Adjoint and Gradient

Generated from SU2 commit 6b54b78464 on 2026-10-23. Do not edit this page. Edit the comment above the option in Common/src/CConfig.cpp instead.

LIMIT_ADJFLOW

Limit value for the adjoint variable.

Type: real number
Default: 1E6
In config_template.cfg: LIMIT_ADJFLOW= 1E6
Mentioned in: Setup Guide: Shape Optimization with the Discrete Adjoint, Unconstrained shape design of a transonic inviscid airfoil at a cte. AoA
Used in 24 test case files, for example: inv_wedge_ROE.cfg (1E15), lam_NACA0012.cfg (1E6)
Code: CConfig.cpp line 2246

MG_ADJFLOW

Multigrid with the adjoint problem. Defualt: YES

More in config_template.cfg: Use multigrid in the adjoint problem (NO, YES)

Type: YES / NO
Default: YES
In config_template.cfg: MG_ADJFLOW= YES
Used in 7 test case files, for example: inv_wedge_ROE.cfg (NO), Arina2KRS.cfg (YES)
Code: CConfig.cpp line 2248

OBJECTIVE_WEIGHT

Adjoint problem boundary condition weights. Applies scaling factor to objective(s)

More in config_template.cfg: List of weighting values when using more than one OBJECTIVE_FUNCTION. Separate by commas and match with MARKER_MONITORING.

Type: list of real numbers
In config_template.cfg: OBJECTIVE_WEIGHT = 1.0
Mentioned in: Setup Guide: Shape Optimization with the Discrete Adjoint, Shape Design With Multiple Objectives and Penalty Functions, Static Conjugate Heat Transfer (CHT)
Used in 41 test case files, for example: inv_wedge_ROE_multiobj.cfg (-1.0E-7,1.0), flow_cylinder.cfg (1.0), fluid.cfg (0.0)
Code: CConfig.cpp line 2251

OBJECTIVE_FUNCTION

Adjoint problem boundary condition

More in config_template.cfg: Objective function in gradient evaluation (DRAG, LIFT, SIDEFORCE, MOMENT_X, MOMENT_Y, MOMENT_Z, EFFICIENCY, BUFFET, EQUIVALENT_AREA, NEARFIELD_PRESSURE, FORCE_X, FORCE_Y, FORCE_Z, THRUST, TORQUE, TOTAL_HEATFLUX, CUSTOM_OBJFUNC MAXIMUM_HEATFLUX, INVERSE_DESIGN_PRESSURE, INVERSE_DESIGN_HEATFLUX, SURFACE_TOTAL_PRESSURE, SURFACE_MASSFLOW, SURFACE_STATIC_PRESSURE, SURFACE_MACH) For a weighted sum of objectives: separate by commas, add OBJECTIVE_WEIGHT and MARKER_MONITORING in matching order.

Type: list of the values below
Values: DRAG, LIFT, SIDEFORCE, EFFICIENCY, INVERSE_DESIGN_PRESSURE, INVERSE_DESIGN_HEATFLUX, MOMENT_X, MOMENT_Y, MOMENT_Z, EQUIVALENT_AREA, NEARFIELD_PRESSURE, FORCE_X, FORCE_Y, FORCE_Z, THRUST, TORQUE, TOTAL_HEATFLUX, MAXIMUM_HEATFLUX, AVG_TEMPERATURE, FIGURE_OF_MERIT, BUFFET, SURFACE_TOTAL_PRESSURE, SURFACE_STATIC_PRESSURE, SURFACE_STATIC_TEMPERATURE, SURFACE_MASSFLOW, SURFACE_MACH, SURFACE_UNIFORMITY, SURFACE_SECONDARY, SURFACE_MOM_DISTORTION, SURFACE_SECOND_OVER_UNIFORM, SURFACE_PRESSURE_DROP, SURFACE_SPECIES_0, SURFACE_SPECIES_VARIANCE, CUSTOM_OBJFUNC, REFERENCE_GEOMETRY, REFERENCE_NODE, VOLUME_FRACTION, TOPOL_DISCRETENESS, TOPOL_COMPLIANCE, STRESS_PENALTY, ENTROPY_GENERATION, TOTAL_PRESSURE_LOSS, KINETIC_ENERGY_LOSS
Default: DRAG_COEFFICIENT
In config_template.cfg: OBJECTIVE_FUNCTION= DRAG
Mentioned in: Changes in Version 8, History and Solution Output, Setup Guide: Shape Optimization with the Discrete Adjoint, Shape Design With Multiple Objectives and Penalty Functions, Static Conjugate Heat Transfer (CHT), Unconstrained shape design of a transonic inviscid airfoil at a cte. AoA
Used in 133 test case files, for example: air_nozzle_restart.cfg (DRAG), inv_wedge_ROE.cfg (SURFACE_TOTAL_PRESSURE), inv_wedge_ROE_multiobj.cfg (SURFACE_TOTAL_PRESSURE, DRAG)
Code: CConfig.cpp line 2253

CUSTOM_OBJFUNC

User-provided definition of a custom objective function.

More in config_template.cfg: Expression used when “OBJECTIVE_FUNCTION= CUSTOM_OBJFUNC”, any history/screen output can be used together with common math functions (sqrt, cos, exp, etc.). This can be used for constraint aggregation (as below) or to compute something SU2 does not, see TestCases/user_defined_functions/.

Type: string
In config_template.cfg: CUSTOM_OBJFUNC= 'DRAG + 10 * pow(fmax(0.4-LIFT, 0), 2)'
Mentioned in: History and Solution Output, Setup Guide: Shape Optimization with the Discrete Adjoint
Used in 5 test case files, for example: configFEA.cfg ('1e2 * (ref_node + TOPOL_COMPLIANCE)'), configFlow.cfg ('1e5 * DRAG'), lam_prem_ch4_hx_ad.cfg (avg_NOx)
Code: CConfig.cpp line 2256

CUSTOM_OUTPUTS

User-provided definitions for custom output.

More in config_template.cfg: User defined functions available on screen and history output. See TestCases/user_defined_functions/.

Type: string
In config_template.cfg: CUSTOM_OUTPUTS= ''
Mentioned in: Aeroacoustic analysis of unsteady pressure, Changes in Version 8, History and Solution Output, Setup Guide: Species Transport and Combustion, Setup Guide: Unsteady Flows and DES, Unsteady von Karman vortex shedding
Used in 6 test case files, for example: configFEA.cfg ('sens_e : Function{1e8 * SENS_E_0};\), lam_prem_ch4_hx_ad.cfg ('avg_NOx : MassFlowAvg{SPECIES[3]}[outlet]'), passive_transport.cfg ('avg_species : MassFlowAvg{SPECIES[0]}[outlet];\)
Code: CConfig.cpp line 2258

DCD_DCL_VALUE

parameter for the definition of a complex objective function

Type: real number
Default: 0.0
In config_template.cfg: DCD_DCL_VALUE= 0.0
Used in 1 test case file, for example: inv_NACA0012_ContAdj.cfg (0.07)
Code: CConfig.cpp line 2261

DCMX_DCL_VALUE

parameter for the definition of a complex objective function

Type: real number
Default: 0.0
In config_template.cfg: DCMX_DCL_VALUE= 0.0
Code: CConfig.cpp line 2263

DCMY_DCL_VALUE

parameter for the definition of a complex objective function

Type: real number
Default: 0.0
In config_template.cfg: DCMY_DCL_VALUE= 0.0
Code: CConfig.cpp line 2265

DCMZ_DCL_VALUE

parameter for the definition of a complex objective function

Type: real number
Default: 0.0
In config_template.cfg: DCMZ_DCL_VALUE= 0.0
Code: CConfig.cpp line 2267

GEO_BOUNDS

Definition of the airfoil section

More in config_template.cfg: Geometrical bounds (Y coordinate) for the wing geometry analysis or fuselage evaluation (X coordinate)

Type: fixed-size list of real numbers
Default: Not extracted; see the code.
In config_template.cfg: GEO_BOUNDS= (1.5, 3.5)
Mentioned in: Setup Guide: Shape Optimization with the Discrete Adjoint, Constrained shape design of a transonic inviscid wing at a cte. C_L
Used in 9 test case files, for example: turb_ONERAM6.cfg ((0.0806, 1.1284)), inv_ONERAM6.cfg ((0, 0.8)), heated_cylinder.cfg ((1.5, 3.5))
Code: CConfig.cpp line 2271

GEO_DESCRIPTION

Identify the body to slice

More in config_template.cfg: Description of the geometry to be analyzed (AIRFOIL, WING)

Type: one of the values below
Values: AIRFOIL, WING, FUSELAGE, NACELLE
Default: WING
In config_template.cfg: GEO_DESCRIPTION= AIRFOIL
Mentioned in: Setup Guide: Shape Optimization with the Discrete Adjoint, Constrained shape design of a transonic inviscid wing at a cte. C_L
Used in 18 test case files, for example: turb_ONERAM6.cfg (WING), naca0012.cfg (AIRFOIL), turb_ahmed.cfg (FUSELAGE)
Code: CConfig.cpp line 2273

GEO_WATERLINE_LOCATION

Z location of the waterline

Type: real number
Default: 0.0
In config_template.cfg: GEO_WATERLINE_LOCATION= 0.0
Code: CConfig.cpp line 2275

GEO_NUMBER_STATIONS

Number of section cuts to make when calculating internal volume

More in config_template.cfg: Number of section cuts to make when calculating wing geometry

Type: integer
Default: 25
In config_template.cfg: GEO_NUMBER_STATIONS= 25
Mentioned in: Constrained shape design of a transonic inviscid wing at a cte. C_L
Used in 4 test case files, for example: naca0012.cfg (25)
Code: CConfig.cpp line 2277

GEO_LOCATION_STATIONS

Definition of the airfoil sections

More in config_template.cfg: Coordinate of the stations to be analyzed

Type: list of real numbers
In config_template.cfg: GEO_LOCATION_STATIONS= (0.0, 0.5, 1.0)
Mentioned in: Setup Guide: Shape Optimization with the Discrete Adjoint, Constrained shape design of a transonic inviscid wing at a cte. C_L
Used in 8 test case files, for example: turb_ONERAM6.cfg ((0.0, 0.2, 0.4, 0.6, 0.8)), heated_cylinder.cfg ((0.0, 0.5, 1.0))
Code: CConfig.cpp line 2279

GEO_NACELLE_LOCATION

Definition of the nacelle location (higlite coordinates, tilt angle, toe angle)

Type: fixed-size list of real numbers
Default: Not extracted; see the code.
In config_template.cfg: GEO_NACELLE_LOCATION= (0.0, 0.0, 0.0, 0.0, 0.0)
Code: CConfig.cpp line 2283

GEO_PLOT_STATIONS

Output sectional forces for specified markers.

More in config_template.cfg: Plot loads and Cp distributions on each airfoil section

Type: YES / NO
Default: NO
In config_template.cfg: GEO_PLOT_STATIONS= NO
Mentioned in: Constrained shape design of a transonic inviscid wing at a cte. C_L
Used in 11 test case files, for example: turb_ONERAM6.cfg (NO), inv_ONERAM6.cfg (YES)
Code: CConfig.cpp line 2285

GEO_MODE

Mode of the GDC code (analysis, or gradient)

More in config_template.cfg: Geometrical evaluation mode (FUNCTION, GRADIENT)

Type: one of the values below
Values: FUNCTION, GRADIENT
Default: FUNCTION
In config_template.cfg: GEO_MODE= FUNCTION
Mentioned in: Constrained shape design of a transonic inviscid wing at a cte. C_L
Used in 20 test case files, for example: turb_ONERAM6.cfg (FUNCTION)
Code: CConfig.cpp line 2287

DRAG_IN_SONICBOOM

Drag weight in sonic boom Objective Function (from 0.0 to 1.0)

Type: real number
Default: 0.0
In config_template.cfg: DRAG_IN_SONICBOOM= 0.0
Used in 1 test case file, for example: NACA64206.cfg (0.0)
Code: CConfig.cpp line 2290

SENS_SMOOTHING

Sensitivity smoothing

Type: one of the values below
Values: NONE, SOBOLEV, BIGRID
Default: NONE
In config_template.cfg: SENS_SMOOTHING= NONE
Used in 6 test case files, for example: turb_ONERAM6.cfg (NONE)
Code: CConfig.cpp line 2292

FROZEN_VISC_CONT

Continuous Adjoint frozen viscosity

Type: YES / NO
Default: YES
In config_template.cfg: % FROZEN_VISC_CONT= NO
Used in 14 test case files, for example: turb_nasa.cfg (YES), HB.cfg (NO)
Code: CConfig.cpp line 2294

FROZEN_VISC_DISC

Discrete Adjoint frozen viscosity

More in config_template.cfg: Frozen the turbulent viscosity in the discrete adjoint formulation (NO, YES)

Type: YES / NO
Default: NO
In config_template.cfg: FROZEN_VISC_DISC= NO
Mentioned in: Setup Guide: Shape Optimization with the Discrete Adjoint
Used in 1 test case file, for example: transonic_stator.cfg (YES)
Code: CConfig.cpp line 2296

FROZEN_LIMITER_DISC

Discrete Adjoint frozen limiter

More in config_template.cfg: Frozen the slope limiter in the discrete adjoint formulation (NO, YES)

Type: YES / NO
Default: NO
In config_template.cfg: FROZEN_LIMITER_DISC= NO
Mentioned in: Setup Guide: Shape Optimization with the Discrete Adjoint, Slope Limiters and Shock Resolution
Code: CConfig.cpp line 2298

INCONSISTENT_DISC

Use an inconsistent (primal/dual) discrete adjoint formulation

More in config_template.cfg: Use an inconsistent spatial integration (primal-dual) in the discrete adjoint formulation. The AD will use the numerical methods in the ADJOINT-FLOW NUMERICAL METHOD DEFINITION section (NO, YES)

Type: YES / NO
Default: NO
In config_template.cfg: INCONSISTENT_DISC= NO
Mentioned in: Setup Guide: Shape Optimization with the Discrete Adjoint, Unconstrained shape design of a transonic inviscid airfoil at a cte. AoA
Used in 2 test case files, for example: inv_cylinder3D.cfg (NO), transonic_stator.cfg (YES)
Code: CConfig.cpp line 2300

FIX_AZIMUTHAL_LINE

Fix an azimuthal line due to misalignments of the near-field (from config_template.cfg)

Type: real number
Default: 90.0
In config_template.cfg: FIX_AZIMUTHAL_LINE= 90.0
Used in 1 test case file, for example: NACA64206.cfg (90.0)
Code: CConfig.cpp line 2302

SENS_REMOVE_SHARP

Remove sharp edges from the sensitivity evaluation Format: SENS_REMOVE_SHARP = YES

More in config_template.cfg: Remove sharp edges from the sensitivity evaluation (NO, YES)

Type: YES / NO
Default: NO
In config_template.cfg: SENS_REMOVE_SHARP = NO
Used in 43 test case files, for example: inv_ONERAM6.cfg (YES), turb_ONERAM6.cfg (NO)
Code: CConfig.cpp line 2305

REORIENT_ELEMENTS

Automatically reorient elements that seem flipped

More in config_template.cfg: Reorient elements based on potential negative volumes (YES/NO)

Type: YES / NO
Default: YES
In config_template.cfg: REORIENT_ELEMENTS= YES
Used in 6 test case files, for example: naca0012.cfg (YES), inv_cylinder3D.cfg (NO)
Code: CConfig.cpp line 2308

Improve this page