Boundary Layer

Description

Analyze wall-resolved boundary layers.

Parameters

• base: Base

  The input base.

• coordinates: list(str)

  The physical Cartesian coordinate names.

• bl_type: str, default= aerodynamic

  Type of boundary layer:

  • ‘aerodynamic’: aerodynamic boundary layer

  • ‘thermal’: thermal boundary layer

• edge_crit: int, default= 7

  Criteria used for the boundary layer edge detection. Add the integer value associated to each criterion to get the chosen combination.

  • 1. \(dV/ds < 1\%\)

  • 2. Stock & Haase

  • 4. isentropic velocity

  • 8. isentropic Mach number

  if edge_crit includes 8, then the edge results from the closest point from the wall above all criteria used. Otherwise, the edge comes from the average of all edges computed independent from each criteria.

  e.g.: edge_crit = 5 means a mix between the criteria 1 and 4.

• wall_type: str, default= resolved

  Type of wall resolution (‘resolved’ or ‘modeled’).

• mesh_type: str in [‘2D’, ‘2.5D’, ‘3D’], default= ‘3D’

  Dimensionality of the mesh.

• max_bl_pts: int, default= 100

  Maximum number of points in the boundary layer.

• added_bl_pts: int, default= 2

  The number of points inside the boundary layer = number of calculated boundary layer points + added_bl_pts

• smoothing_cycles: int, default= 0

  Number of smoothing cycles for boundary layer results.

• families: dict(str: int) or list(str)

  Surface parts on which the boundary layer is to be computed.

  If the value is a dictionary: The key value is a family name. This family name must be refered by the boundary condition objects. The boundary condition associated to this family name will be processed. The value associated to a key is a marker of the corresponding family. Two families may have the same marker.

  As an example,

  t[‘families’] = {‘WingExt’: 3, ‘Flap’: 4, ‘WingInt’: 3}

  Both ‘WingExt’ and ‘WingInt’ families have the same marker. Then, all corresponding mesh points will be processed alltogether. They will be concatenated in the output base.

  If the value is a list, then a different marker will be assigned to each item of the list.

  As an example,

  t[‘families’] = [‘WingExt’, ‘Flap’, ‘WingInt’]

  Value 0 is special, and applied to all points that do not belong to an input family. So if you give a marker 0 to a family, then you will get all boundary conditions in the ouput base.

• only_profiles: bool, default= False

  Only the profiles selected by profile_points will be calculated.

• profile_points: list(list(float)), default= []

  Profiles will be plotted according to the surface points. If you give the coordinates of a point in the mesh, the boundary layer profile issued from the nearest point to the surface will be computed.

• bl_parameters: dict, default= {}

  Dictionary with keys in [“2D_offset_vector”, “reference_Mach_number”, “reference_velocity”, “reference_pressure”, “reference_density”, “reference_temperature”, “Prandtl_number”, “gas_constant_gamma”, “gas_constant_R”]

• offsurf: int, default= -1

  Distance from the wall. To get values on a surface lying at this distance from the wall.

• eps_Mis: float, default= 0.01

  Threshold on the isentropic Mach number for edge_crit = 8.

• variables: list(str), default= [‘density’, ‘x_velocity’, ‘y_velocity’, ‘z_velocity’, ‘pressure’]

  Names of the input primitive variables (density, velocity components, static pressure).

• nb_zone_passingthrough: int, default= 50

  Maximum number of zone a vector could pass through. When the maximum is reached, the process will stop to propagate the vector (will be cut before the end).

• keep_interpolation_data: bool, default= False

  Allow intermediary results (interpolation coefficients, C objects to remain instantiated in the treatment, in order not to be recomputed in a further call of the same treatement. Warning, the mesh must remain the same.

• instant_selection: int or str, default= 0

  Instant name (or index) to execute in the base.

• change_input_data_type: bool, default= False

  C routines need to have a specific type for the provided variables. Mostly float64 for float, and int32 for int. And C ordered tables. If the provided base does not have the correct type and ordering, a retyped/ordered copy is performed. To save memory, this copy could replace the input variable. This will avoid to keep both tables instantiated at the same time, but may increase the memory used in the initial base (as new copies will mostly use a bigger type size). This copy will only apply to coordinates and retrieved variables.

Preconditions

Zones must be unstructured.

Input flow field values must be dimensionalized. The standard variables named density, x_velocity, y_velocity, z_velocity and pressure must be given in instants.

If edge_crit >= 8, then the variables total_temperature and Mach_number must also be given.

Postconditions

Three output bases.

If you give the coordinates of a point in the mesh, it will compute the boundary layer profile issued from the nearest point to the surface.

A profile is a line issued from a point (mesh point) on a surface that is perpendicular to this surface. The points along the boundary layer profile are computed from the volume grid points. A point along the boundary layer profile is the result of the intersection between the normal and a face of a volume element.

Example

import antares
myt = antares.Treatment('Bl')
myt['base'] = base
myt['coordinates'] = ['points_xc', 'points_yc', 'points_zc']
myt['families'] = {'Slat': 2, 'Flap': 4, 'Main': 3}
myt['mesh_type'] = '2.5D'
myt['max_bl_pts'] = 50
myt['added_bl_pts'] = 2
myt['smoothing_cycles'] = 0
myt['only_profiles'] = False
myt['bl_type'] = 'aerodynamic' # 'aerodynamic' 'thermal'
myt['offsurf'] = 0.5
myt['profile_points'] = [[1.0 , 1.0 , 0.0],]
t['bl_parameters'] = {'2D_offset_vector': 2, 'gas_constant_R': 287.0,
                      'reference_Mach_number': 0.1715, 'reference_velocity': 36.4249853252,
                      'reference_pressure': 101325.0, 'reference_density': 3.14466310931,
                      'reference_temperature': 112.269190122, 'gas_constant_gamma': 1.4,
                      'Prandtl_number': 0.72, }
res_base, profbase, offsurfbase = t.execute()

Main functions

class antares.treatment.codespecific.boundarylayer.TreatmentBl.TreatmentBl

This class is used to analyze boundary layers.

The treatment is divided in many parts. Some of them are developed in C language to get CPU performance. C routines are made available in python with the module ‘ctypes’.

1. read results and user parameters

2. build a volume element connectivity structure

3. compute the surface normals

4. interpolate the flow field values to the surface normals

5. compute the boundary layer edge see file calculate_boundary_layer_edge.c

6. define the local (s, n) coordinate system

7. compute the boundary layer values in the (s, n) coordinate system see file calculate_boundary_layer.c

8. modify the boundary layer edge

9. define the local (s, n) coordinate system

10. compute the boundary layer values in the (s, n) coordinate system

static asvoid(arr)

View the array as dtype np.void (bytes).

Based on https://stackoverflow.com/a/16973510/190597(Jaime, 2013-06) The items along the last axis are viewed as one value. This allows comparisons to be performed which treat entire rows as one value.

compute_constant_data(base)

Compute normal vectors and interpolation coefficients.

compute_interzonetopo(base, cstzone, zone_loc_bnds)

Create additional boundary to consider elements coming from other zone when doing normal vector calculation.

execute()

Analyze the boundary layer.

Returns:

the base containing the results

Return type:

Base

initialization()

Set the constant part.

Only needed once (both global then cstzone)

merge_prism_layer(nb_zone_passingthrough)

Merge prism layer layer between additional vector and main one.

---

Notations:

Indices \(\displaystyle \delta\) and \(\displaystyle \infty\) are used respectively for the boundary layer edge and for the free stream conditions.

‘s’ denotes the coordinate of the streamwise direction, and ‘n’ the coordinate of the normal to the streamwise direction. ‘s’ and ‘n’ form the (s, n) streamline-oriented coordinate system. This coordinate system is curvilinear, orthogonal and attached to the surface.

Boundary layer profile (normal to the wall):

If asked, the profiles are output in a formated ascii file containing:

• a first section (header) with free stream conditions (given as parameters):

Ma	Mach number	\(\displaystyle M_\infty\)
P	Pressure	\(\displaystyle p_\infty\)
Rho	Density	\(\displaystyle \rho_\infty\)
T	Temperature
V	Velocity modulus	\(\displaystyle V_\infty\)
Cp	Pressure coefficient
Ptot	Total pressure
Pdyn	Dynamic pressure

• a section that is repeated as long as there are profiles:

Profile number
requested	Requested coordinates of the profile origin
real	True coordinates of the profile origin. Profiles are computed at surface grid points
Normal vector	Normal unit vector at the profile origin

then come the integral values:

Del	Boundary layer thickness
Dels	Displacement thickness [s-dir]
Deln	Displacement thickness [n-dir]
Delt	Thermal boundary layer thickness approximation
Thtss	Momentum thickness [s-dir]
Thtnn	Momentum thickness [n-dir]
H12	Incompressible shape factor

then the wall values:

FirstLayer height	a
P	Pressure
cp	Pressure coefficient
cfsI, cfnI	Skin friction coefficients [s, n dirs], \(\displaystyle V_\infty\)
cfrI	Resulting skin friction coefficient, \(\displaystyle V_\infty\)
cfsD, cfnD	Skin friction coefficients [s, n dirs], \(\displaystyle V_\delta\)
cfrD	Resulting skin friction coefficient, \(\displaystyle V_\delta\)
Tw	Temperature
Rhow	Density
muew	Dynamic viscosity
nuew	Kinematic viscosity
Taus	Shear stress component [s-dir]
Taun	Shear stress component [n-dir]
Taur	Resulting shear stress

then the edge values:

Prism height	Total height of the prism layer
Vdel	Velocity modulus
Ma	Mach number
ptot	Total pressure
cp	Pressure coefficient

finally, come different values along the profile:

S	Distance to the wall
S/S_edge	adimensionalized distance to the wall
Vx, Vy, Vz	Cartesian velocity components
Vres/V_inf	Velocity modulus normalized by the free stream velocity modulus
Vs/Ve	s normalized velocity component
Vn/Ve	n normalized velocity component
beta	Skew angle
log10(Y+)	logarithm of the nondimensional wall distance
Vr/V	Velocity modulus normalized by the shear stress velocity modulus
Vr/V*_theo	Velocity modulus normalized by the shear stress velocity modulus based on a flow over a flat plate (theory)
Rho	Density
cp_static	Static pressure coefficient
cp_dyn	Dynamic pressure coefficient
Ptot	Total pressure
PtotLoss	Total pressure loss
T	Temperature
Ttot	Total temperature
Trec	to be defined
Vi	Isentropic velocity modulus
Vi-Vr	Isentropic velocity modulus - velocity modulus
MassFlow	Mass flow
tke	Turbulent Kinetic Energy
mu_t/mu_l	ratio of viscosities
mu_t	Eddy viscosity
Mach	Mach number
Mach_is	Isentropic Mach Number

Boundary layer surface:

A surface contains in the mesh can be divided in many parts. A marker is a set of triangular and quadrilateral faces.

All the surfaces of a computation may be defined as families (or markers).

It can return a 2D field on the surface (or selected markers) with the following boundary layer quantities:

X, Y, Z	Cartesian coordinates of the surface point
VecNx, VecNy, VecNz	Cartesian coordinates of the surface normal vector
VxSurf, VySurf, VzSurf	Cartesian velocity components close to the wall [first grid layer] for wall streamlines
VxEdge, VyEdge, VzEdge	Cartesian velocity components at the boundary layer edge for inviscid streamlines
Vs, Vn	s, n velocity components close to the wall
beta_w	Profile skew angle between wall and boundary layer edge in the s, n coordinate system
Y+	Nondimensional wall distance
delta	Boundary layer thickness, \(\displaystyle \delta\)
deltas	Displacement thickness [s-dir] based on the velocity at the boundary layer edge, \(\displaystyle \delta^{*}_s\)
deltan	Displacement thickness [n-dir] based on the velocity at the boundary layer edge, \(\displaystyle \delta^{*}_n\)
deltat	Thermal boundary layer thickness, \(\displaystyle \delta_T\) (approximation)
deltaes	Kinetic energy thickness [s-dir]
deltaen	Kinetic energy thickness [n-dir];
thetass	Momentum thickness [s-dir], \(\displaystyle \theta_{ss}\)
thetann	Momentum thickness [n-dir], \(\displaystyle \theta_{nn}\)
H12	Incompressible shape factor. A shape factor is used in boudary layer flow to determine the nature of the flow. The higher the value of H, the stronger the adverse pressure gradient. A large shape factor is an indicator of a boundary layer near separation. A high adverse pressure gradient can greatly reduce the Reynolds number at which transition into turbulence may occur. Conventionally, H = 2.59 (Blasius boundary layer) is typical of laminar flows, while H = 1.3 - 1.4 is typical of turbulent flows.
Ptot	Total pressure
cp	Pressure coefficient
Ma_edge	Mach number at the boundary layer edge
CfsInf, CfnInf	Skin friction coefficients [s, n dirs], \(\displaystyle V_\infty\)
CfrInf	Resulting skin friction coefficient, \(\displaystyle V_\infty\)
CfsDel, CfnDel	Skin friction coefficient [s, n dirs], \(\displaystyle V_\delta\)
CfrDel	Resulting skin friction coefficient, \(\displaystyle V_\delta\)
maxEddyVisc	Maximum eddy viscosity along the profile
T_wall	Wall temperature
T_edge	Temperature at the boundary layer edge
T_rec	Temperature at the boundary layer edge, TO BE DETAILED
profTag	Profile tag
pointMeshIndex	Index of the point in the array of grid points
heightPrismLayers	Total height of the prism layer
nbPrismLayers	Number of prism layers
BlThickOutPrismThick,	(delta - heightPrismLayers)/delta*100
firstLayerHeight	Height of the first prism layer
ptsExceeded	1 if the number of profile points is close to the maximum number of boundary layer points given as input, NOT YET AVAILABLE
profFound	100 tells that the boundary layer edge is found
Lam/Turb	Transition Laminar/Turbulent
nbBlPts	Number of points in the boundary layer

---

The mathematical definitions are given below:

Displacement thicknesses (based on the velocity at the boundary layer edge):	\(\displaystyle \delta^{*}_s = \int_0^\delta (1-\frac{\rho}{\rho_\delta}\frac{V_s}{V_\delta}) ds\) (deltas), \(\displaystyle \delta^{*}_n = -\int_0^\delta (1-\frac{\rho}{\rho_\delta}\frac{V_n}{V_\delta}) ds\) (deltan)
Momentum thicknesses:	\(\displaystyle \theta_{ss} = \int_0^\delta \frac{\rho}{\rho_\delta}\frac{V_s}{V_\delta}(1-\frac{V_s}{V_\delta}) ds\) (thetass), \(\displaystyle \theta_{nn} = \int_0^\delta \frac{\rho}{\rho_\delta}\frac{V_n}{V_\delta}(1-\frac{V_n}{V_\delta}) ds\) (thetann)
Kinetic energy thicknesses:	\(\displaystyle \theta_{es} = \int_0^\delta \frac{\rho}{\rho_\delta}\frac{V_s}{V_\delta}(1-\frac{V_s^2}{V_\delta^2}) ds\) (deltaes), \(\displaystyle \theta_{en} = \int_0^\delta \frac{\rho}{\rho_\delta}\frac{V_n}{V_\delta}(1-\frac{V_n^2}{V_\delta^2}) ds\) (deltaen)
Shape factors:	\(\displaystyle H12 = \frac{\delta^{*}_s}{\theta_{ss}}\), \(\displaystyle H22 = \frac{\delta^{*}_n}{\theta_{nn}}\)
Pressure coefficients:	\(\displaystyle Cp = \frac{p - p_{\infty}}{\frac{1}{2}\rho_{\infty} V_{\infty}^2}\) (cp_static), \(\displaystyle Kp = \frac{p_{tot} - p_{\infty}}{\frac{1}{2}\rho_{\infty} V_{\infty}^2}\) (cp_dyn)
Skin friction coefficients:	\(\displaystyle Cf_{\delta} = (\mu_w + \mu_t) \frac{{\frac{\partial V}{\partial w}}_{|_w}}{\frac{1}{2}\rho_w V_\delta^2}\), \(\displaystyle Cf_{\infty} = (\mu_w + \mu_t) \frac{{\frac{\partial V}{\partial w}}_{|_w}}{\frac{1}{2}\rho_\infty V_\infty^2}\)
	\(\displaystyle Y^+ = Y \frac{V^\star}{\nu}\)
Velocity normalized by the shear stress velocity:	\(\displaystyle V^+ = \frac{V}{V^\star}\)
Shear stress at the wall:	\(\displaystyle \tau_w = \mu_w \frac{\partial u}{\partial y}\)
Shear stress velocity (friction velocity) at the wall:	\(\displaystyle V^\star = \sqrt{\frac{\tau_w}{\rho_w}}\)
Total pressure loss:	\(\displaystyle \frac{\Delta P_t}{P_t} = \frac{P_t - {P_t}_\infty}{{P_t}_\infty}\) (PtotLoss)
Skew angles:	\(\displaystyle \beta = \gamma - \gamma_\delta\), \(\displaystyle \beta_w = \gamma_w - \gamma_\delta\)

---

Users must keep in mind the following restrictions when analyzing their results.

Restrictions:

Regions:

There are two types of regions where an exact boundary layer edge can not be found:

1. The velocity component Vs becomes very small or zero:

   1. Stagnation point regions

      At a stagnation point, Vs=0. Then, there is no boundary layer thickness. Near stagnation points, it is also difficult to find a boundary layer thickness accurately.

   2. Trailing edge regions

      In thick trailing edge regions, Vs becomes very small.

2. The surface normals do not leave the boundary layer:

   1. Intersection regions (wing-body, body-tail, etc)

      1. The surface normal does not leak from the boundary layer.

      2. The surface normal passes the boundary layer from another component first.

      3. The surface normal hits another surface.

   No accurate boundary layer edge can be found in these cases.

Thickness:

The best situation is when the boundary layer edge lies in the prism layer. In this case, the difference between the true boundary layer edge and the grid point laying on the profile that is upper (or lower) than this true edge is small.

If the boundary layer edge lies outside the prism layer, then the diffence may be more important. Of course, that depends on the mesh quality, but, in general, mesh cell ratio increase quite fast outside the prism layer. In this case, the computed boundary layer thickness can be overestimated, and the step between two successive points on the profile can be large. Then, it is necessary to check the displacement (or momentum) thickness to check the lower pointsand the upper point (respective to the boundary layer edge).

Boundary layer edge:

The edge of the boundary layer is detected with the following multistage algorithm.

A first guess of the boundary layer edge is made with four different methods (see input parameter edge_crit) (function calculate_boundary_layer_edge()):

1. the first derivative of the boundary layer profile velocity is smaller than the 1% and the step between one derivative and the previous one is smaller than the 0.01%. (function edge_velocity_slope())

2. \(\displaystyle \delta = \epsilon y_{max}\) with \(\displaystyle y_{max}\) is the wall distance for which the diagnostic function \(\displaystyle y^a |\frac{\partial u}{\partial y}|^b\) is maximum. The first 10 points are omitted from this maximum search. The constants for a turbulent boundary layer are evaluated from Coles velocity profiles: \(\displaystyle a=1, b=1, \epsilon = 1.936\). The constants for a laminar boundary layer come from quasi-similar compressible solutions including heat transfer effects: \(\displaystyle a=3.9, b=1, \epsilon = 1.294\). The reference is [STOCKHAASE].

4. the relative difference between the isentropic velocity and the velocity of the profile is smaller than a threshold.

   \(\displaystyle V_r = \sqrt{V_x^2 + V_y^2 + V_z^2}\) and \(\displaystyle V_i = V_{\infty} \sqrt{\frac{2}{(\gamma-1)*M_{\infty}^2} *(1-(0.5*\frac{2*(P-P_{\infty})}{\rho_{\infty}*V_{\infty}^2}*\gamma*M_{\infty}^2 + 1)^{(\gamma-1)/\gamma}) + 1}\) and

   The edge is the first point giving \(\displaystyle \frac{| V_r - V_i|}{Vr} < \epsilon_{V_i}\) with \(\displaystyle \epsilon_{V_i} = 2.5 \, 10^{-3}\)

   If none, then the edge is given by \(\displaystyle \min{\frac{| V_r - V_i|}{Vr}}\)

8. hypothesis: the isentropic Mach number is constant in the boundary layer. Comparing with the Mach number gives the edge. \(\displaystyle P_{t_{is}} = P_{ref} (T_t/T_{ref})^{\gamma/(\gamma-1)}\), \(\displaystyle T_{t_{is}} = (P_{t_{is}}/P)^{(\gamma-1)/\gamma}\), \(\displaystyle M_{is} = \sqrt{2 |T_{t_{is}}-1| / (\gamma-1)}\)

   \(\displaystyle \delta M = M_{is} - M\), \(\displaystyle \delta M_{min} = \min(\delta M)\), \(\displaystyle M_{is_{cor}} = M_{is} - \delta M_{min}\).

   The edge is the first point giving \(\displaystyle | M - M_{is_{cor}}| / M_{is_{cor}} < \epsilon_{M_{is}}\) with \(\displaystyle \epsilon_{M_{is}} = 0.01\) by default.

The boundary layer edge is at the position where the first derivative of “V+” becomes approximately zero.

Start defining a range for the edge search. This range, stored in “points”, will be the 160% of the first guess of the boundary layer edge.

The first step is to search the first maximum of the first derivative of “V+” and then, starting from this maximum, search where the first derivative is smaller than two. Once this first edge is found, the next maximum is compared to the first one. If it is bigger, search again the boundary layer edge starting from this new maximum.

When the final edge is found, perform two checks. If the thickness of the boundary layer is greater than two, move the edge back to the last maximun or minimum of the first derivate of “V+” before the distance to the wall is two.

The second check is done using the isentropic velocity. Search for the point where the isentropic velocity meets the profile velocity but only until three points below the current boundary layer edge. If the point is found, average it with the current edge to get the final and corrected boundary layer edge.

Interpolation:

The interpolation of the CFD values to the normal (profile) is done with a polynomial expansion of the form \(\displaystyle G^i = a + b x + c y\), which represents a linear variation of G in both x and y directions within a triangular element. The three constants a, b, and c are computed with the three triangle points. more details can be found in [CHUNG].

Note

Sutherland law for molecular viscosity: \(\displaystyle \mu = \frac{C_1 \sqrt{T_w}}{1+\frac{S}{T_w}}\) with \(\displaystyle C_1 = 1.46 10^{-6} kg.m^{-1}.s^{-1}.K^{-1/2}\) and \(\displaystyle S = 112. K\) (functions Skinfric() of calculate_boundary_layer.c and bl_profile(), clear_values() of calculate_boundary_layer_edge.c)

Note

Function \(\displaystyle U^+\) versus \(\displaystyle y^+\):

linear part: \(\displaystyle U^+ = y^+\) for \(\displaystyle y^+ < 11.63\)

log part: \(\displaystyle U^+ = 5.75 * \log(y^+) + 5.5\) for \(\displaystyle y^+ > 11.63\)

Note

if not given, T_wall is computed with the equation of state of perfect gas: \(\displaystyle T_w = \frac{\gamma}{(\gamma-1)C_p} \frac{P}{\rho}\) where \(\displaystyle C_p = 1,003.41 J.kg^{-1}.K\) is here the specific heat at constant pressure (at T = 250 K), and \(\displaystyle \gamma = \frac{C_p}{C_v}\) given as input, with \(\displaystyle C_v\) the specific heat at constant volume

Note

The boundary layer thickness \(\displaystyle (\delta)\) is the average of a numerical value (distance of the boundary layer edge from the wall) and a correlation value coming from the shape factor (function Thickness() of calculate_boundary_layer.c).

\(\displaystyle \delta = \frac{\delta_{sf}+\delta_{num}}{2}\)

\(\displaystyle \delta_{sf} = \delta^{*I} * (\frac{\theta_s^{I}}{\delta_s^{*I}} H_{1}+1)\) where the superscript I means that the incompressible formula is used and with \(\displaystyle H_{1}\) the mass flow shape factor computed with a correlation of Green [GREEN].

Note

\(\displaystyle \delta_T\) (deltat) is the E. Polhausen approximation of the thermal boundary layer thickness for Prandtl number greater than 0.6

\(\displaystyle \delta_T = \frac{\delta}{{Pr}^{1/3}}\)

Note

Numerical values that are not available as inputs are computed with:

\(\displaystyle M_a = \frac{V}{\sqrt(\gamma*R*T)}\)

\(\displaystyle P_{tot} = (1+.2*M_a*M_a)^{3.5} * P\) (so \(\displaystyle \gamma = 1.4\))

[GREEN]

PROCUREMENT EXECUTIVE MINISTRY OF DEFENCE AERONAUTICAL RESEARCH COUNCIL REPORTS AND MEMORANDA no 3791 Prediction of Turbulent Boundary Layers and Wakes in Compressible Flow by a Lag-Entrainment Method By J. E. GREEN, D. J. WEEKS AND J. W. F. BRODMAN, Aerodynamics Dept., R.A.E, Farnborough

[STOCKHAASE]

Hans W. Stock and Werner Haase. “Feasibility Study of e^N Transition Prediction in Navier-Stokes Methods for Airfoils”, AIAA Journal, Vol. 37, No. 10 (1999), pp. 1187-1196. https://doi.org/10.2514/2.612

[CHUNG]

Computational Fluid Dynamics. T. J. Chung. Cambridge University Press, 2002. pp 273–275

Example

import os
if not os.path.isdir('OUTPUT'):
    os.makedirs('OUTPUT')

import antares

r = antares.Reader('netcdf')
r['filename'] = os.path.join('..', 'data', 'BL', 'EUROLIFT_2D', 'grid1_igs_Local_Sweep.nc')
base = r.read()

r = antares.Reader('netcdf')
r['base'] = base
r['filename'] = os.path.join('..', 'data', 'BL', 'EUROLIFT_2D', 'sol_Local_Sweep_ETW1.pval.30000')
r.read()

t = antares.Treatment('Bl')
t['base'] = base
t['coordinates'] = ['points_xc', 'points_yc', 'points_zc']
t['families'] = {'Slat_2': 2, 'Flap_4': 4, 'Main_3': 3}
t['mesh_type'] = '2.5D'
t['max_bl_pts'] = 50
t['added_bl_pts'] = 2
t['smoothing_cycles'] = 0
t['only_profiles'] = False
t['bl_type'] = 'thermal' # 'aerodynamic' 'thermal'
t['bl_type'] = 'aerodynamic' # 'aerodynamic' 'thermal'
t['offsurf'] = 0.5
t['profile_points'] = [[1.0 , 1.0 , 0.0],
                       [2.0 , 1.0 , 0.0],
                       [3.0 , 1.0 , 0.0],
                       [5.0 , 1.0 , 0.0],
                       [7.0 , 1.0 , 0.0],
                       [9.0 , 1.0 , 0.0],
                       [10.0, 1.0,  0.0],
                       [15.0, 1.0,  0.0]
                      ]
t['bl_parameters'] = {'2D_offset_vector': 2, 'gas_constant_R': 287.0,
                      'reference_Mach_number': 0.1715, 'reference_velocity': 36.4249853252,
                      'reference_pressure': 101325.0, 'reference_density': 3.14466310931,
                      'reference_temperature': 112.269190122, 'gas_constant_gamma': 1.4,
                      'Prandtl_number': 0.72, }
res_base, profbase, offsurfbase = t.execute()

w = antares.Writer('bin_tp')
w['base'] = res_base
w['filename'] = os.path.join('OUTPUT', 'ex_blsurf.plt')
w.dump()

if offsurfbase:
    w = antares.Writer('bin_tp')
    w['base'] = offsurfbase
    w['filename'] = os.path.join('OUTPUT', 'ex_offblsurf.plt')
    w.dump()