Penn State

Mechanical & Nuclear Engineering

Turbine Heat Transfer and Aerodynamics Group

Film Cooling Projects

Film Cooling
Effects of
Sand Deposition

Internal Cooling Projects

Pin Fin
Pin Fin Channels
Ribbed Channels

Additive Manufacturing Projects

Heat Exchanger

External Flow Projects

Conjugate Heat Transfer
Rim Seal
Particle Separation
Rust Deposition
Endwall Film Cooling
Endwall Contouring
Interface Leakage
Horshoe Vortex Flowfield
and Heat Transfer Analysis

Film Cooling

[Back to Top]

The public shaped hole is a laidback fanshaped design

Literature contains many different shaped hole designs, however few geometries are common between publications of different researchers. Most published film cooling studies compare novel cooling hole designs to the performance of a cylindrical hole, which is not necessarily the most helpful standard given the well-known jet detachment that occurs at high momentum flux ratios. Designers and researchers would benefit from defining a baseline shaped hole geometry to be used instead of cylindrical holes for comparison purposes. At PSU ExCCL we performed a literature review that identified over 120 different shaped hole geometries. We then designed a public, baseline shaped hole representative of geometries in literature. The resulting public baseline is the "7-7-7 shaped hole" geometry.

Meter/Diffuser Misalignment

[Back to Top]

Film-cooling test coupons with diagonal aft-right
and fore-right 1/4D offsets.

Film cooling holes are often manufactured using something called electrical discharge machining, or EDM. The EDM drill first creates the metering section of the film cooling hole and then the part is moved to a different EDM drill in order to create the diffuser. If the part isn't perfectly aligned on the second drill, there will be an offset between the centerline of the meter and the centerline of the diffuser. The separation of the flow at this offset creates a deficit in film cooling effectiveness. This deficit is quantified in our wind tunnel using the 7-7-7 Public Shaped Film Cooling Hole.

Effect of Blockage

[Back to Top]

Pictured is an infrared image of the temperature of a
stryrofoam plate cooled by a blocked diffuser.

Thermal barrier coating (TBC) is a ceramic coating that is sprayed on blades and vanes as an additional level of protection from hot gas-path temperatures. This TBC is sprayed after film cooling holes are drilled, and the TBC can build up on the wall of the diffuser, which augments the shape of the diffuser and increases the ejection angle of the coolant. This has a tendency to increase separation on the endwall and reduce film cooling performance. In our lab, we replicated the TBC spraying process in order to conduct experiments that aimed to measure the decrease in cooling effectiveness due to a sprayed-in blockage. These tests were conducted on a 7-7-7 Public Shaped film cooling hole.

Sand Deposition

[Back to Top]

Film and impingement cooling holes are particularly susceptible to particle deposition and flow blockage. These cooling holes are found in regions exposed to the hottest temperatures, including the combustor liner, blades and vanes, and the blade outer air seal.

Impingement- and film-cooling test coupons with
sand deposition [1].

Several previous studies at PSU ExCCL investigated sand deposition in engine components and test coupons with cooling hole geometries representative of engine hardware.

Test facilities were constructed to flow compressed air and inject sand particles through components at pressure ratios realistic to engine conditions. To achieve engine realistic temperatures, some tests were performed inside a kiln capable of heating parts to 1100°C. Flow was measured before and after sand injection, and blockage was characterized in terms of the reduction in flow function.

[1] Cardwell, N. D., Thole, K. A., and Burd, S. W., "Investigation of Sand Blocking Within Impingement and Film Cooling Holes," J. Turbomach., Vol. 132, pp. 021020, 2010.

Bio-Inspired Cooling

[Back to Top]

Branching structures, like those in a leaf
are inspiration for cooling channels.

Dendretic-like transport systems are ubiquitous in nature. From the human circulatory system to leaf venation, branching structures are evolution’s choice for the most efficient, widespread transport of nutrients. The study of such efficient transport architectures in heat transfer applications has gained mild traction in the last decades, but the specific application to gas turbine heat transfer is only recently being explored. Current manufacturing processes impose strict constraints on the types of cooling technologies that can be constructed inside of gas turbine blades and vanes. With advances in additive manufacturing technology, however, complex, reticulate cooling channels inside turbine blades and vanes are a distinct possibility. PSU ExCCL is investigating optimal bio-mimetic branching architectures for internally cooling turbine components.

Pin Fin Horseshoe Vortex Investigations

[Back to Top]

A visualization of the horseshoe vortex
rendered by stereo PIV.

When flow impinges upon pin fins situated between endwalls, a horseshoe vortex system develops near the junction between the pin fin and the endwall as the impinging flow moves around the pin fin obstruction. Previous literature has shown that the horseshoe vortex system is unstable in a quasi-periodic nature, and that the system as a whole is directly related to enhanced endwall heat transfer. While the horseshoe vortex system has been well investigated with respect to a single cylinder in crossflow, the development of the horseshoe vortex with respect to position within pin-fin arrays is not well understood.

Experimental setup of the large scale pin fin
flowfield rig for stereo PIV measurements.

Currently, the ExCCL's large scale pin fin flowfield rig is being utilized in conjunction with stereo particle image velocimetry to better understand exactly what happens to the horseshoe vortex system at different array row locations, array geometries, and Reynolds numbers. With a better understanding of how the horseshoe vortex changes with these parameters, designers will be better equipped to design pin fin arrays that provide the greatest heat transfer under engineering constraints.

Normalized turbulent kinetic energy of the horseshoe
vortex shown at three different Reynolds numbers.

Pin Fin Channels

[Back to Top]

TRDPIV being used to
measure the turbulent
pin-fin flow

In the trailing edge of gas turbine airfoils, the external hot-gas path imposes a large thermal load, and the parent material is characteristically thin. In this region, pin-fin arrays provide a robust solution due to high heat transfer rates and structural integrity from bridging the pressure and suction sides of the airfoil.

In PSU ExCCL, detailed flow measurements are made in pin-fin channels using time-resolved, digital particle image velocimetry (TRDPIV) to determine which turbulent features are best for cooling the airfoils. Using TRDPIV results, pin-fin channel heat transfer tests are then conducted to conclude which geometric pin-fin layouts provide the most efficient cooling.

In the first image, TRDPIV is being used to measure the turbulent pin-fin flow. The resulting time-averaged flow and turbulence, shown in the second image, indicate how the flowfield may influence the heat transfer performance. Finally, heat transfer measurements are made using infrared thermography, and the resulting heat transfer contours are shown in the third image. By comparing the turbulent flow and the surface heat transfer, we can identify which features provide the best heat transfer performance.

Heat transfer measured with IR camera

Flow (top) and turbulence intensity (bottom)
measured with TRDPIV

Ribbed Channels

[Back to Top]

Instrumented ribs (left); and ribs installed in a typical ribbed
channel (right).

In the body, or mid-section, of the airfoil, internal channels are typically lined with ribs since the channel height is too great for using pin-fins. In modern airfoils, the ribbed channels may take on complex 3D shapes. And obtaining the heat transfer on the rib surface typically requires the use of copper models which are expensive and require extensive machining lead times. In PSU ExCCL, complex rib channels are tested quickly and easily using a low-melt temperature alloy to cast a high thermal conductivity rib. The molds are created using stereo-lithography (SLA) rapid prototyping to accommodate the complex shapes typical in modern airfoils.

The left side of the image at right shows instrumentation of the ribs with thermocouples and a foil heater. The right side of the image shows the ribs installed in a typical ribbed channel. The ribs were seated on foam and a small air gap to minimize conduction losses.

Heat Exchanger

[Back to Top]

Heat Exchangers additively manufactured
by the CIMP-3D Lab.

Heat exchangers for aircraft must be light-weight, compact, and efficient. However, traditional manufacturing limitations can sometimes prevent the implementation of highly efficient heat exchange features. Instead, additive manufacturing (AM) techniques can be used to create complex geometries that increase the performance of aircraft heat exchangers without a large increase in manufacturing complexity or cost.

The objective of this study is to compare the performance of a mass-produced, traditionally-built heat exchanger (aircraft oil cooler) to an additively manufactured heat exchanger of the same geometry. 3D X-ray computed tomography (CT) scans were performed on the traditional heat exchanger in order to develop a solid model for AM fabrication using a laser-based powder bed fusion process. The heat exchanger were printed at the Center for Innovative Materials Processing through Direct Digital Deposition (CIMP-3D) in collaboration with the Applied Research Laboratory (ARL).

A performance test rig has been designed and built to characterize heat transfer and pressure drop across the liquid and air streams of both the traditional and additively built heat exchangers. The metal additively manufactured heat exchangers have about 15% more heat transfer but twice the air-side pressure drop. The surface roughness could be the cause of this performance deviation because the additive manufactured surfaces are an order of magnitude more rough than the traditionally built surfaces.

Video of CIMP-3D printing the heat exchanger HERE

Schematic of test rig.

Conjugate Heat Transfer

[Back to Top]

Contours of measured overall cooling effectiveness, with
internal impingement cooling and external film cooling [2].

While adiabatic effectiveness measurements provide a good indication of how well coolant covers the surface, they alone cannot predict the actual endwall temperature. Typically, measurements of adiabatic effectiveness and internal and external heat transfer coefficients are combined to estimate the heat flux into an actual endwall. However, there can be errors with this method for certain areas of the endwall due to conjugate heat transfer effects. The effects of conduction within the wall and internal and external convection can be properly modeled if the geometry, Biot number and ratio of external to internal heat transfer coefficients are matched. This involves using a material with a moderate conductivity to construct the endwall. Endwall overall effectiveness (φ) measurements, or normalized conducting wall temperature, show the cooling effect of conduction and internal convection upstream of the film cooling holes, and more uniform cooling across the passage [2]. A conjugate heat transfer experiment also allows measurement of the thermal performance with and without a thermal barrier coating. The improvement in overall effectiveness with TBC is found to be much greater than other factors such as increasing the coolant flowrate. Additionally, conjugate CFD predictions of endwall and TBC temperatures have good agreement to the corresponding effectiveness measurements [3].

[2] Mensch, A., Thole, K. A., "Overall Effectiveness of a Blade Endwall with Jet Impingement and Film Cooling," J. Eng. Gas Turb. Pwr., Vol. 136, pp. 031901, 2014.

[3] Mensch, A., Thole, K. A., Craven, B. A., "Conjugate Heat Transfer Measurements and Predictions of a Blade Endwall with a Thermal Barrier Coating," J. Turbomach., Vol. 137, 2015.

Combustor/Turbine Interaction

[Back to Top]

Fig. 1. Cartoon depicting four types of vortical structures
associated with transverse jet near-field. [4]

Demands for high-performance gas turbines can be reached by increasing the combustion temperature, this increases the power output of the turbine. The increase in power output comes with a harsher flow field that is often detrimental to the turbine vanes downstream of the combustor. The combustor flow field carries with it temperature gradients, high amounts of turbulence, and large-scale coherent structures, all of which have a large impact on the turbine vanes downstream. While combustor flow fields have been investigated to some degree, the effects of the large coherent structures caused by various aspects of the combustor exit flow have not been heavily studied. The two main causes of the large-scale structures originate from the swirling flow and the dilution jets. The swirling flow, helps to stabilize the flame exiting the combustor, the dilution jets oxidize the remaining fuel and also dilute the product stream before it enters the turbine vane cavity. The swirling flow showcases two types of hydrodynamic instabilities, the Kelvin-Helmholtz instability in the shear layer creates vortices, as well as the large recirculation zone created by the inlet swirl can shed large-scale vortices off it's trailing edge. The dilution jets act as a jet-in-cross-flow, which has a number of vortical structures stemming from it as seen in Figure 1. The JICF has both large and small-scale structures originating from it, the large-scale structures are in the form of a pair counter-rotating vortices, while also creating small-scale features like wake and shear layer vortices. Experiments are conducted in a low-speed recirculating wind tunnel using a linear vane cascade shown in Figure 2. High fidelity measurements are taken using particle image velicometry to aid in the understanding of these coherent structures and their effect on the vane flowfield, an example can be seen in Figure 3.

Fig. 2. Linear vane cascade
and example dilution jets.

Fig. 3. PIV Results

[4] Vortical structure in the wake of a transverse jet, Fric, T. and Roshko A., Journal of Fluid Mechanics Vol. 2798, pp. 1-47, Figure 1

Rim Seal Ingestion

[Back to Top]

Contours of adiabatic effectiveness for
several different purge rates.

Rim seals are turbine features that are used to prevent the ingress of hot gas into the wheel-space below the main gas path. The process of hot gas ingress into a turbine rim seal, known as ingestion, is driven by a complicated set of unsteady conditions in the main gas path as well as in the wheel-space. Linear cascades provide a test bed that is flexible and inexpensive to study the trends in ingestion due to seal geometry. Bluff bodies were placed downstream of the seal to provide a more realistic pressure field in the main gas path. The bluff bodies were designed to match the unsteady pressure distortion of a rotating blade in a stationary cascade. Gas concentration measurements, thermocouples, and infrared thermography were used to measure ingestion levels in the seal and the cooling of the blade endwall. Contours of adiabatic effectiveness on the blade endwall for several different purge rates are shown in the figure. Regions of high effectiveness correspond to locations where the flow leaving the seal, known as purge flow, dominates. Low effectiveness levels indicate regions where the hot gas is present. The contours of adiabatic effectiveness illustrate the rapid mixing that occurs at the seal exit. The large scale of the cascade facilitates flow field measurements near the seal exit. Laser Doppler Velocimetry measurements of the local flow field confirmed the high level of unsteady mixing near the seal exit.

Particle Separation

[Back to Top]

Particle collector circulation patterns [5].

Mechanical separation methods can be implemented to capture and remove particles entrained in the flow with minimal added weight and pressure drop. One such method is that of a static louvered particle separator, which uses the inertia of particles to extract them as the flow streamlines make one or more sharp turns.

CFD studies were performed to assess the effectiveness of louvered particle separators with various louver shapes and spacing. A test facility was then constructed to flow air seeded with sand particles past a louvered particle separator. The effectiveness of the particle separator was assessed based on the proportion of the injected sand that was collected in the separator.

[5] Musgrove, G. O., Barringer, M. D., Thole, K. A., Grover, E. and Barker, J., "Computational Design of a Louver Particle Separator for Gas Turbines," J. Eng. Gas Turb. Pwr., Vol. 135, pp. 012001, 2013.

Rust Deposition

[Back to Top]

Engine hardware with rust deposition [6].

In land-based gas turbines, components such as piping, casings, and rotor hardware can rust, releasing small rust particles into a secondary air system. While some particles may be removed from the air with large fabric filters, others enter the engine and can block flow passages such as the seal/damper pin between two adjacent blade platforms.

A particle characterization study showed that the effects of centrifugal force dominate the effects of temperature for rust particles depositing near the seal/damper pin. Following a method similar to experiments performed with sand, static and rotating test facilities were constructed to flow air and rust particles through engine hardware and test geometries representative of the seal/damper pin area between two blades. Tests were performed at engine realistic pressure ratios and, in the rotating facility, at an engine-representative centrifugal acceleration.

[6] Barringer, M. D., Thole, K. A., Breneman, D., Tham, K-M, and Laurello, V., "Effects of Centrifugal Forces on Particle Deposition for a Representative Seal Pin Between Two Blades," J. Eng. Gas Turb. Pwr., Vol. 135, pp. 032601, 2013.

Endwall Film Cooling

[Back to Top]

Top: Regions of high heat
transfer require cooling [7].
Bottom: Good coolant coverage [8].

At PSU ExCCL, the cooling of a first stage, nozzle guide vane endwall is investigated using various experimental measurements. Spatially resolved temperature data is captured with an Infrared camera to measure cooling effectiveness and convection heat transfer coefficients. We are continually testing new combinations of cooling schemes to determine the best configuration for reducing heat transfer at the airfoil endwalls.

On the image at right, the top figure illustrates the Nusselt number distribution on a vane endwall [7]. Regions of high heat transfer (warm colors) are observed in the stagnation region and would require cooling to prevent thermal damage. An example of a film-cooling scheme that would be employed on a vane endwall is shown in the contours of film-cooling effectivenss in the bottom figure [8].

[7] Thrift, A. A., Thole, K. A., and Hada, S., "Effects of an Axisymmetric Contoured Endwall on a Nozzle Guide Vane: Convective Heat Transfer Measurements," J. Turbomach., Vol. 133, pp. 041008, 2011.

[8] Thrift, A. A., Thole, K. A., and Hada, S., "Effects of an Axisymmetric Contoured Endwall on a Nozzle Guide Vane: Adiabatic Effectiveness Measurements," J. Turbomach., Vol. 133, pp. 041007, 2011.

Endwall Contouring

[Back to Top]

Axisymmetric Endwall Contouring

passage [8]

Planar passage [8]

Gas turbine designs seek improved performance by modifying the endwalls of nozzle guide vanes in the engine hot section. Within the nozzle guide vanes these modifications can be in the form of an axisymmetric contour as the area contracts from the combustor to the turbine.

At PSU ExCCL, the performance of axisymmetric passages are evaluated and compared with traditional planar passages. Measurements have shown that axisymmetric contouring reduces the performance of film cooling as the approaching flow must impinge and turn at the start of the contour. Heat transfer, however, was shown to improve with an axisymmetric contour.

The first image at right shows the cooling effectiveness for a planar passage. The second image shows the cooling effectiveness for an axisymmetric passage. The cooling effectiveness was reduced for the planar passage in comparison with the axisymmetric passage.

Non-Axisymmetric Endwall Contouring

A three-dimensional contour, otherwise known as a non-axisymmetric contour, makes use of localized hills and valleys in the endwall to reduce the radial pressure gradient at the endwall.

Reducing the pressure gradient at the endwall from the pressure to the suction side reduces endwall secondary flows and improves passage aerodynamic losses as measured with a 5-hole probe in PSU ExCCL.

Experimental measurements of endwall heat transfer in a non-axisymmetric contour was shown to improve heat transfer near the pressure side of the passage but increase heat transfer near the blade leading edge versus a flat endwall.

The image below at left illustrates a non-axisymmetric endwall contour in a blade cascade [9]. The image at right illustrates a comparison of secondary flows in the exit plane of a planar and non-axisymmetric contoured passage [10]. By adding endwall contouring, aerodynamic efficiency was increased by suppressing secondary flows.

Non-axisymmetric endwall contour

Comparison of secondary flows

[8] Thrift, A. A., Thole, K. A., and Hada, S., "Effects of an Axisymmetric Contoured Endwall on a Nozzle Guide Vane: Adiabatic Effectiveness Measurements," J. Turbomach., Vol. 133, pp. 041007, 2011.

[9] Lynch, S. P., Sundaram, N., Thole, K. A., Kohli, A., and Lehane, C., "Heat Transfer for a Turbine Blade with Non-Axisymmetric Endwall Contouring," J. Turbomach., Vol. 133, pp. 011019, 2011.

[10] Lynch, S. P., Thole, K. A., Kohli, A., and Lehane, C., "Computational Predictions of Heat Transfer and Film-Cooling for a Turbine Blade with Non-Axisymmetric Endwall Contouring," J. Turbomach., Vol. 133, pp. 041003, 2011.

Combustor/Turbine Interace Leakage

[Back to Top]

Comparison of time-averaged flow fields [11]

Although generally overlooked in comparison to film cooling holes, leakage flow from the combustor-turbine interface slot can provide essential cooling to the endwalls of the first stage, nozzle guide vane. The ability of the leakage flow to reduce the heat load to the endwall is critical as the coolant is a parasitic drain on the compressor, possibly reducing the overall efficiency of the engine. A net benefit to engine efficiency can be gained, however, if the performance of the leakage flow in reducing the endwall heat load allows for increased turbine inlet temperatures. Studies at PSU ExCCL have shown that altering the interface slot orientation can greatly improve endwall durability while also altering the secondary endwall flows.

Comparison of net heat flux reductions [11]

Flow field measurements in the stagnation plane of a vane with leakage flow from an upstream interface slot were made in PSU ExCCL using a time-resolved, high-image-density, digital particle image velocimetry (TRDPIV) system. Velocity measurements indicated the formation of a large leading edge vortex for coolant injected at 90° and 65° while coolant injected at 45° and 30° flows along the endwall and washes up the vane surface at the endwall junction.

Combining the measurements of heat transfer and adiabatic effectiveness, a net heat flux reduction (NHFR) to the endwall can be calculated that indicates the reduction in heat load relative to a case with no leakage coolant. The net heat flux reduction was shown to improve with a reduction in slot injection angle down to 45° where the results reach a plateau. A weakening of endwall secondary flows as a result of 45° and 30° injection is evident in the reduction of sweeping of high neat heat flux values from pressure to suction side in comparison to the 90° and 65° slots.

The top image at left compares time-averaged flow fields with overlaid streamlines and contours of turbulence intensity between (a) 90°, (b) 65°, (c) 45°, and (d) 30° interface slot. The bottom image at left compares net heat flux reductions on the endwall of a vane for several different interface slot orientations.

[11] Thrift, A. A. and Thole, K. A., "Impact of the Combustor-Turbine Interface Slot Orientation on the Durability of a Nozzle Guide Vane Endwall," J. Turbomach., vol. 135, pp. 041019, 2013.

Horshoe Vortex Flowfield and Heat Transfer Analysis

[Back to Top]

A depiction of the Rood Wing test section and large recirculating
wind tunnel, with optical access for PIV measurements.

Time-resolved video of the junction flow vorticity field, showing
in detail a coherent horseshoe vortex feature transitioning
between dynamic modes.

The horseshoe vortex system is a common feature of industry-relevant flows. A test section has been constructed at the PSU ExCCL Lab to study the fundamental effect of freestream turbulence on the dynamic behavior of the horseshoe vortex in front of a single symmetric Rood wing.

Time-resolved stereo particle image velocimetry (SPIV) measurements of the velocity flowfield have been made in the junction flow region in front of the wing leading edge, capturing the dynamic behavior of the horseshoe vortex flowfield for varying freestream turbulence intensity and length scale conditions. The resulting time-resolved flowfield data has shown that at low turbulent Reynolds numbers, freestream turbulence has a large effect in increasing the turbulent kinetic energy of the horseshoe vortex, however, at high Reynolds number, the horseshoe vortex is not sensitive to freestream turbulence.

Currently, time-resolved heat transfer behavior of the horseshoe vortex is being captured using high speed heat flux sensors embedded in the endwall of the wing junction at the leading edge. This data is collected simultaneously with Stereo PIV measurements, providing insight into how the mechanics of the horseshoe vortex may contribute to dynamic events of high or reduced heat transfer.