|
| float | momentum_flux_ratio (float rho_O, float u_O, float rho_F, float u_F) |
| float | thrust_momentum_ratio (float J, float MR) |
| float | spray_angle_from_J (float J, float k, float n) |
| float | spray_angle_from_TMR (float TMR) |
| float | weber_number (float rho, float u, float d_char, float sigma) |
| float | ohnesorge_number (float mu, float rho, float sigma, float d_or) |
| float | smd_lefebvre (float d_or, float We, float Oh, float C, float m, float p) |
| float | tn4222_property_factor (str transfer, *, float rho_liq, float mu_liq, float sigma, float rho_gas) |
| str | tn4222_transfer_model (smd_cfg) |
| float | smd_impinging_tn4222 (float d_jet, float v_jet, float rho_liq, float mu_liq, float sigma, float rho_gas, *, Optional[float] dv=None, bool scale_properties=True, Optional[str] transfer=None) |
| float | evaporation_constant_m2_s (*, float Tc, float Pc, float rho_g, float rho_l, float L_vap, float T_boil, float cp_g=2200.0, float D0=2.0e-5, float C_evap=1.562) |
| float | tau_evap_from_k (float D32, float k_evap_m2_s) |
| float | spray_axial_velocity (*, float mdot_O, float u_O, float theta_O_deg, float mdot_F, float u_F, float theta_F_deg) |
| float | tau_evap (float D32, float K) |
| float | xstar (float U_rel, float tau_evap) |
| Tuple[bool, List[str]] | check_spray_constraints (float We_O, float We_F, float x_star, SprayConfig config) |
| float | smd_pintle (float L_open, float V_rel, float rho_f, float mu_f, float sigma_f, float C, float B, float n, float p) |
Spray and mixing models (J, We, SMD, x*; TMR and theta are diagnostic only)
| float engine.core.spray.evaporation_constant_m2_s |
( |
* | , |
|
|
float | Tc, |
|
|
float | Pc, |
|
|
float | rho_g, |
|
|
float | rho_l, |
|
|
float | L_vap, |
|
|
float | T_boil, |
|
|
float | cp_g = 2200.0, |
|
|
float | D0 = 2.0e-5, |
|
|
float | C_evap = 1.562 ) |
d²-law evaporation constant k_evap [m²/s], derived from propellant properties.
Replaces a single hardcoded ``K`` that was identical in all 14 configs and for every
propellant. That constant's VALUE was defensible for ethanol at ~3100 K (it implies
k_evap = 3.33e-6 m²/s, within 10% of a T^2 extrapolation of measured ethanol data), but its
SHAPE was wrong: it could not distinguish ethanol from methane from RP-1, and it did not
move with chamber temperature or pressure at all -- so a 200 psi design and a 600 psi design
evaporated at identical rates.
k_evap = C * (8 * rho_g * D_v / rho_l) * ln(1 + B_M)
B_M = cp_g * (Tc - T_boil) / L_vap (Spalding mass transfer number)
D_v = D0 * (Tc/300)^1.75 * (101325/Pc) (vapor diffusivity)
Every input except C and D0 already exists in the propellant presets, so this needs no new
tables. ``C_evap`` is the one calibration constant, anchored so ethanol at 3094 K / 450 psi
reproduces the 3.33e-6 m²/s the old fixed K implied -- i.e. this is a re-shaping of a value
that was already right for the propellant it was tuned on, not a change of magnitude.
The same form is already written (and disabled) in combustion_physics.py under
``SPALDING_DIAGNOSTIC_ENABLED``, where it also hardcodes ``fuel="RP-1"``.
Definition at line 332 of file spray.py.
References max.
| float engine.core.spray.smd_impinging_tn4222 |
( |
float | d_jet, |
|
|
float | v_jet, |
|
|
float | rho_liq, |
|
|
float | mu_liq, |
|
|
float | sigma, |
|
|
float | rho_gas, |
|
|
* | , |
|
|
Optional[float] | dv = None, |
|
|
bool | scale_properties = True, |
|
|
Optional[str] | transfer = None ) |
Sauter mean diameter of an impinging-jet pair, Ingebo, NACA TN 4222 (1958).
Dj / D30 = 2.64 (Dj Vj)^(1/2) + 0.97 Dj dV (Dj, D30 in inches; Vj, dV in ft/s)
D32 = (5 / 3.915) D30 (Nukiyama-Tanasawa, as fitted)
``v_jet`` is the stream's own jet velocity and ``dv`` the jet-to-gas velocity difference; a
face-injected jet enters gas at rest, dV = Vj (TN 4222 eq. 11), the default.
TN 4222 measured n-heptane only. ``transfer`` (see ``tn4222_property_factor``) carries it to
other liquids and gas densities; None means "tn4087" when ``scale_properties`` else "none".
That transfer is an assumption: measure D32 and set ``spray.smd.smd_scale``.
Raises ValueError on a non-positive or non-finite input (it used to return d_jet, a drop the
size of the orifice, without saying so).
Replaces an earlier function that cited this report but ran TN 4087's crossflow form with the
We*Re product on the jets' closing speed, and read its D30 as D32: about half the drop size
TN 4222 measured.
Definition at line 286 of file spray.py.
References tn4222_property_factor().
| float engine.core.spray.smd_pintle |
( |
float | L_open, |
|
|
float | V_rel, |
|
|
float | rho_f, |
|
|
float | mu_f, |
|
|
float | sigma_f, |
|
|
float | C, |
|
|
float | B, |
|
|
float | n, |
|
|
float | p ) |
Calculate Sauter Mean Diameter (SMD) for pintle injector using relative velocity physics.
Formula: SMD = C * L_open * We_rel^(-n) * (1 + B * Oh_f)^p
Parameters:
-----------
L_open : float
Characteristic length (Pintle Opening / Gap Height) [m]
V_rel : float
Relative velocity magnitude between streams [m/s]
rho_f : float
Fuel (sheet) density [kg/m³]
mu_f : float
Fuel (sheet) dynamic viscosity [Pa·s]
sigma_f : float
Fuel surface tension [N/m]
C, B, n, p : float
Correlation constants
Returns:
--------
D32 : float [m]
Definition at line 503 of file spray.py.
| float engine.core.spray.spray_angle_from_J |
( |
float | J, |
|
|
float | k, |
|
|
float | n ) |
Calculate spray angle from momentum flux ratio J.
Uses half-angle tangent form: tan(θ/2) = k × J^n
PHYSICS INTUITION:
------------------
Imagine the spray as a cone coming out of the injector:
1. **Momentum Vector Decomposition**:
- The fuel (RP-1) flows RADIALLY outward (perpendicular to axis)
- The oxidizer (LOX) flows AXIALLY (along the axis, with some angle)
- When they collide, the combined spray has both radial and axial momentum
2. **Why Half-Angle?**:
- The spray is SYMMETRIC about the axis
- We measure angles FROM THE AXIS, not from the opposite edge
- A particle at the edge of the spray travels at angle θ/2 from the axis
- Its momentum vector makes angle θ/2 with the axis (not θ!)
- When we decompose momentum: p_radial = p × sin(θ/2), p_axial = p × cos(θ/2)
- Full angle θ = 2×(θ/2) is just a convention - we use θ/2 in calculations
3. **Why Tangent?**:
- In a conical spray, momentum vectors decompose as:
* Radial component: p_radial = p_total × sin(θ/2)
* Axial component: p_axial = p_total × cos(θ/2)
- The RATIO of radial to axial momentum is:
p_radial / p_axial = sin(θ/2) / cos(θ/2) = tan(θ/2)
- This ratio determines how "spread out" the spray is
4. **Connection to J**:
- J = (ρ_O u_O²) / (ρ_F u_F²) measures relative momentum flux
- High J → more oxidizer momentum → spray more axial (smaller angle)
- Low J → more fuel momentum → spray more radial (larger angle)
- The spray angle emerges from balancing these momentum components
5. **Why Not Direct θ = k×J^n?**:
- Direct form treats angle as arbitrary scalar
- But spray angle comes from VECTOR balance of momenta
- tan(θ/2) captures this vector relationship naturally
- Also provides natural bounds: tan(θ/2) → ∞ as θ → 90°
Parameters:
-----------
J : float
Momentum flux ratio
k : float
Model coefficient
n : float
Model exponent
Returns:
--------
theta : float [rad]
Definition at line 51 of file spray.py.
| float engine.core.spray.spray_axial_velocity |
( |
* | , |
|
|
float | mdot_O, |
|
|
float | u_O, |
|
|
float | theta_O_deg, |
|
|
float | mdot_F, |
|
|
float | u_F, |
|
|
float | theta_F_deg ) |
Axial velocity of the combined spray after the doublet collides [m/s].
Momentum balance on the pair: the two jets arrive canted at theta_O / theta_F from the
chamber axis, and what survives downstream is the momentum-weighted axial component
u_axial = (mdot_O*u_O*cos(theta_O) + mdot_F*u_F*cos(theta_F)) / (mdot_O + mdot_F)
This is the velocity that TRANSPORTS droplets down the chamber, and therefore the one that
belongs in x* = u * tau_evap. The code previously used ``u_rel``, the jet-to-jet RELATIVE
velocity -- correct for atomisation (it is what shears the sheet, and it belongs in the
Ingebo Weber number) but wrong for transport: droplets do not travel downstream at the
closing speed of two jets. Measured on a converged design, u_rel was 28.4 m/s against a true
axial 13.8 m/s, so x* -- and hence the required chamber length -- was over-predicted 2.06x.
Definition at line 382 of file spray.py.
| float engine.core.spray.tn4222_property_factor |
( |
str | transfer, |
|
|
* | , |
|
|
float | rho_liq, |
|
|
float | mu_liq, |
|
|
float | sigma, |
|
|
float | rho_gas ) |
D32(liquid, gas) / D32(n-heptane, air at TN 4222's conditions), same jets and speeds.
- "tn4087": Ingebo's crossflow exponents (NACA TN 4087: D30 ~ (We Re)^(-1/4)), so
(sigma mu_l / rho_l)^(1/4) rho_g^(-1/4).
- "dombrowski_johns": the inviscid sheet-breakup length scale of Dombrowski & Johns (Chem. Eng.
Sci. 18, 1963), D ~ (sigma^2 / (rho_g rho_l U^4))^(1/6) at fixed sheet geometry and speed, so
(sigma^2 / (rho_g rho_l))^(1/6) relative to heptane/air.
- "none": 1 (heptane in air as measured).
The spread between them is the uncertainty of carrying a single-liquid correlation to
LOX/ethanol at chamber density; a measured D32 (smd_scale) replaces it.
Definition at line 251 of file spray.py.
Referenced by smd_impinging_tn4222().
1.15.0