Imported from ashfordeOU/aero-agent-skills (skills/aerodynamics/high-speed/shock-tube/SKILL.md). Install upstream with npx skills add ashfordeOU/aero-agent-skills --skill shock-tube. Copyright stays with the author (Apache-2.0).

Shock-Tube Wave System (aerodynamics/high-speed/shock-tube)

Use when the task is the classical one-dimensional shock-tube problem: a high-pressure driver gas (region 4) separated by a diaphragm from a low-pressure driven gas (region 1), both initially at rest, with the diaphragm burst releasing an incident shock into the driven gas, a contact surface between the two gases, and a centered expansion wave running back into the driver gas. This leaf recovers the incident-shock Mach number from the diaphragm pressure state alone, then resolves the four-region state. It pairs with normal-shock for the stationary-shock ratio context at a given upstream Mach number and with prandtl-meyer for the steady turning-fan context; the moving-shock, contact-surface and unsteady-wave content here has no other home in the high-speed pack. Pure stdlib, deterministic, offline.

Domain quick reference

• Incident shock into gas at rest at shock Mach Ms: p2/p1 = 1 + 2 gamma1 (Ms^2 - 1) / (gamma1 + 1), rho2/rho1 = (gamma1 + 1) Ms^2 / (2 + (gamma1 - 1) Ms^2), T2/T1 = (p2/p1) / (rho2/rho1); the shock speed is Ws = Ms a1 with a1 = sqrt(gamma1 R t1).
• Induced flow behind the shock (contact-surface velocity): u2 = 2 a1 (Ms - 1/Ms) / (gamma1 + 1). The shock-frame downstream Mach number Mn2 = sqrt((1 + (gamma1 - 1) Ms^2 / 2) / (gamma1 Ms^2 - (gamma1 - 1) / 2)) stays below one; u2 = Ms a1 - Mn2 a2 links the frames.
• Unsteady centered expansion into the driver gas: p3/p4 = (1 - (gamma4 - 1) u3 / (2 a4)) ** (2 gamma4 / (gamma4 - 1)), T3/T4 = (1 - (gamma4 - 1) u3 / (2 a4))^2, rho3/rho4 = (p3/p4) / (T3/T4).
• Implicit shock-tube equation (the diaphragm match p2 = p3): the normal-shock ratio p2/p1 at Ms equals p4/p1 times the expansion ratio at the shared contact velocity, with u2/a4 = 2 (Ms - 1/Ms) / ((gamma1 + 1) a4/a1). The residual is strictly increasing in Ms, so the root is unique.
• Expansion bracket ceiling: the driver expansion stays physical only while (Ms - 1/Ms) < C = (gamma1 + 1) (a4/a1) / (gamma4 - 1), which caps Ms at Ms_lim = (C + sqrt(C^2 + 4)) / 2 (6.162277660 for air/air at a4/a1 = 1).
• Both fan boundaries are Mach waves: the head runs leftward at speed a4 into the quiescent driver gas and the tail runs at u3 - a3 in the lab frame, right-positive.
• Absolute driver state: same specific gas constant R for both gases (documented assumption), so T4 = t1 (a4/a1)^2 gamma1 / gamma4.
• Module constants: GAMMA_AIR = 1.4, R_AIR = 287.0 J/(kg K), BISECT_TOL = 1e-12, MAX_ITER = 200. Units are SI (Pa, K, kg/m3, m/s); ratios are unitless.
• NACA-TR-824 frames the compressible-flow relations; the equations above are standard engineering methodology, summary-only.

Workflow

1. Fix the diaphragm state: driver-to-driven pressure ratio p4_p1, driver sound-speed ratio a4_a1, specific heat ratios gamma1 and gamma4, and the driven-gas absolute state p1, t1, r_gas; shock_tube_state validates every input and rejects non-physical values with ValueError.
2. Recover the incident-shock Mach number Ms with incident_shock_mach, the deterministic bisection of the shock_tube_residual diaphragm-match equation on the bracket below the ceiling Ms_lim.
3. Build the post-shock driven-gas state: normal_shock_ratios at Ms returns p2_p1, rho2_rho1, T2_T1, and multiplying by the region-1 absolute state gives p2, T2, rho2, a2 and the shock speed Ws = Ms a1.
4. Compute the contact-surface velocity u2 = u3 with induced_velocity from Ms and a1, then the region-2 lab-frame flow Mach number M2_lab = u2 / a2 (just subsonic for a moderate run).
5. Expand the driver gas: expansion_pressure_ratio at u3 / a4 gives p3_p4, T3_T4 = bracket squared, rho3_rho4, and the region-3 absolute state with a3 and M3_lab = u3 / a3 (supersonic in the lab frame for a moderate run).
6. Read the fan wave speeds from shock_tube_state: fan_head_speed = a4 into the quiescent driver gas and fan_tail_speed = u3 - a3, and assemble the full four-region state table.
7. Close the contact: check the residual at Ms, p3 - p2 and u3 - u2, plus the closed-form identities (ideal gas p = rho R T, the shock-frame Mn2 recovery of u2, and p3_p4 times p4_p1 equals p2_p1).
8. Confirm determinism (bit-identical reruns) and ValueError rejection of every non-physical input: p4_p1 <= 1, a4_a1 <= 0, gamma <= 1, p1/t1/r_gas <= 0, and expansion velocities at or beyond the full-expansion limit.
9. Run the contract test offline: python3 scripts/test_shock_tube.py.

Worked example

Air driver and driven gas, gamma1 = gamma4 = 1.4, R = 287.0, a4/a1 = 1 (T4 = T1), p1 = 101325 Pa, T1 = 288.15 K, p4/p1 = 40:

• Ms = 2.057576505, shock speed Ws = Ms a1 = 700.1164 m/s; sound speeds a1 = a4 = 340.2626 m/s.
• Post-shock ratios: p2/p1 = 4.772557918, T2/T1 = 1.734842373, rho2/rho1 = 2.751003776. Region 2: p2 = 483579.43 Pa, T2 = 499.89 K, rho2 = 3.370600480 kg/m3, a2 = 448.1716 m/s.
• Contact velocity: u2 = u3 = 445.6215 m/s, identical to the closed form 2 a1 (Ms - 1/Ms) / (gamma1 + 1); M2_lab = 0.994310161 with Mn2 = 0.567851523; M3_lab = 1.774406595 (shocked driven stream just subsonic, expanded driver stream supersonic, meeting at the contact surface).
• Driver expansion ratios: p3/p4 = 0.119313948, T3/T4 = 0.544750315, rho3/rho4 = 0.219025019. Region 3: p3 = 483579.43 Pa (p3 - p2 = -6.04e-8 Pa at the bisected root), T3 = 156.97 K, rho3 = 10.734203118 kg/m3, a3 = 251.1383 m/s.
• Fan boundaries: head speed 340.2626 m/s leftward into the quiescent driver gas (a Mach wave, M_head = 1.0 relative to region 4); tail speed u3 - a3 = 194.4832 m/s rightward in the lab frame. Region 4: p4 = 4053000.00 Pa, T4 = 288.15 K, rho4 = 49.009027310 kg/m3.
• Four-region state table (p Pa, T K, rho kg/m3, u m/s): region 1: 101325.00, 288.15, 1.225225683, u = 0; region 2: 483579.43, 499.89, 3.370600480, u = 445.6215, M = 0.994310161; region 3: 483579.43, 156.97, 10.734203118, u = 445.6215, M = 1.774406595; region 4: 4053000.00, 288.15, 49.009027310, u = 0.
• Contact checks from the state dict: p3 - p2 = -6.041955e-08 Pa, u3 - u2 = 0.000000e+00 m/s, pressure-match residual at Ms = 5.959677e-13.
• Driver-strength family (air/air, a4/a1 = 1): p4/p1 = 1.2 gives Ms = 1.039831292, u2 = 22.1559 m/s, T2 = 295.71 K; 2.0 gives 1.159478862, 84.2214 m/s, 317.70 K; 5.0 gives 1.402408030, 195.4664 m/s, 361.99 K; 10.0 gives 1.607525211, 279.4268 m/s, 401.44 K; 100.0 gives 2.371054059, 552.7285 m/s, 580.01 K; 1000.0 gives 3.150486214, 803.3246 m/s, 824.23 K. Ms and u2 grow monotonically with the driver overpressure, always below the air/air ceiling Ms_lim = 6.162277660.
• General-gamma case, helium driver (gamma4 = 5/3) at p4/p1 = 40, a4/a1 = 2.4 over air: Ms = 2.698174185, p2/p1 = 8.326834585, T2/T1 = 2.340950220, u2 = 659.9828 m/s, p3/p4 = 0.208170865, T3 = 744.20 K. The light fast driver nearly doubles the contact velocity of the air/air case at the same pressure ratio.

Verification

• Confirm incident_shock_mach(40.0, 1.0) returns Ms 2.057576505 with the residual shock_tube_residual at that root below 1e-10, and that the driver-strength and helium-driver anchors above all reproduce within the tolerances listed in the contract test.
• Confirm the contact-surface match at the root: p3 - p2 below 1e-6 Pa and u3 - u2 exactly 0.0 in the state dict, with p3_p4 * p4_p1 equal to p2_p1 within 1e-9.
• Confirm the identities: p2 / (rho2 R T2) = 1.0 (ideal gas), u2 recovered from 2 a1 (Ms - 1/Ms) / (gamma1 + 1) and from the shock-frame Ms a1 - Mn2 a2, and p2_p1 from the closed form.
• Confirm ValueError rejection of non-physical inputs: p4_p1 <= 1, a4 /a1 <= 0, gamma <= 1, p1/t1/r_gas <= 0, shock_mach <= 1, and expansion velocities at or beyond 2 / (gamma4 - 1).
• Confirm determinism: rerunning incident_shock_mach on the same input returns identical bits; no RNG, pure stdlib math only.
• Run the contract test offline: python3 scripts/test_shock_tube.py (31 tests, deterministic).

Related leaves

• aerodynamics/high-speed/normal-shock: stationary-shock ratios given the upstream Mach number only; it never determines the shock Mach number from a diaphragm pressure state and has no driver gas or moving-shock content.
• aerodynamics/high-speed/oblique-shock: steady single-turn waves with the theta-beta-M solve; its implicit-solve precedent is the pattern this leaf follows for the diaphragm-match bisection.
• aerodynamics/high-speed/prandtl-meyer: steady turning fans; the unsteady driver expansion wave here is the moving counterpart.
• aerodynamics/high-speed/isentropic-flow-relations: steady area-Mach and choked-flow inversions, no unsteady diaphragm wave motion.
• aerodynamics/high-speed/regular-shock-reflection: wall interactions after the incident shock reflects, beyond this leaf's single-transit scope.

Pitfalls

• Solving the stationary-shock problem instead of the shock-tube problem: this leaf's post-shock ratios are evaluated at the incident-shock Mach number recovered from the diaphragm pressure state, not at a Mach number you supply; feed a given upstream Mach number to normal-shock instead.
• Forgetting the driver gas: the expansion side depends on gamma4 and a4/a1, so an air/air answer is wrong for a helium driver at the same p4/p1; the worked example shows u2 rising from 445.62 to 659.98 m/s when the light fast driver replaces air.
• Treating region-2 flow as always subsonic in the lab frame: M2_lab = u2/a2 approaches one at p4/p1 = 40 and exceeds it for stronger runs; the invariant is the shock-frame Mn2 below one, checked via u2 = Ms a1 - Mn2 a2.
• Reading the fan head as the fastest wave in the lab frame: the head (speed a4 leftward) and tail (u3 - a3 rightward) are both Mach waves relative to the gas they move through; a strong driver sweeps the tail downstream, so the two boundaries travel in opposite directions in the lab frame.
• Extrapolating past the full-expansion limit: (Ms - 1/Ms) must stay below C = (gamma1 + 1) (a4/a1) / (gamma4 - 1); beyond it the driver expansion is unphysical and the solver raises ValueError.
• Modeling wall reflections: the classical single-transit solution stops at the first arrival of the incident shock and expansion wave at the end walls; regular-shock-reflection covers the interactions that follow.

Behavior contract (gate 3)

Run the deterministic contract test (stdlib unittest, offline):

python3 scripts/test_shock_tube.py

The 31 tests cover the air/air worked example at p4/p1 = 40 (Ms, post-shock ratios and region-2 absolute state, contact-surface velocity, driver expansion ratios and region-3 absolute state, fan head and tail speeds, four-region state table keys), the contact close-out identities (residual below 1e-10, p3 - p2 below 1e-6 Pa, u3 - u2 exactly 0, p3_p4 * p4_p1 equal to p2_p1, ideal gas, shock- frame Mn2 recovery), the driver-strength family anchors from p4/p1 = 1.2 to 1000 with monotonic Ms and u2 below the Ms_lim ceiling, the helium-driver general-gamma case, ValueError rejection of every non-physical input, and bit-identical determinism reruns. All numeric asserts are order-safe (assertAlmostEqual with delta or math.isclose).

Compliance

• Standards referenced, not reproduced: NACA-TR-824 frames the normal-shock and compressible-flow relations; only names and summary equations appear above, per standards-map.yaml.
• compliance: STANDARDS-REF, gated: false.