Contains most of the options for running the code, and how to handle the equations. More...

#include <mushyLayerOpt.h>

Public Attributes
string	output_dir
	Output directory for plot and checkpoint files.

string	plotfile_prefix
	Prefix for plot files.

bool	minimalOutput
	Turn on to only produce a minimal ammount of output.

bool	debug
	Turn on to produce lots of output.

Real	steadyStateCondition
	Maximum value by which the solution may have changed for us to have reached steady state.

bool	ignoreVelocitySteadyState
	Turn on to ignore changes in velocity when deciding if we've reached steady state.

bool	ignoreBulkConcSteadyState
	Turn on to ignore changes in bulk concentration when deciding if we've reached steady state.

Real	domainWidth
	Domain width.

Real	domainHeight
	Domain height.

Real	cfl
	Maximum allowed CFL number.

bool	skipUnsafeCFL
	TODO: remove this.

bool	forceUseUChiForCFL
	Force the code to compute a CFL condition based on the size of $\mathbf{U}/\chi$ , rather than $\mathbf{U}$ .

Real	min_time
	Minimum time to run simulations for (even if the steady state condition has been met)

Real	max_dt_growth
	Maximum allowed fractional increase in dt from one timestep to the next.

Real	init_dt_scale
	Multiply the dt computed during initialisation procedures by this factor before using it. More...

Real	fixedDt
	Enforce a fixed timestep (if > 0)

Real	initial_dt_multiplier
	Scale initial dt's by this factor.

Real	minDt
	Minimum dt allowed.

Real	max_dt
	Maximum dt allowed.

Real	max_init_dt
	Maximum initial dt allowed.

Real	accelCFL
	CFL number for acceleration. Only used if useAccelDt = true. More...

bool	printAccelDt
	Turn on to print some diagnostics about the dt computed according to acceleration considerations.

bool	useAccelDt
	Use the timesteps computed according to the cfl condition applied to accelerations.

bool	useInitAccelDt
	For the first timestep, use dt computed according to the cfl condition applied to accelerations.

int	max_possible_level
	Maximum level allowed in this AMR hierarchy.

bool	computeVorticityStreamFunction
	Turn on to compute the vorticity/streamfunction as a diagnostic.

bool	useFortranRegularisationFace
	Turn on to use fortran routines for regularising the solution on cell faces (i.e. ensuring porosity is not 0) More...

bool	useFortranRegularisation
	Turn on using fortran routines for regularising the solution on cell centres (i.e. ensuring porosity is not 0)

bool	viscousBCs
	Whether or not we want to use viscous boundary conditions.

int	num_init_passes
	Number of loops of the pressure initialisation routine.

Real	restart_new_time
	If >= 0, and we're restarting, set the simulation time to this.

bool	init_use_prev_pressure_for_Ustar
	Whether to use previous estimates of the pressure to compute the initial cell centred $\mathbf{U}^*$ . More...

bool	init_compute_uDelu
	Whether to compute $\mathbf{U} \cdot \nabla \mathbf{U}/\chi$ terms during initialisation. More...

Real	spongeHeight
	Distance from the bottom of the domain that a sponge region will extend.

bool	skipSaltUpdate
	Turn on to set the advective source term in the salt equation to 0.

bool	skipHCUpdate
	Turn on to stop evolving the enthalpy and bulk concentration fields.

bool	doDiffusionSrc
	Use a diffusive source term to compute the advective terms more accurately. More...

Real	rampBuoyancy
	Fractional increase in the buoyancy force with each timestep.

Real	initRaC
	Initial RaC when we're changing the buoyancy force at each timestep.

Real	initRaT
	Initial RaT when we're changing the buoyancy force at each timestep.

Real	maxRaC
	Max RaC when we're changing the buoyancy force at each timestep.

Real	maxRaT
	Max RaT when we're changing the buoyancy force at each timestep.

Real	initAdvVelCentering
	At which time to compute advection velocities during initialisation. More...

Real	adv_vel_centering_growth

bool	ignoreSolveFails
	Whether or not we want to ignore solvers failing. More...

int	solverFailRestartMethod
	What to do if a solver fails and we're not ignoring the problem.

int	steadyStateNormType
	When testing for steady state can consider different norms of the solution. More...

bool	load_advVel
	Whether we want to try and load the advection velocity from a restart file. More...

int	CFinterpOrder_advection
	Whether to do 1st order (set to 1) or 2nd order (set to 2) interpolation at coarse-fine boundaries.

bool	nonlinearHCOpSuperOptimised
	Whether we want to try and do some dodgy stuff to make the enthalpy-bulk concentration nonlinear solver quicker. More...

bool	apply_diagnostic_bcs
	Whether to apply BCs to temperature, porosity etc. explicitly during multigrid solves. More...

bool	useSubcycling
	Whether or not to do subcycling.

int	verbosity
	How much text output to produce on a scale of 0 -> infinity.

bool	useLimiting
	Use slope limiting in advection calculations?

int	tagBufferSize
	Number of cells to add around tagged cells before computing new meshes.

Real	refineThresh
	Refinement threshold.

bool	doEulerPart

bool	computeDiagnostics
	Whether or not to compute diagnostics ad hoc.

bool	doProjection
	Whether or not to do projection.

bool	useIncrementalPressure
	Whether or not to use the previous pressure when solving Darcy's equation. More...

bool	useIncrementalPressureRefinedLevels
	Whether or not to solve for an incremental pressure change on fine levels, rather than the full pressure. More...

bool	doSyncOperations
	Whether or not to do synchronisation operations over multiple levels have reached the same time.

bool	enforceAnalyticSoln
	Whether or not to enforce an analytic solution at the start of the simulation.

int	analyticSolution
	Define the analytic solution to apply. More...

Real	maxDivUFace
	Try and make sure the maximum divergence of the face centred velocity is less than this.

bool	scaleP_MAC
	Whether or not the MAC (face-centred) projection should solve for a pressure that is scaled by the porosity or permeability. More...

bool	scaleP_CC
	Whether or not the CC (cell-centred) projection should solve for a pressure that is scaled by the porosity. More...

bool	usePiAdvectionBCs
	Use the cell centred pressure $\pi$ as the boundary condition when solving for the face centred pressure $\phi$ . More...

int	projection_verbosity
	How much output the projection object should generate.

bool	multiCompUStarSolve
	Solve for all velocity componenents at once. Not currently implemented.

Real	solidPorosity
	Porosity below which we enforce zero velocity. More...

Real	advPorosityLimit
	Porosity below which we ensure advection velocities are 0.

Real	chiLimit
	If > 0, use Darcy's equation to compute velocities in cells with porosity less than this value.

Real	ccVelPorosityLimit
	Porosity below which we ensure cell-centred velocities are 0.

Real	advVelsrcChiLimit
	Porosity below which we ensure that the src term for computing advection velocities is 0.

Real	uDelU_porosityLimit
	Porosity below which we ensure that the $\mathbf{U} \cdot \nabla \mathbf{U}$ part of the cell-centred source term is 0.

Real	advVelChiLimit
	Porosity below which we ensure that velocities are zero when computing $\mathbf{U} \cdot \nabla \mathbf{U}$ .

bool	useOldAdvVel
	Use advection velocity from previous timestep for tracing advection velocities. More...

bool	enforceAnalyticVel
	Whether to enforce an analytic velocity field.

bool	projectAnalyticVel
	Whether to project the enforced analytic velocity field. More...

int	analyticVelType
	If enforcing an analytic velocity, decide what to use.

bool	initAnalyticVel
	Whether or not to initialise the velocity to an analytically specified value.

int	maxProjBaseLevel
	Maximum number of projections to do on levle 0.

int	maxNumMACProj
	Maximum number of MAC (face cented) projections to do.

int	lapVelBCOrder
	Accuracy for boundary conditions applied to laplacian(velocity) in the advection velocity source term.

Real	CCVelSrcTermCentering
	Where to compute the CC velocity source term (default = 0.5 i.e. halfway through the timestep)

int	uDeluMethod

int	uDelU_grow

bool	uDelUConservativeForm

bool	advVelPressureSrc
	Whether or not to include pressure in the advection velocity source term.

bool	advVelDarcySrc
	Whether or not to include the Darcy term in the advection velocity source.

bool	advVelViscousSrc
	Whether or not to include the viscous term in the advection velocity source.

bool	advVelBuoyancySrc
	Whether or not to include buoyancy in the advection velocity source term.

bool	CCAdvSrc
	Whether or not to include $\mathbf{U} \cdot \nabla \mathbf{U}/\chi$ in the cell centred velocity source term.

bool	CCBuoyancySrc
	Whether or not to include buoyancy in the cell centred velocity source term.

bool	CCPressureSrc
	Whether or not to include pressure in the cell centred velocity source term. More...

bool	CCPressureSrcOverride
	Turn to make sure we use MushyLayerOptions::CCPressureSrc to decide whether or not to include pressure. More...

bool	reflux_momentum
	Whether or not to reflux momentum.

bool	reflux_normal_momentum
	Whether or not to just reflux momentum in the normal direction.

bool	reflux_enthalpy
	Whether or not to reflux the enthalpy field.

bool	reflux_concentration
	Whether or not to reflux the bulk concentration field.

bool	reflux_lambda
	Whether or not to reflux lambda. More...

bool	refluxAverageDown
	Whether or not to average down from fine levels to coarser levels after refluxing. More...

RefluxMethod	refluxMethod
	Method to use for implicit refluxing.

Real	refluxBetaSign
	Coefficient of the laplacian operator during implicit reflux solves. More...

Real	refluxCorrSign
	Whether to add (+1) or subtract (-1) the correction. More...

Real	viscous_solver_tol
	Solver tolerance for the viscous solver, which we use for smoothing scalar fields and doing momentum reflux.

int	viscous_num_smooth_down
	Viscous solver - number of smoothing steps during a down sweep.

int	viscous_num_smooth_up
	Viscous solver - number of smoothing steps during an up sweep.

int	AMRMultigridVerb
	Verbosity for the enthalpy-bulk concentration reflux solver.

Real	AMRMultigridTolerance
	Solver tolerance.

Real	AMRMultigridHang
	Solver hang condition.

Real	AMRMultigridNormThresh
	If residual norm is less than this absolute value, solver decides we've converged.

int	velMGNumSmooth
	Multigrid for implicit cell centred $\mathbf{U}^*$ solve - num smoothing steps.

Real	velMGTolerance
	Solver tolerance.

Real	velMGHang
	Solver tolerance.

int	velMGNumMG
	Number of V-cycles to perform.

Real	velMGNormThresh
	Residual norm condition for convergence.

int	VelMGMaxIter
	Max number of multigrid iterations.

int	HCMultigridNumSmoothUp
	Multigrid for the coupled enthalpy-bulk concentration solve: number of smooths on each up sweep.

int	HCMultigridNumSmoothDown
	Number of smooths on each down sweep.

int	HCMultigridNumMG
	Number of V-cycles to do.

int	HCMultigridMaxIter
	Max number of multigrid iterations.

int	HCMultigridVerbosity
	How much output to print.

int	HCMultigridBottomSolveIterations
	Number of iterations of the bottom solver.

Real	HCMultigridTolerance
	Solver tolerance.

Real	HCMultigridHang
	Hang condition.

Real	HCMultigridNormThresh
	Condition on the norm of the residual for convergence.

int	HCMultigridRelaxMode
	How to do relaxation. More...

bool	HCMultigridUseRelaxBottomSolverForHC
	Whether to use a relaxation scheme as the bottom solve. More...

int	velAdvNormalPredOrder
	Order of the normal predictor for velocity advection solves.

bool	velAdvUseFourthOrderSlopes
	Whether to use 4th order slopes for velocity advection solves.

bool	velAdvHigherOrderLimiter
	Whether to use a higher order flux limiter for velocity advection solves.

bool	velAdvUseArtVisc
	Whether to use artifical viscosity for velocity advection solves.

Real	velAdvArtVisc
	Value of artifical viscosity to use for velocity advection solves. More...

bool	HCUseArtVisc
	Whether to use artificial viscosity for enthalpy-bulk concentration advection.

Real	HCArtVisc
	Value of artificial viscosity to use for enthalpy-bulk concentration advection.

int	HCNormalPredOrder
	Order of the normal predictor for enthalpy-bulk concentration advection.

bool	HCUseFourthOrderSlopes
	Whether to use 4th order slopes for enthalpy-bulk concentration advection.

bool	HCHigherOrderLimiter
	Whether to use a higher order flux limiter for enthalpy-bulk concentration advection.

int	taggingVar
	Scalar variable to use for working out where to do refinement.

int	taggingVectorVar
	Vector variable to use for working out where to do refinement.

TaggingMethod	taggingMethod
	Strategy for tagging cells. See TaggingMethod for the options.

Real	min_regrid_time
	Don't start regridding until time > min_regrid_time.

Real	fixed_grid_time
	If this is > 0, then when time > fixed_grid_time we will switch to the grid specified by 'main.regrid_gridfile'.

bool	tagDomainBoundary
	Whether we should tag the domain boundary for refinement.

bool	tagMLboundary
	Whether we should tag the mush-liquid boundary for refinement.

RefinementMethod	refinementMethod
	Method for refinement.

bool	onlyTagPorousCells
	Turn on to only tags cells where other refinement criteria are met and porosity $\chi < 1$ .

int	porousCellsShrink
	Filter porous cells to remove anomalies. More...

bool	tag_velocity
	Whether we should tag cells where the fluid velocity (magnitude) is greater than MushyLayerOptions::vel_thresh.

bool	compositeChannelTagging
	New option for a refinement method based on a composite criteria.

Real	vel_thresh
	If MushyLayerOptions::tag_velocity is true, specify the velocity magnitude criteria here.

bool	tag_plume_mush
	Refine mushy regions on level 1, the refine around plumes according to some empirical criteria. More...

bool	tag_channels
	Just refine around channels.

Real	plumeVelThreshold
	Velocity theshold value for identifying plumes.

Real	plumeSalinityThreshold
	Salinity threshold value for identifying plumes.

Real	taggingMarginalPorosityLimit
	If less than 1, use this limit to identify cells with taggingMarginalPorosityLimit < porosity < 1.0. More...

int	tagCenterBoxSize
	Tag cells in the centre of the domain to make a box of this size. Leave as 0 (default) to do nothing.

Real	tagCenterBoxRegridTime
	If MushyLayerOptions::tagCenterBoxSize > 0, then regrid once time > tagCenterBoxRegridTime.

bool	testRegridCoarsening
	For testing: if turned on, refined levels will only exist for one regrid cycle before being removed. More...

bool	scalarHOinterp
	Whether to do higher order interpolationg for scalars to fill new cells following regridding.

bool	vectorHOinterp
	Whether to do higher order interpolation for vectors to fill new cells following regridding.

bool	makeRegridPlots
	Whether or not to produce lots of plot files during regridding. More...

Real	regrid_dt_scale
	Factor by which to scale dt for re-initialising pressure following regridding.

bool	initLambdaPostRegrid
	Whether or not to recalculate lambda after regridding.

bool	regrid_advect_before_freestream
	Whether or not to advect lambda after regridding, in order to recompute the freestream correction.

bool	regrid_freestream_subcycle
	Whether or not to do subcycling when advecting lambda after regridding. More...

Real	regrid_eta_scale
	Option to scale $\eta$ before recomputing the freestream correction after regridding. More...

bool	do_postRegrid_smoothing
	Whether or not to do some smoothing after regridding.

Real	regrid_smoothing_coeff
	Coefficient to determine how aggressively to do smoothing (if MushyLayerOptions::do_postRegrid_smoothing = true)

bool	project_initial_vel
	Whether or not to project the initial velocity field. More...

bool	initialize_pressures
	Whether or not to initialize pressures.

bool	regrid_init_pressure
	Whether or not to initialize pressure after regridding.

bool	regrid_linear_interp
	Whether to do bi-linear interpolation when regridding.

bool	usePrevPressureForUStar
	Set the initial value for AMRLevelMushyLayer::m_usePrevPressureForUStar. More...

Real	diagnostics_period
	Time period between successive diagnostic reports.

int	customInitData
	Some custom options for initial data.

bool	writePressureInitFields
	Whether to write out plotfiles during pressure initialisation. More...

bool	initResetStates
	When initialising pressure via multiple iterations, decide whether or not to reset the other fields between iterations.

bool	allowMulticompAdvection
	Advect both enthalpy and bulk concentration at the same time (slightly more efficient)

bool	compute_initial_VD_corr
	Whether or not to compute the volume discrepancy correction at the start of a simulation.

int	timeIntegrationOrder
	Order of accuracy for time integration (1 = backward euler, 2 = TGA)

Real	lowerPorosityLimit
	How much output multigrid operators should print. More...

Real	initialPerturbation
	Size of the initial perturbation (default = 0.0)

bool	doScalarAdvectionDiffusion
	Whether or not we should do advection and diffusion of scalars (enthalpy, bulk concentration)

Real	initVel
	Size of initial velocity if required.

Real	perturbationPhaseShift
	Phase shift for sinusoidal perturbations.

Real	delayedPerturbation = 0.0
	Size of perturbation to add at time MushyLayerOptions::perturbationTime.

Real	perturbationTime = 0.0
	Specific time at which to introduce a perturbation.

Real	perturbationWavenumber
	Wavenumber of sinusoidal perturbations (default = 0.0, i.e. just a uniform value)

bool	perturbationSin = false
	If true, sinusoidal perturbations take the form of a sin wave. Otherwise a cosine wave. More...

Real	initialRandomPerturbation = 0.0
	Add an initial random perturbation to enthalpy.

bool	seedRandomPert = true

Real	fixedPorosity = -1.0
	Porosity value to set if we're using a fixed porosity. More...

Real	porosityTimescale
	Timescale for porosity variations if enforcing a fixed porosity.

PorosityFunctions	porosityFunction
	Functional form to use for the porosity if using a fixed porosity (rather than a porosity computed from the solution).

int	maxRestartWavenumbers
	When restarting, if MushyLayerOptions::restartPerturbation > 0.0 then add a perturbation which if the sum of waves with wavenumbers from 1->maxRestartWavenumber.

Real	restartPerturbation
	Size of perturbation to add when restarting a simulation.

Real	porousHoleRadius
	Create a porous hole of this radius during initialisation.

Real	initVelScale
	Velocity scale for the initial conditions.

bool	linearGradient
	Whether to initialise with a linear enthalpy gradient in the vertical direction.

Real	mushHeight
	Approximate depth of mushy layer to create as part of an initial condition approximating some sea ice.

int	summerProfile
	Different options for initialising data that looks a bit like year old sea ice. More...

Real	meltPondDepth
	Depth of melt pond to add.

bool	horizAverageRestart
	Whether or not to horizontally average all fields before restarting.

int	restart_perturbation_var
	When restarting, add a perturbation to this variable.

Real	refTemp
	Reference temperature for nondimensionalisation.

Real	refSalinity
	Reference salinity for nondimensionalisation.

Real	prevRefSalinity
	Previous reference salinity for nondimensionalisation (if restarting and different from the new value)

Real	prevRefTemp
	Previous reference temperature for nondimensionalisation (if restarting and different from the new value)

Real	dtReductionFactor
	Facotr by which to reduce dt after restarting: .

Real	fixedPorosityMaxChi
	Fixed porosity maximum value.

Real	FixedPorositySTD
	Standard deviation of the gaussian used to create a fixed porosity field.

Real	fixedPorosityFractionalInnerRadius
	For a fixed porosity which varies linearly from some constant value inside the domain to the edge, this is the radius of the inner region.

Real	fixedPorosityEndTime
	If using a time dependent enforced porosity, stop changing it after this time has passed.

bool	scalarExchangeCorners
	Whether to exchange corner cells by default for scalar fields.

Real	buoyancy_zero_time
	Set buoyancy forces to zero after this time.

MGmethod	MGtype
	If this is set to greater than 0,. More...

velocityAdvectionTypes	advectionMethod
	Specify how to treat the $\mathbf{U} \cdot \nabla \left( \mathbf{U}/\chi \right)$ term.

bool	includeTracers
	Whether to include tracer dynamics.

Real	surfaceIrradiance
	If > 0, compute penetration of irradiance down through the ice.

Detailed Description

Contains most of the options for running the code, and how to handle the equations.

Member Data Documentation

◆ accelCFL

Real MushyLayerOptions::accelCFL

CFL number for acceleration. Only used if useAccelDt = true.

When velocities are changing rapidly due to a large acceleration, the timestep computed based on the current velocities may not be a appropriate. Instead, we should estimate the new velocities given the acceleration:

$\Delta t = \sqrt{\sigma \Delta x / a}$ ,

where $\sigma$ is the cfl number and $ a$ is the maximum acceleration.

◆ adv_vel_centering_growth

Real MushyLayerOptions::adv_vel_centering_growth

When computing the advection velocity during multiple initialisation steps, grow MushyLayerOptions::initAdvVelCentering by this factor with each iteration

◆ analyticSolution

int MushyLayerOptions::analyticSolution

Define the analytic solution to apply.

By default, analyticSolution = MushyLayerParams::physicalProblem

◆ apply_diagnostic_bcs

bool MushyLayerOptions::apply_diagnostic_bcs

Whether to apply BCs to temperature, porosity etc. explicitly during multigrid solves.

If false, use BC comptued from enthalpy/bulk concentration

◆ CCPressureSrc

bool MushyLayerOptions::CCPressureSrc

Whether or not to include pressure in the cell centred velocity source term.

Only matters if MushylayerOptions::CCPressureSrcOverride is turned on

◆ CCPressureSrcOverride

bool MushyLayerOptions::CCPressureSrcOverride

Turn to make sure we use MushyLayerOptions::CCPressureSrc to decide whether or not to include pressure.

Otherwise, we will use AMRLevelMushyLayer::m_usePrevPressureForUStar. This is turned on by default if MushyLayerOptions::CCPressureSrc is defined through the inputs file

◆ doDiffusionSrc

bool MushyLayerOptions::doDiffusionSrc

Use a diffusive source term to compute the advective terms more accurately.

When computing things like $\mathbf{U} \cdot \nabla T$ , we upwind temperature to cell faces. As the temperature also obeys diffuses, we can add a diffusive source term to estimate the upwinded temperatures more accurately. Similarly for other advected and diffused quantities.

◆ fixedPorosity

Real MushyLayerOptions::fixedPorosity = -1.0

Porosity value to set if we're using a fixed porosity.

If < 0, use some functional form for the porosity as defined by MushyLayerOptions::porosityFunction

◆ HCMultigridRelaxMode

int MushyLayerOptions::HCMultigridRelaxMode

How to do relaxation.

See AMRPoissonOp::relax in Chombo for the various options. In particular, 1 = Gauss-Seidel Red-Black, 4 = Jacobi

◆ HCMultigridUseRelaxBottomSolverForHC

bool MushyLayerOptions::HCMultigridUseRelaxBottomSolverForHC

Whether to use a relaxation scheme as the bottom solve.

We use this here as it treats the nonlinearity more sensibly than linear solvers like BiCGStab

◆ ignoreSolveFails

bool MushyLayerOptions::ignoreSolveFails

Whether or not we want to ignore solvers failing.

If false, we'll try and do some thing about the situation as dictated by MushyLayerOptions::solverFailRestartMethod

◆ init_compute_uDelu

bool MushyLayerOptions::init_compute_uDelu

Whether to compute $\mathbf{U} \cdot \nabla \mathbf{U}/\chi$ terms during initialisation.

Turned on by default

◆ init_dt_scale

Real MushyLayerOptions::init_dt_scale

Multiply the dt computed during initialisation procedures by this factor before using it.

Useful for using a smaller dt initially for stability TODO: can probably remove this, it's just a duplicate of initial_dt_multiplier really

◆ init_use_prev_pressure_for_Ustar

bool MushyLayerOptions::init_use_prev_pressure_for_Ustar

Whether to use previous estimates of the pressure to compute the initial cell centred $\mathbf{U}^*$ .

Turned off by default

◆ initAdvVelCentering

Real MushyLayerOptions::initAdvVelCentering

At which time to compute advection velocities during initialisation.

Default is 0.1, i.e. at time 0 + 0.1*dt

◆ load_advVel

bool MushyLayerOptions::load_advVel

Whether we want to try and load the advection velocity from a restart file.

Turned off by default, as this won't always work with some of the restart files we generate (e.g. after refinement/adding new cells). This is only really useful for when we want to do post processing without recomputing velocity fields.

◆ lowerPorosityLimit

Real MushyLayerOptions::lowerPorosityLimit

How much output multigrid operators should print.

We don't let the porosity go below this value

◆ makeRegridPlots

bool MushyLayerOptions::makeRegridPlots

Whether or not to produce lots of plot files during regridding.

Useful for debugging

◆ MGtype

MGmethod MushyLayerOptions::MGtype

If this is set to greater than 0,.

Whether to use linear or FAS geometric multigrid Should really be FAS as we're solving nonlinear problems, but leaving the option here to change back to linear multigrid.

◆ nonlinearHCOpSuperOptimised

bool MushyLayerOptions::nonlinearHCOpSuperOptimised

Whether we want to try and do some dodgy stuff to make the enthalpy-bulk concentration nonlinear solver quicker.

Turned off by default, use with extreme caution.

◆ perturbationSin

bool MushyLayerOptions::perturbationSin = false

If true, sinusoidal perturbations take the form of a sin wave. Otherwise a cosine wave.

E.g, perturbation $= \alpha \sin (2 \pi N x / L)$ where $\alpha$ is MushyLayerOptions::initialPerturbation, $ N $ is MushyLayerOptions::perturbationWavenumber, and $ L $ is the domain width

◆ porousCellsShrink

int MushyLayerOptions::porousCellsShrink

Filter porous cells to remove anomalies.

Select all cells with $\chi < 1$ , then shift vertically by this value and tag only the intersection of the two

◆ project_initial_vel

bool MushyLayerOptions::project_initial_vel

Whether or not to project the initial velocity field.

Projecting ensures it is divergence free

◆ projectAnalyticVel

bool MushyLayerOptions::projectAnalyticVel

Whether to project the enforced analytic velocity field.

Useful for testing if the projection works as expected

◆ reflux_lambda

bool MushyLayerOptions::reflux_lambda

Whether or not to reflux lambda.

If you don't do this, you won't be able to calculate the freestream preservation correction.

◆ refluxAverageDown

bool MushyLayerOptions::refluxAverageDown

Whether or not to average down from fine levels to coarser levels after refluxing.

As the solution will have changed following refluxing, doing this ensures consistency between levels

◆ refluxBetaSign

Real MushyLayerOptions::refluxBetaSign

Coefficient of the laplacian operator during implicit reflux solves.

This might have to change depending on the operator being used, so left as an option.

◆ refluxCorrSign

Real MushyLayerOptions::refluxCorrSign

Whether to add (+1) or subtract (-1) the correction.

This might need to be changed depending on the operator being used, so left as an option.

◆ regrid_eta_scale

Real MushyLayerOptions::regrid_eta_scale

Option to scale $\eta$ before recomputing the freestream correction after regridding.

Default value = 1.0.

◆ regrid_freestream_subcycle

bool MushyLayerOptions::regrid_freestream_subcycle

Whether or not to do subcycling when advecting lambda after regridding.

Only used if MushyLayerOptions::regrid_advect_before_freestream is true

◆ scaleP_CC

bool MushyLayerOptions::scaleP_CC

Whether or not the CC (cell-centred) projection should solve for a pressure that is scaled by the porosity.

If scaled, we solve something like $\nabla \cdot \chi \nabla \pi = \nabla \cdot \mathbf{U}^*$ , else we'll solve $\nabla^2 \pi = \nabla \cdot \mathbf{U}^*$ .

◆ scaleP_MAC

bool MushyLayerOptions::scaleP_MAC

Whether or not the MAC (face-centred) projection should solve for a pressure that is scaled by the porosity or permeability.

If scaled, we solve something like $\nabla \cdot \chi \nabla \phi = \nabla \cdot \mathbf{U}^*$ , else we'll solve $\nabla^2 \phi = \nabla \cdot \mathbf{U}^*$ .

◆ seedRandomPert

bool MushyLayerOptions::seedRandomPert = true

Turn on to seed random perturbations on different boxes to prevent the same 'random' number patterns in each box

◆ solidPorosity

Real MushyLayerOptions::solidPorosity

Porosity below which we enforce zero velocity.

This needs to be greater than 0

◆ steadyStateNormType

int MushyLayerOptions::steadyStateNormType

When testing for steady state can consider different norms of the solution.

0 -> compute the max value 1 -> compute the L1 norm of the solution (i.e. some of absolute values) 2 -> compute the L2 norm of the solution (i.e. some of values squared)

◆ summerProfile

int MushyLayerOptions::summerProfile

Different options for initialising data that looks a bit like year old sea ice.

See AMRLevelMushyLayer::initialDataMushyLayer() for more details

◆ tag_plume_mush

bool MushyLayerOptions::tag_plume_mush

Refine mushy regions on level 1, the refine around plumes according to some empirical criteria.

Specifically, identify plumes as regions where either a) the downwards velocity is greater than MushyLayerOptions::plumeVelThreshold, and b) the bulk concentration is greater than MushyLayerOptions::plumeSalinityThreshold

◆ taggingMarginalPorosityLimit

Real MushyLayerOptions::taggingMarginalPorosityLimit

If less than 1, use this limit to identify cells with taggingMarginalPorosityLimit < porosity < 1.0.

These cells are then not considered part of the mushy layer for tagging purposes, but also not liquid.

◆ testRegridCoarsening

bool MushyLayerOptions::testRegridCoarsening

For testing: if turned on, refined levels will only exist for one regrid cycle before being removed.

This let's us test levels being created then destroyed in a controlled manner

◆ useFortranRegularisationFace

bool MushyLayerOptions::useFortranRegularisationFace

Turn on to use fortran routines for regularising the solution on cell faces (i.e. ensuring porosity is not 0)

Turned off by default as doesn't work yet.

◆ useIncrementalPressure

bool MushyLayerOptions::useIncrementalPressure

Whether or not to use the previous pressure when solving Darcy's equation.

If true, we then only solve for a small extra pressure correction which is usually a lot quicker

◆ useIncrementalPressureRefinedLevels

bool MushyLayerOptions::useIncrementalPressureRefinedLevels

Whether or not to solve for an incremental pressure change on fine levels, rather than the full pressure.

This may not work fully yet...

◆ useOldAdvVel

bool MushyLayerOptions::useOldAdvVel

Use advection velocity from previous timestep for tracing advection velocities.

Default is false, in which case we average the cell centred velocities from the previous timestep to cell faces and use these (as they are calculated at a later time, they should be more accurate).

◆ usePiAdvectionBCs

bool MushyLayerOptions::usePiAdvectionBCs

Use the cell centred pressure $\pi$ as the boundary condition when solving for the face centred pressure $\phi$ .

Default is true.

◆ usePrevPressureForUStar

bool MushyLayerOptions::usePrevPressureForUStar

Set the initial value for AMRLevelMushyLayer::m_usePrevPressureForUStar.

Determines if we should use the previous $\nabla P$ to calculate $\mathbf{u}^*$

◆ velAdvArtVisc

Real MushyLayerOptions::velAdvArtVisc

Value of artifical viscosity to use for velocity advection solves.

Only used if MushylayerOptions::velAdvUseArtVisc is turned on

◆ writePressureInitFields

bool MushyLayerOptions::writePressureInitFields

Whether to write out plotfiles during pressure initialisation.