Oblique shock and Prandtl-Meyer expansion calculator
What a supersonic flow does at a corner. Turning into itself it forms an oblique shock: the wave angle, weak and strong solutions, the largest turn before the shock detaches. Turning away from itself it expands through a Prandtl-Meyer fan. Free, no sign-up, runs in your browser.
The third tab gives the exact outlet of Anderson's Prandtl-Meyer duct (Mach 2, 5.352° corner), the "Prandtl Meyer shock problem" Skill-Lync's CFD students solve in CONVERGE. The fourth checks a report's numbers against it, after Skill-Lync read 297 of those reports.
θ-β-M and ν(M) as in NACA Report 1135 · checked in Python with a different method · Updated 20 Nov 2026
Defaults: Mach 2 air over a 10° wedge. θ is the flow deflection, the wedge half-angle for a symmetric wedge. Perfect gas, constant γ.
θ is the turn away from the flow at a sharp convex corner. Defaults: the Prandtl-Meyer problem's Mach 2, 5.352°, 1.01 bar and 286.1 K.
Anderson's duct: flat lower wall to x = E, then a straight wall turned down by θ to the outlet at x = L; straight top wall at y = H. Inviscid, slip walls. Mach, θ and γ come from the boxes above.
Type what your report states; empty boxes are skipped. The duct and the inflow come from the boxes above and the Duct outlet tab (default: Anderson's).
- The shock sees only the normal component: Mn₁ = M₁ sin β = 1.267, so the jump is a normal shock at Mach 1.267 (p₂/p₁ 1.707). The tangential velocity carries through unchanged.
- Behind the weak shock the flow stays supersonic (M₂ 1.641). It turns subsonic above θ = 22.71° for this Mach number.
- Total pressure kept: p₀₂/p₀₁ = 0.9846 (1.54% lost). The strong solution (β 83.7°, M₂ 0.6037) is what a high back pressure forces; a free wedge in open flow gets the weak one.
What does the Prandtl-Meyer shock problem ask?
Simulate a 2-D supersonic flow past a convex corner in CONVERGE, refine the mesh with adaptive mesh refinement (AMR) on temperature, and see how the sub-grid criterion changes the captured fan. The case is the one in chapter 8 of John D. Anderson's Computational Fluid Dynamics: The Basics with Applications: air enters at Mach 2 with p = 1.01 × 10⁵ Pa, T = 286.1 K and ρ = 1.23 kg/m³. The domain is 65 m long and 40 m high. The lower wall is flat for 10 m and then turns down by 5.352°, so the outlet is 45.15 m high.
Skill-Lync has 298 public reports titled after this problem in its PG Program in CFD. 160 of them set the inlet to 680 m/s, 25 to 678 m/s, and 134 also run a subsonic 100 m/s case. The findings below come from reading all of them.
Oblique shock relations: β, M₂ and the pressure jump
When a supersonic flow meets a wedge or a concave corner, it turns through θ across a straight shock inclined at β to the oncoming flow. β follows from θ and M₁:
Only the velocity component normal to the shock is slowed. So the jumps are those of a normal shock at the normal Mach number, and the tangential component carries straight through:
Below θmax every θ has two answers. The weak shock (smaller β) is what a wedge in open flow gets, usually with supersonic flow behind it. The strong shock (β nearer 90°) leaves subsonic flow and appears when a high downstream pressure forces it. For Mach 2 in air:
| Weak shock | Strong shock | |||||
|---|---|---|---|---|---|---|
| θ | β | M₂ | p₂/p₁ | p₀₂/p₀₁ | β | M₂ |
| 5° | 34.30° | 1.821 | 1.315 | 0.998 | 86.97° | 0.584 |
| 10° | 39.31° | 1.641 | 1.707 | 0.985 | 83.70° | 0.604 |
| 15° | 45.34° | 1.446 | 2.195 | 0.952 | 79.83° | 0.644 |
| 20° | 53.42° | 1.210 | 2.843 | 0.893 | 74.27° | 0.728 |
| 22.97° | θmax: the weak and strong solutions meet at β = 64.67°; above it the shock detaches | |||||
The flow behind the weak shock turns subsonic just before θmax, at θ = 22.71° for Mach 2. At Mach 3 the limit rises to 34.07°. The chart in the first tab shows the θ-β curves for several Mach numbers.
The Prandtl-Meyer expansion fan
At a convex corner the flow turns away from itself. It does this through a fan of Mach waves centred on the corner, each turning the flow a little and speeding it up, with no loss of total pressure. The Prandtl-Meyer function measures how far a flow has turned from Mach 1:
A turn by θ adds θ to ν. For the problem's corner:
Anderson chose 5.352° so that M₂ comes out at 2.2. Pressure and temperature follow the isentropic relations:
With the problem's inflow that is p₂ = 73,911 Pa, T₂ = 261.7 K and ρ₂ = 0.9841 kg/m³. The fan starts on the Mach line at μ₁ = 30.0° to the incoming flow and ends at μ₂ − θ = 21.68° to it.
| M | ν(M) | Mach angle μ |
|---|---|---|
| 1.5 | 11.905° | 41.81° |
| 2 | 26.380° | 30.00° |
| 2.2 | 31.732° | 27.04° |
| 3 | 49.757° | 19.47° |
| 5 | 76.920° | 11.54° |
120 reports call the expansion isentropic, and 40 write out ν(M). Despite the exercise's name, there is no shock in it. One report evaluates ν(M₁) as 49.756°, which is ν(3): its formula has 3 inside the square roots instead of M₁ = 2. It then carries 26.452° (0.072° high) and reads M₂ = 2.16 off a table, where ν = 31.084° gives 2.175. Its simulation's 2.17 looked "close enough" to that.
Why the outlet Mach number is 2.14, not 2.2
The usual validation compares the Mach number at the outlet with the 2.2 from the Prandtl-Meyer relation. That only works if the whole outlet lies downstream of the fan, and in Anderson's duct it does not. The fan's leading Mach line leaves the corner at 30° and climbs 31.75 m over the 55 m to the outlet. It reaches the 40 m top wall only at x = 79.3 m, beyond the end of the domain. So the outlet plane cuts through all three regions:
The duct to scale. Lines: the leading and trailing Mach lines of the fan and five in between; the dashed line continues the leading line to the top wall past the outlet.
| Outlet plane (corner at y = 0) | Height | Share | Mach |
|---|---|---|---|
| Lower wall to rear Mach line | -5.15 to 21.87 m | 60% | 2.200 |
| Inside the fan | 21.87 to 31.75 m | 22% | 2.200 → 2.000 |
| Above the leading Mach line | 31.75 to 40 m | 18% | 2.000 |
| Average over the outlet | 2.141 by area · 2.134 by mass flow · p 81,800 Pa · T 268.9 K | ||
Exact inviscid Mach number across the outlet plane, with the area average and one report's outlet value (2.12851).
15 reports quote an outlet Mach number. 13 of them land between 2.12 and 2.17, and 8 say 2.15. One report found 2.12851, called it a 3.25% error against 2.2 and blamed the mesh. Against the exact average it is 0.27% off by mass flow, 0.57% by area. The reports were closer than they thought. Another report found 2.13 on average with a local maximum of 2.2, which is exactly the picture above. To validate M₂ itself, put a probe or a line below the rear Mach line, where the flow is uniform at 2.2.
The 1-D area-ratio estimate for the outlet (45.15/40 of the inlet area) gives 2.142, close to the area average. It only becomes the actual mean flow in a longer duct, where the fan reflects off the top wall and the waves settle out.
Boundary conditions for supersonic inflow and outflow
Information in a supersonic flow travels only downstream. At the inlet all the characteristics enter the domain, so all the flow variables are prescribed there: velocity, temperature and pressure. At the outlet they all leave, so every variable is extrapolated from inside (zero normal gradient). 59 reports set the inlet pressure to zero normal gradient and 35 fix it. With a zero-gradient inlet pressure nothing pins the pressure level: it is whatever the initialisation leaves in the first cells (reports use 101,000 Pa or 50,000 Pa). The Mach number and the pressure ratios across the fan can still be right while the absolute pressures are not.
The textbook problem is inviscid with slip walls. A no-slip wall with a turbulence model adds a boundary layer, which at 680 m/s over 65 m is thin: a flat-plate estimate (Re ≈ 3 × 10⁹) gives about 0.3 m at the outlet, against 45 m of height. It changes the averages a little, not the picture.
Which slips does the check tab catch?
- Outlet average compared with M₂: the exact outlet averages, and whether your number matches them.
- A wrong ν(M₁): which Mach number your ν belongs to, and the M₂ it leads to.
- M₂ from a table between rows: the exact M₂ for your turn.
- 680 m/s called Mach 2: the Mach number for air and for CONVERGE's O₂/N₂ mixture (680 m/s at 286.1 K is Mach 2.005; Mach 2 is 678.2 m/s).
- Zero-gradient pressure at a supersonic inlet.
What this calculator does not cover
- Detached and curved shocks. Above θmax the shock stands off the body; its shape and stand-off distance need a numerical solution or correlations.
- Reflections and interactions. The duct tab tells you when the fan reaches the top wall, but does not trace the reflected waves (the method of characteristics does).
- Viscous effects. Real corners separate when the turn is large, and shocks interact with boundary layers.
- Real-gas effects. γ is constant; at high Mach numbers and temperatures it is not.
Setting up a compressible case, choosing boundary conditions that match the physics and validating against an exact solution is daily work for a CFD engineer. The Prandtl-Meyer problem and the shock tube are both cases in the CONVERGE module of Skill-Lync's PG Program in CFD.
Frequently asked questions
How do you calculate an oblique shock angle?
Solve the θ-β-M relation tan θ = 2 cot β (M₁² sin²β − 1) / (M₁² (γ + cos 2β) + 2) for β. Each deflection below θmax has two answers, a weak and a strong shock. For Mach 2 and a 10° wedge in air: β = 39.31° (weak), p₂/p₁ = 1.707 and M₂ = 1.641; the strong solution is β = 83.70° with subsonic flow behind it.
What is the maximum deflection angle for an attached oblique shock?
It depends on the Mach number and γ: 22.97° at Mach 2 and 34.07° at Mach 3 for γ = 1.4. A wedge steeper than that cannot hold an attached shock; the shock detaches and stands in front of the body as a curved bow shock.
What is the Prandtl-Meyer function?
ν(M) is the angle through which a sonic flow must turn, expanding isentropically, to reach Mach M. A turn by θ from M₁ takes the flow to the Mach number with ν(M₂) = ν(M₁) + θ. ν(2) = 26.380° and ν(3) = 49.757°; the largest possible value for γ = 1.4 is 130.45°, an expansion to zero pressure.
Is there a shock in the Prandtl-Meyer shock problem?
No. A supersonic flow turning away from itself at a convex corner expands through a fan of Mach waves; the process is isentropic and there is no shock. The name of the CONVERGE exercise is a misnomer. Shocks form where a supersonic flow turns into itself, at a concave corner or a wedge.
Why is my CFD outlet Mach number lower than 2.2 in the Prandtl-Meyer problem?
Because the outlet plane is not all downstream of the fan. In Anderson's duct (65 m long, 40 m high, corner at x = 10 m, 5.352°) the fan's leading Mach line reaches the top wall only at x = 79.3 m, so at the outlet 18% of the height is still at Mach 2 and 22% is inside the fan. The exact outlet average is 2.141 by area and 2.134 by mass flow. A run that reports 2.13 to 2.14 there is right.
What boundary conditions should a supersonic inlet and outlet have?
At a supersonic inlet every characteristic enters the domain, so velocity, temperature and pressure (or density) are all fixed. At a supersonic outlet every characteristic leaves, so all variables are extrapolated from inside (zero normal gradient). Walls are slip walls for the inviscid textbook problem.
Related on Skill-Lync
- Student project: Prandtl-Meyer expansion fan
- NACA airfoil calculator (supersonic lift of a flat plate from these shock and expansion relations)
- Shock tube calculator (the other CONVERGE compressible case)
- Converging-diverging nozzle calculator (normal shocks, back pressure, MacCormack)
- y+ calculator
- Backward-facing step calculator
- Pipe flow calculator
- All free engineering tools
Sources
- Ames Research Staff, Equations, Tables, and Charts for Compressible Flow, NACA Report 1135 (1953): oblique-shock relations, the cubic in sin²β for the wave angle (used by the checking script), the Prandtl-Meyer function and its tables (ν(2) = 26.380°, ν(3) = 49.757°).
- J. D. Anderson, Computational Fluid Dynamics: The Basics with Applications (McGraw-Hill, 1995), ch. 8, the two-dimensional supersonic Prandtl-Meyer expansion problem: inflow, corner angle and duct dimensions.
- J. D. Anderson, Modern Compressible Flow, ch. 4: the θ-β-M relation, weak and strong shocks, θmax, the Prandtl-Meyer expansion.
- Skill-Lync, 298 public Prandtl-Meyer shock problem reports, read 20 Nov 2026: inflow settings, boundary conditions, outlet values and analytical comparisons.
- Skill-Lync, scripts/verify/oblique_shock.py: the wave angle from NACA 1135's cubic (numpy roots), θmax by numerical maximisation, Prandtl-Meyer inverses by root-finding, and the outlet averages by adaptive quadrature. The page re-checks every number against it when it is built.