 2); else *wlte = PetscMax(enormX, enormV); if (order) *order = 2; PetscFunctionReturn(PETSC_SUCCESS); } static PetscErrorCode TSRollBack_Alpha(TS ts) { TS_Alpha *th = (TS_Alpha *)ts->data; PetscFunctionBegin; PetscCall(VecCopy(th->X0, ts->vec_sol)); PetscCall(VecCopy(th->V0, ts->vec_dot)); PetscFunctionReturn(PETSC_SUCCESS); } /* static PetscErrorCode TSInterpolate_Alpha(TS ts,PetscReal t,Vec X,Vec V) { TS_Alpha *th = (TS_Alpha*)ts->data; PetscReal dt = t - ts->ptime; PetscFunctionBegin; PetscCall(VecCopy(ts->vec_dot,V)); PetscCall(VecAXPY(V,dt*(1-th->Gamma),th->A0)); PetscCall(VecAXPY(V,dt*th->Gamma,th->A1)); PetscCall(VecCopy(ts->vec_sol,X)); PetscCall(VecAXPY(X,dt,V)); PetscCall(VecAXPY(X,dt*dt*((PetscReal)0.5-th->Beta),th->A0)); PetscCall(VecAXPY(X,dt*dt*th->Beta,th->A1)); PetscFunctionReturn(PETSC_SUCCESS); } */ static PetscErrorCode SNESTSFormFunction_Alpha(PETSC_UNUSED SNES snes, Vec X, Vec F, TS ts) { TS_Alpha *th = (TS_Alpha *)ts->data; PetscReal ta = th->stage_time; Vec Xa = th->Xa, Va = th->Va, Aa = th->Aa; PetscFunctionBegin; PetscCall(TSAlpha_StageVecs(ts, X)); /* F = Function(ta,Xa,Va,Aa) */ PetscCall(TSComputeI2Function(ts, ta, Xa, Va, Aa, F)); PetscCall(VecScale(F, th->scale_F)); PetscFunctionReturn(PETSC_SUCCESS); } static PetscErrorCode SNESTSFormJacobian_Alpha(PETSC_UNUSED SNES snes, PETSC_UNUSED Vec X, Mat J, Mat P, TS ts) { TS_Alpha *th = (TS_Alpha *)ts->data; PetscReal ta = th->stage_time; Vec Xa = th->Xa, Va = th->Va, Aa = th->Aa; PetscReal dVdX = th->shift_V, dAdX = th->shift_A; PetscFunctionBegin; /* J,P = Jacobian(ta,Xa,Va,Aa) */ PetscCall(TSComputeI2Jacobian(ts, ta, Xa, Va, Aa, dVdX, dAdX, J, P)); PetscFunctionReturn(PETSC_SUCCESS); } static PetscErrorCode TSReset_Alpha(TS ts) { TS_Alpha *th = (TS_Alpha *)ts->data; PetscFunctionBegin; PetscCall(VecDestroy(&th->X0)); PetscCall(VecDestroy(&th->Xa)); PetscCall(VecDestroy(&th->X1)); PetscCall(VecDestroy(&th->V0)); PetscCall(VecDestroy(&th->Va)); PetscCall(VecDestroy(&th->V1)); PetscCall(VecDestroy(&th->A0)); PetscCall(VecDestroy(&th->Aa)); PetscCall(VecDestroy(&th->A1)); PetscCall(VecDestroy(&th->vec_sol_prev)); PetscCall(VecDestroy(&th->vec_dot_prev)); PetscCall(VecDestroy(&th->vec_lte_work[0])); PetscCall(VecDestroy(&th->vec_lte_work[1])); PetscFunctionReturn(PETSC_SUCCESS); } static PetscErrorCode TSDestroy_Alpha(TS ts) { PetscFunctionBegin; PetscCall(TSReset_Alpha(ts)); PetscCall(PetscFree(ts->data)); PetscCall(PetscObjectComposeFunction((PetscObject)ts, "TSAlpha2SetRadius_C", NULL)); PetscCall(PetscObjectComposeFunction((PetscObject)ts, "TSAlpha2SetParams_C", NULL)); PetscCall(PetscObjectComposeFunction((PetscObject)ts, "TSAlpha2GetParams_C", NULL)); PetscFunctionReturn(PETSC_SUCCESS); } static PetscErrorCode TSSetUp_Alpha(TS ts) { TS_Alpha *th = (TS_Alpha *)ts->data; PetscBool match; PetscFunctionBegin; PetscCall(VecDuplicate(ts->vec_sol, &th->X0)); PetscCall(VecDuplicate(ts->vec_sol, &th->Xa)); PetscCall(VecDuplicate(ts->vec_sol, &th->X1)); PetscCall(VecDuplicate(ts->vec_sol, &th->V0)); PetscCall(VecDuplicate(ts->vec_sol, &th->Va)); PetscCall(VecDuplicate(ts->vec_sol, &th->V1)); PetscCall(VecDuplicate(ts->vec_sol, &th->A0)); PetscCall(VecDuplicate(ts->vec_sol, &th->Aa)); PetscCall(VecDuplicate(ts->vec_sol, &th->A1)); PetscCall(TSGetAdapt(ts, &ts->adapt)); PetscCall(TSAdaptCandidatesClear(ts->adapt)); PetscCall(PetscObjectTypeCompare((PetscObject)ts->adapt, TSADAPTNONE, &match)); if (!match) { PetscCall(VecDuplicate(ts->vec_sol, &th->vec_sol_prev)); PetscCall(VecDuplicate(ts->vec_sol, &th->vec_dot_prev)); PetscCall(VecDuplicate(ts->vec_sol, &th->vec_lte_work[0])); PetscCall(VecDuplicate(ts->vec_sol, &th->vec_lte_work[1])); } PetscCall(TSGetSNES(ts, &ts->snes)); PetscFunctionReturn(PETSC_SUCCESS); } static PetscErrorCode TSSetFromOptions_Alpha(TS ts, PetscOptionItems *PetscOptionsObject) { TS_Alpha *th = (TS_Alpha *)ts->data; PetscFunctionBegin; PetscOptionsHeadBegin(PetscOptionsObject, "Generalized-Alpha ODE solver options"); { PetscBool flg; PetscReal radius = 1; PetscCall(PetscOptionsReal("-ts_alpha_radius", "Spectral radius (high-frequency dissipation)", "TSAlpha2SetRadius", radius, &radius, &flg)); if (flg) PetscCall(TSAlpha2SetRadius(ts, radius)); PetscCall(PetscOptionsReal("-ts_alpha_alpha_m", "Algorithmic parameter alpha_m", "TSAlpha2SetParams", th->Alpha_m, &th->Alpha_m, NULL)); PetscCall(PetscOptionsReal("-ts_alpha_alpha_f", "Algorithmic parameter alpha_f", "TSAlpha2SetParams", th->Alpha_f, &th->Alpha_f, NULL)); PetscCall(PetscOptionsReal("-ts_alpha_gamma", "Algorithmic parameter gamma", "TSAlpha2SetParams", th->Gamma, &th->Gamma, NULL)); PetscCall(PetscOptionsReal("-ts_alpha_beta", "Algorithmic parameter beta", "TSAlpha2SetParams", th->Beta, &th->Beta, NULL)); PetscCall(TSAlpha2SetParams(ts, th->Alpha_m, th->Alpha_f, th->Gamma, th->Beta)); } PetscOptionsHeadEnd(); PetscFunctionReturn(PETSC_SUCCESS); } static PetscErrorCode TSView_Alpha(TS ts, PetscViewer viewer) { TS_Alpha *th = (TS_Alpha *)ts->data; PetscBool iascii; PetscFunctionBegin; PetscCall(PetscObjectTypeCompare((PetscObject)viewer, PETSCVIEWERASCII, &iascii)); if (iascii) PetscCall(PetscViewerASCIIPrintf(viewer, " Alpha_m=%g, Alpha_f=%g, Gamma=%g, Beta=%g\n", (double)th->Alpha_m, (double)th->Alpha_f, (double)th->Gamma, (double)th->Beta)); PetscFunctionReturn(PETSC_SUCCESS); } static PetscErrorCode TSAlpha2SetRadius_Alpha(TS ts, PetscReal radius) { PetscReal alpha_m, alpha_f, gamma, beta; PetscFunctionBegin; PetscCheck(radius >= 0 && radius <= 1, PetscObjectComm((PetscObject)ts), PETSC_ERR_ARG_OUTOFRANGE, "Radius %g not in range [0,1]", (double)radius); alpha_m = (2 - radius) / (1 + radius); alpha_f = 1 / (1 + radius); gamma = (PetscReal)0.5 + alpha_m - alpha_f; beta = (PetscReal)0.5 * (1 + alpha_m - alpha_f); beta *= beta; PetscCall(TSAlpha2SetParams(ts, alpha_m, alpha_f, gamma, beta)); PetscFunctionReturn(PETSC_SUCCESS); } static PetscErrorCode TSAlpha2SetParams_Alpha(TS ts, PetscReal alpha_m, PetscReal alpha_f, PetscReal gamma, PetscReal beta) { TS_Alpha *th = (TS_Alpha *)ts->data; PetscReal tol = 100 * PETSC_MACHINE_EPSILON; PetscReal res = ((PetscReal)0.5 + alpha_m - alpha_f) - gamma; PetscFunctionBegin; th->Alpha_m = alpha_m; th->Alpha_f = alpha_f; th->Gamma = gamma; th->Beta = beta; th->order = (PetscAbsReal(res) < tol) ? 2 : 1; PetscFunctionReturn(PETSC_SUCCESS); } static PetscErrorCode TSAlpha2GetParams_Alpha(TS ts, PetscReal *alpha_m, PetscReal *alpha_f, PetscReal *gamma, PetscReal *beta) { TS_Alpha *th = (TS_Alpha *)ts->data; PetscFunctionBegin; if (alpha_m) *alpha_m = th->Alpha_m; if (alpha_f) *alpha_f = th->Alpha_f; if (gamma) *gamma = th->Gamma; if (beta) *beta = th->Beta; PetscFunctionReturn(PETSC_SUCCESS); } /*MC TSALPHA2 - ODE/DAE solver using the implicit Generalized-Alpha method for second-order systems Level: beginner References: . * - J. Chung, G.M.Hubert. "A Time Integration Algorithm for Structural Dynamics with Improved Numerical Dissipation: The Generalized-alpha Method" ASME Journal of Applied Mechanics, 60, 371:375, 1993. .seealso: [](ch_ts), `TS`, `TSCreate()`, `TSSetType()`, `TSAlpha2SetRadius()`, `TSAlpha2SetParams()` M*/ PETSC_EXTERN PetscErrorCode TSCreate_Alpha2(TS ts) { TS_Alpha *th; PetscFunctionBegin; ts->ops->reset = TSReset_Alpha; ts->ops->destroy = TSDestroy_Alpha; ts->ops->view = TSView_Alpha; ts->ops->setup = TSSetUp_Alpha; ts->ops->setfromoptions = TSSetFromOptions_Alpha; ts->ops->step = TSStep_Alpha; ts->ops->evaluatewlte = TSEvaluateWLTE_Alpha; ts->ops->rollback = TSRollBack_Alpha; /*ts->ops->interpolate = TSInterpolate_Alpha;*/ ts->ops->snesfunction = SNESTSFormFunction_Alpha; ts->ops->snesjacobian = SNESTSFormJacobian_Alpha; ts->default_adapt_type = TSADAPTNONE; ts->usessnes = PETSC_TRUE; PetscCall(PetscNew(&th)); ts->data = (void *)th; th->Alpha_m = 0.5; th->Alpha_f = 0.5; th->Gamma = 0.5; th->Beta = 0.25; th->order = 2; th->predictor = NULL; th->predictor_ctx = NULL; PetscCall(PetscObjectComposeFunction((PetscObject)ts, "TSAlpha2SetRadius_C", TSAlpha2SetRadius_Alpha)); PetscCall(PetscObjectComposeFunction((PetscObject)ts, "TSAlpha2SetParams_C", TSAlpha2SetParams_Alpha)); PetscCall(PetscObjectComposeFunction((PetscObject)ts, "TSAlpha2GetParams_C", TSAlpha2GetParams_Alpha)); PetscFunctionReturn(PETSC_SUCCESS); } /*@ TSAlpha2SetRadius - sets the desired spectral radius of the method for `TSALPHA2` (i.e. high-frequency numerical damping) Logically Collective Input Parameters: + ts - timestepping context - radius - the desired spectral radius Options Database Key: . -ts_alpha_radius - set the desired spectral radius Level: intermediate Notes: The algorithmic parameters $\alpha_m$ and $\alpha_f$ of the generalized-$\alpha$ method can be computed in terms of a specified spectral radius $\rho$ in `[0, 1]` for infinite time step in order to control high-frequency numerical damping\: $$ \alpha_m = (2-\rho)/(1+\rho) \alpha_f = 1/(1+\rho) $$ .seealso: [](ch_ts), `TS`, `TSALPHA2`, `TSAlpha2SetParams()`, `TSAlpha2GetParams()` @*/ PetscErrorCode TSAlpha2SetRadius(TS ts, PetscReal radius) { PetscFunctionBegin; PetscValidHeaderSpecific(ts, TS_CLASSID, 1); PetscValidLogicalCollectiveReal(ts, radius, 2); PetscCheck(radius >= 0 && radius <= 1, ((PetscObject)ts)->comm, PETSC_ERR_ARG_OUTOFRANGE, "Radius %g not in range [0,1]", (double)radius); PetscTryMethod(ts, "TSAlpha2SetRadius_C", (TS, PetscReal), (ts, radius)); PetscFunctionReturn(PETSC_SUCCESS); } /*@ TSAlpha2SetParams - sets the algorithmic parameters for `TSALPHA2` Logically Collective Input Parameters: + ts - timestepping context . alpha_m - algorithmic parameter . alpha_f - algorithmic parameter . gamma - algorithmic parameter - beta - algorithmic parameter Options Database Keys: + -ts_alpha_alpha_m - set alpha_m . -ts_alpha_alpha_f - set alpha_f . -ts_alpha_gamma - set gamma - -ts_alpha_beta - set beta Level: advanced Notes: Second-order accuracy can be obtained so long as\: $$ \gamma = 1/2 + alpha_m - alpha_f \beta = 1/4 (1 + alpha_m - alpha_f)^2 $$ Unconditional stability requires\: $$ \alpha_m >= \alpha_f >= 1/2 $$ Use of this function is normally only required to hack `TSALPHA2` to use a modified integration scheme. Users should call `TSAlpha2SetRadius()` to set the desired spectral radius of the methods (i.e. high-frequency damping) in order so select optimal values for these parameters. .seealso: [](ch_ts), `TS`, `TSALPHA2`, `TSAlpha2SetRadius()`, `TSAlpha2GetParams()` @*/ PetscErrorCode TSAlpha2SetParams(TS ts, PetscReal alpha_m, PetscReal alpha_f, PetscReal gamma, PetscReal beta) { PetscFunctionBegin; PetscValidHeaderSpecific(ts, TS_CLASSID, 1); PetscValidLogicalCollectiveReal(ts, alpha_m, 2); PetscValidLogicalCollectiveReal(ts, alpha_f, 3); PetscValidLogicalCollectiveReal(ts, gamma, 4); PetscValidLogicalCollectiveReal(ts, beta, 5); PetscTryMethod(ts, "TSAlpha2SetParams_C", (TS, PetscReal, PetscReal, PetscReal, PetscReal), (ts, alpha_m, alpha_f, gamma, beta)); PetscFunctionReturn(PETSC_SUCCESS); } /*@ TSAlpha2GetParams - gets the algorithmic parameters for `TSALPHA2` Not Collective Input Parameter: . ts - timestepping context Output Parameters: + alpha_m - algorithmic parameter . alpha_f - algorithmic parameter . gamma - algorithmic parameter - beta - algorithmic parameter Level: advanced Note: Use of this function is normally only required to hack `TSALPHA2` to use a modified integration scheme. Users should call `TSAlpha2SetRadius()` to set the high-frequency damping (i.e. spectral radius of the method) in order so select optimal values for these parameters. .seealso: [](ch_ts), `TS`, `TSALPHA2`, `TSAlpha2SetRadius()`, `TSAlpha2SetParams()` @*/ PetscErrorCode TSAlpha2GetParams(TS ts, PetscReal *alpha_m, PetscReal *alpha_f, PetscReal *gamma, PetscReal *beta) { PetscFunctionBegin; PetscValidHeaderSpecific(ts, TS_CLASSID, 1); if (alpha_m) PetscAssertPointer(alpha_m, 2); if (alpha_f) PetscAssertPointer(alpha_f, 3); if (gamma) PetscAssertPointer(gamma, 4); if (beta) PetscAssertPointer(beta, 5); PetscUseMethod(ts, "TSAlpha2GetParams_C", (TS, PetscReal *, PetscReal *, PetscReal *, PetscReal *), (ts, alpha_m, alpha_f, gamma, beta)); PetscFunctionReturn(PETSC_SUCCESS); }