Impulse vs Reaction Turbine

Legacy context

The site’s documented heritage in vortex methods and turbine cascade analysis provides a precise lens for examining modern turbomachinery questions. Early work on adaptive, fast parallel vortex methods for turbulent separated flows, alongside investigations into compression shock waves and boundary layer separation in supersonic cascades, established a foundation for understanding loss mechanisms in rotating machinery. These legacy efforts focused on how pressure gradients and shock-boundary layer interactions govern efficiency in blade rows.

This background directly informs the distinction between impulse and reaction turbines. In an impulse turbine, pressure drop occurs entirely across the nozzle, leaving the rotor to convert kinetic energy without further expansion. Conversely, a reaction turbine distributes pressure drop across both stator and rotor, altering the velocity triangle and the nature of secondary flows. The vortex dynamics and wake behavior that the site’s earlier cascade studies examined become critical when comparing these designs. Boundary layer development, trailing edge shedding, and shock formation—if present—differ markedly between the two configurations. Understanding these flow physics, rooted in the site’s computational heritage, is essential before selecting a blade profile or predicting off-design performance. The transition from historical solver development to contemporary blade row analysis is therefore direct and practical.

Side-by-Side Comparison Table

---

Fundamental Distinction: Where Does the Pressure Drop Occur?

The defining difference between impulse and reaction turbines lies in the location of the enthalpy (pressure) drop. In an impulse turbine, the entire pressure drop occurs across the stationary nozzle or stator row. The rotor experiences no static pressure change; it merely deflects a high-velocity jet, converting kinetic energy to shaft work through momentum exchange. In a reaction turbine, the pressure drop is distributed—part occurs across the stator and part across the rotor itself. The rotor blades are shaped as converging passages, so the working fluid accelerates through them, producing a reactive thrust in addition to the impulse from the stator-exit jet.

For CFD practitioners, this distinction dictates the boundary conditions and modeling strategy. An impulse rotor can often be modeled with a single rotating reference frame and a prescribed inlet velocity triangle, since the rotor passage does not sustain an adverse pressure gradient that drives separation. A reaction rotor, by contrast, requires careful treatment of the inter-row interface, particularly if you are resolving unsteady blade-row interactions. The degree of reaction—the ratio of rotor enthalpy drop to total stage enthalpy drop—is the key dimensionless parameter. A 50% reaction stage splits the expansion equally, which typically yields high efficiency because both rows operate under similar diffusion factors.

Aerodynamic Loading and Blade Design Implications

From an airfoil-design perspective, impulse blades are essentially turning vanes. They are thick, symmetric, and designed to turn the flow through a large angle (often 120–170 degrees) without significant acceleration within the passage. The loading is concentrated at the leading edge where the jet impinges. Reaction blades, in contrast, are true airfoils with camber and a converging-diverging passage shape. They operate with continuous acceleration along the suction surface, and the pressure distribution resembles that of a highly loaded compressor or turbine airfoil.

This distinction matters for separation control. Low-pressure turbine stages—which are typically reaction designs—operate at low Reynolds numbers where laminar separation bubbles form on the suction surface. Research on low-pressure turbine airfoils has examined passive and active flow-control devices, including synthetic vortex generator jets, to suppress this separation [2]. The unsteady loading environment in reaction stages also raises structural concerns. Pulsed or unsteady excitation can produce higher instantaneous lift coefficients, but this benefit is offset by load unsteadiness that causes structural vibration and fatigue [1]. For a reaction turbine rotor, where the blade loading is distributed and the passage is sensitive to separation, these unsteady effects are more consequential than in an impulse rotor where the loading is localized at the jet impingement point.

CFD Modeling Considerations

For impulse turbines, the CFD challenge is capturing the jet-mixing region downstream of the nozzle and the subsequent impingement on the rotor buckets. The flow is highly non-uniform at the rotor inlet, and the jet may be supersonic if the nozzle pressure ratio is high. Steady-state frozen-rotor approaches often suffice because the impulse rotor does not strongly influence the upstream nozzle flow. However, if you are analyzing a multi-stage impulse turbine (e.g., a Curtis stage), you must resolve the inter-stage mixing and the carry-over of kinetic energy.

Reaction turbines demand more sophisticated treatment. The blade-row interaction is inherently unsteady, and the clocking between adjacent rows—the relative circumferential position of stator vanes—affects the time-averaged performance. Studies of combustor–turbine interaction have defined simulation cases that examine different relative orientations between the combustor exit pattern and the turbine inlet [3]. This clocking sensitivity is a hallmark of reaction stages, where the wake from one row impinges on the next and modifies the boundary-layer transition and separation behavior. For low-pressure reaction turbines, the Reynolds number is low enough that transition modeling becomes critical. The separated-flow transition under low-pressure turbine airfoil conditions has been studied extensively, with turbulence spectra measurements informing transition models [2].

Loss Mechanisms and Efficiency Trade-offs

The loss inventory differs fundamentally between the two types. In an impulse turbine, the dominant losses are jet mixing losses (as the high-velocity jet shears against the lower-velocity surrounding fluid), bucket friction, and windage losses on the rotor disk. The kinetic energy leaving the stage is often large because the rotor exit velocity is high—this is the "carry-over" loss that makes single-stage impulse turbines less efficient than multi-stage reaction turbines for the same overall pressure ratio.

In a reaction turbine, the losses are distributed along the blade surfaces: profile loss from boundary-layer growth and separation, endwall loss from secondary flows, and tip clearance loss. The efficiency advantage of reaction turbines arises because the velocity is more uniform across the stage and the kinetic energy leaving each stage is lower. However, reaction turbines require more stages for the same total pressure drop, increasing mechanical complexity and cost. The trade-off is between the higher stage loading of an impulse design (fewer stages, lower cost) and the higher efficiency of a reaction design (more stages, higher cost). The cost difference is qualitative—reaction turbines with many stages are more expensive to manufacture and maintain, but the efficiency gain can justify the expense in large-scale power generation or high-bypass turbofan applications.

Practical Selection Guidance

For a given application, the choice hinges on the pressure ratio per stage and the flow coefficient. If you need a large enthalpy drop in a single stage with minimal mechanical complexity, an impulse turbine is appropriate. If you need high efficiency over a wide operating range and can accommodate multiple stages, a reaction turbine is preferable. In aircraft engine low-pressure turbines, reaction stages are standard because the efficiency penalty of impulse stages is unacceptable at cruise. In rocket turbopumps, where the pressure ratio is extreme and the working fluid is dense, impulse stages or partial-admission impulse turbines are often used because they can extract a large enthalpy drop in a compact package.

For CFD engineers, the practical advice is to match your modeling fidelity to the dominant physics. An impulse turbine can often be analyzed with steady RANS and a mixing-plane interface, provided the jet-mixing region is adequately resolved. A reaction turbine, particularly a low-pressure stage at altitude Reynolds numbers, requires unsteady RANS or scale-resolving simulations to capture the separation and transition phenomena that govern performance. The computational cost is higher, but the physics demands it.