Experimental validation of a hot gas turbine particle deposition facility

...3Figure 2 - Turbine Accelerated Deposition Facility TADF at Brigham Young University [11] ...7Figure 3 - TuRFR schematic showing main flow path.. Figure 10 - Total pressure measuremen

EXPERIMENTAL VALIDATION OF A HOT GAS TURBINE PARTICLE

DEPOSITION FACILITY

A Thesis

Presented in Partial Fulfillment of the Requirements for the Degree of Master of Science

in the Graduate School of The Ohio State University

By:

Christopher Stephen Smith, B.S

Graduate Program in Aeronautical and Astronautical Engineering

The Ohio State University

2010

Master’s Examination Committee:

Dr Jeffrey P Bons, Advisor

Dr James Gregory

Dr Ali Ameri

A new turbine research facility at The Ohio State University Aeronautical and Astronautical Research Lab has been constructed The purpose of this facility is to re-create deposits on the surface of actual aero-engine Nozzle Guide Vane (NGV) hardware

in an environment similar to what the hardware was designed for This new facility is called the Turbine Reacting Flow Rig (TuRFR) The TuRFR provides air at temperatures

up to 1200 °C and at inlet Mach numbers comparable to those found in an actual turbine (~0.1) Several validation studies have been undertaken which prove the capabilities of the TuRFR These studies show that the temperature entering the NGV cascade is uniform, and they demonstrate the capability to provide film cooling air to the NGV cascade at flow rates and density ratios comparable to the NGV design Deposition patterns have also been created on the surface of actual NGV hardware Deposition was created at different flow temperatures, and it was found that deposition levels decrease with decreasing gas temperature Also, film cooling levels were varied from 0% film cooling to 4% film cooling It was found that with increased rates of film cooling deposition decreased With the TuRFR capabilities demonstrated, research on the effects

Integrated non-dimensional total pressure loss values were calculated in an exit Rec range

of 0.2x106 to 1.7x106 for a deposit roughened NGV cascade and a smooth cascade The data suggests that deposition causes increased losses across the NGV cascade and possibly earlier transition The data also suggests a possible region of separated flow in the NGV cascade which disappears at higher exit Reynolds numbers These results are similar to those found in the literature

ACKNOWLEDGMENTS

I would like to take this opportunity to thank my advisor Dr Jeffrey Bons for his patience, guidance, and continual support during this endeavor It has truly been privilege learning from him

I would also like to acknowledge the help of Brett Barker, Carey Clum, and Josh Webb Without their knowledge and willingness to solve problems the TuRFR would not be operational today

I would also like to thank the staff of AARL, especially Ken Copely, Ken Fout, Jeff Barton and Cathy Mitchell for their undying patience and help

I would also like to express my appreciation to Dr James Gregory and Dr Ali Ameri for serving on my graduate committee

Finally I would like to thank my friends and family for their support and encouragement along the way

VITA

July 15, 1984 ……… Born in Texas City, Texas

August 2007……….B.S Mechanical Engineering,

The University of Texas at Austin

2007 – Present ………Graduate Research Associate,

Department of Aerospace Engineering

The Ohio State University

PUBLICATIONS

1 Smith, C., Barker, B., Clum, C., Bons, J., “Deposition in a Turbine Cascade with Combusting Flow,” to be presented at the 2010 Turbo Expo in Glasgow, Scotland June 2010, paper # GT2010-22855

FIELDS OF STUDY

Major Fields of Study: Aeronautical and Astronautical Engineering

TABLE OF CONTENTS

ABSTRACT ii

ACKNOWLEDGMENTS iv

VITA v

FIELDS OF STUDY v

TABLE OF CONTENTS vi

LIST OF FIGURES ix

LIST OF TABLES xii

LIST OF EQUATIONS xiii

NOMENCLATURE xiv

CHAPTER 1 1

INTRODUCTION 1

1.1 Background 1

1.2 Literature Review 3

1.2.1 Other Deposition Research 6

CHAPTER 2 10

TuRFR DESCRIPTION AND VAILDATION STUDIES 10

2.1 Primary Flow Path Description 10

2.2 Particulate Feed Sub-System 17

2.3 TuRFR Validation Studies 19

2.3.1 NGV Inlet Temperature Survey 21

2.3.2 Film Cooling Validation Survey 22

2.3.3 Exit Total Pressure Surveys 24

CHAPTER 3 30

ASH PARTICULATE 30

3.1 Description of Particulate 30

CHAPTER 4 34

DEPOSITION TESTING RESULTS 34

4.1 Deposition Testing 34

4.1.1 Temperature Variation 37

4.1.2 Film Cooling Variation 38

4.2 Surface Roughness Measurements 39

4.3 Deposit Structure and Chemical Composition 42

CHAPTER 5 45

AERODYNAMIC PERFORMANCE ASSESSMENT 45

5.1 Aerodynamic Performance Assessment Background 45

5.1.1 Total Pressure and Exit Temperature Measurement Setup 46

5.1.2 Inlet Total Pressure Measurement 48

5.2 Aerodynamic Performance Results 52

5.2.1 Rough Aerodynamic Performance 53

5.2.2 Smooth Aerodynamic Performance 59

5.3 Discussion of Results 64

CHAPTER 6 70

6.1 Conclusions 70

6.2 Future Work 71

REFERENCES 73

APPENDIX A: 76

UNCERTAINTY ANALYSIS OF THE γ CALCULATION 76

LIST OF FIGURES

Figure 1 - Leading edge of turbine nozzle guide vanes exposed to high volcanic ash concentrations Note the effects of increased surface heat transfer caused by clogged film cooling holes [2] .3Figure 2 - Turbine Accelerated Deposition Facility (TADF) at Brigham Young

University [11] 7Figure 3 - TuRFR schematic showing main flow path 12Figure 4 - Flameholder inside combustion section of TuRFR 13Figure 5 - 3D cutaway view of upper section of TuRFR showing measurement locations 15Figure 6 - Top view of NGV cascade and film cooling reservoir showing film cooling temperature measurement locations 16Figure 7 - 3D CAD schematic of particulate feeder Pressure equalization tube is not shown in the diagram 18Figure 8 - NGV cascade inlet temperature survey 22Figure 9 - Dimensionless temperature at vane exit with 4% film heating, red box shows approximate location of measurement plane 24

Figure 10 - Total pressure measurements at the exit of the NGV cascade with 0% film heating applied: a) Non-dimensional total pressure (PTOT/Pamb) and b) lines of total

pressure loss at various span locations 26Figure 11 - Non-dimensional total pressure loss [(PTOTin – PTOT)/0.5ρu2] collected 13% true chord downstream of NGV exit plane indicating passage periodicity 27Figure 12 – Total pressure measurements at the exit of the NGV cascade with 4% film heating applied: a) Non-dimensional total pressure (PTOT/Pamb) and b) lines of total

pressure loss at various span locations 29Figure 13 - Particle size distribution results from Coulter counter for Bituminous coal fly ash 32Figure 14 - SEM image of Bituminous coal fly ash 33Figure 15 - Post test images of deposit formation for Low Temperature test, No Film Cooling test, and Film Cooling test 37Figure 16 - Representative surface roughness traces and roughness regions for deposition formed during NO Film Cooling test CAD image of NGV doublet is not to scale 41Figure 17 - SEM photo highlighting deposit structure The vane surface is not indicated

in the image, but it is below the deposit flake 42Figure 18 - EDS element maps of deposit flake highlighting Cu, Fe, Al, Si, and O layers SEM photo highlighting deposit structure The vane surface is not indicated in the

images, but it is below the deposit flake in all of the images 43Figure 19 - Ptotin vs curve showing the experimental data, the curve fit, and the

equation of the fit m in the fit equation stands for inlet mass flow 49

Figure 20 - Schematic of TuRFR used in the transit time estimates 51

Figure 21 - Photo of exit plane indicating the approximate location and direction of the probe head travel 53

Figure 22 - Photograph of NGV cascade surface with and without deposition 55

Figure 23 - Non-Dimensional Total Pressure Loss vs Position for Rec exit ≈ 0.5x106 56

Figure 24 – γ vs Rec exit for Blade 3 of the NGV cascade for the rough cases only 57

Figure 25 - Total skin friction coefficient of sand roughened plate at zero incidence angle 59

Figure 26 - γ vs Rec exit for Blade 3 of the NGV cascade 60

Figure 27 - γ vs Rec exit for Blade 2 of the NGV cascade 61

Figure 28 - Yarn tuft test showing that separation is not occurring in the Reynolds number range considered 63

Figure 29 - Area averaged total pressure loss coefficient vs Exit Reynolds number from Boyle and Senyitko [9] (filled symbols) with avearge γ between data presented in Figure 26 and Figure 27 (open symbols) 65

LIST OF TABLES

Table 1 - Test conditions for TuRFR validation studies 20Table 2 - Particulate mass mean diameter and bituminous coal ash molecular

composition 31Table 3 - Deposition testing conditions 35Table 4 - Operating conditions for deposition creation and surface characteristics of the NGV cascade before and after deposition 54Table 5 - Exit Reynolds numbers and gamma for Blade 2 and Blade 3 of the NGV

cascade 58

Table 6 - Typical values of the N and γ uncertainties using the error propagation analysis.

80Table 7 - Values used in the γ variance determination 81

LIST OF EQUATIONS

Equation 1……… 4

Equation 2……… 27

Equation 3……… 36

Equation 4……… 36

Equation 5……… 36

Equation 6……… 36

Equation 7……… 39

Equation 8……… 45

Equation 9……… 48

Equation 10……… 51

Equation 11……… 52

A* Reference cross sectional area required to reach sonic velocity

Ain NGV inlet cross sectional area

Avs Cross sectional area of view section of TuRFR

Cu2 Absolute tangential flow velocity in turbine rotor passage

CD Drag Coefficient

Cu Chemical Symbol for Copper

CAD Computer Aided Drawing

EDS Energy dispersive spectroscopy

L Total width of wake used in the γ calculation

LDV Laser Doppler Velocimetry

Mguess Mach number at NGV inlet guess

Mvs Mach number in view section of TuRFR

NGV Nozzle guide vane

Pamb Ambient pressure

Pinlet Average total pressure at the cascade inlet

Pop Pressure in main air line just downstream of the orifice plate flow meters

Ppipe Pressure in the main air line

PTOT meas Measured total pressure at the exit

PTOT Total pressure at NGV exit

PTOTin Average total pressure at NGV inlet

PSP Pressure Sensitive Paint

R Specific gas constant

Ra Average relative roughness

Re Reynolds number

Rec Reynolds number based on the true chord of the NGV

Recx Reynolds number based on the axial chord of the NGV

Rec exit Exit Reynolds number based on the true chord of the NGV

Rep Reynolds number of flow around a spherical particle SEM Scanning electron microscope

Tex Temperature of air exiting NGV cascade

Average inlet temperature

TFC Film cooling temperature

Tmeas Measured temperature

TADF Turbine Accelerated Deposition Facility

TuRFR Turbine Reacting Flow Rig

U2 Turbine rotor tangential velocity

Ideal shaft work done by turbine

X Pitchwise grid dimension

XRD X-ray diffraction

Y Spanwise grid dimension

a Speed of sound in TuRFR during operation

c NGV true chord (6.8cm est.)

cx NGV axial chord (3.6cm est.)

dp Diameter of coal ash particle

dx Spacing between measurement locations of PTOT meas

k Ratio of specific heats

Mass flow of fluid ppmw-hr Parts per million by weight hour

t Total transit time for air in TuRFR to go from the orifice plates to the

NGV exit plane

uexit Flow velocity at the NGV exit

uvs Flow velocity in view section of TuRFR

wax Axial flow velocity in a turbine cascade

x Location along main air line being considered for the liner drop in Ppipe

Δg Total head loss

τ Time constant of particle reaction to abrupt change in gas flow

γ Integrated Non-Dimensional Total Pressure Loss

as an alternative fuel cite its relatively low cost and availability as the primary reasons for converting to coal energy However, there are opponents who have justifiable reasons not to use coal as an alternative energy source Among the arguments against coal usage are the negative impacts on the environment from coal combustion and coal acquisition Other reasons coal is having trouble gaining momentum in its fight to become the fuel of choice is the negative effect of coal combustion products on power generation hardware

When coal is burned it can sometimes produce airborne particulate known as fly ash Often coal is used in steam power generation plants as the fuel source for steam generation The fly ash produced by the combustion of coal, deposits onto the walls of the boiler system and degrades the system over time Periodically the entire plant must

be shut down and the boiler over-hauled in order to lengthen the life of the power generation system and to protect the workers from an unexpected catastrophic failure If coal is used as an alternative fuel in a gas turbine the combustion products could deposit onto the surface of the turbine hardware Removing deposition on turbine hardware requires weeks of intense labor to disassemble the turbine, clean the components, and reassemble the turbine Not to mention the risks to personnel and equipment safety Understanding how deposition forms on the surface of turbine hardware can help to reduce it, therefore increasing the time between overhauls and reducing costs In much the same way liquid and gaseous coal based fuels can create deposits when burned in the combustion chamber of a jet engine or power generation system [1] Turbine machinery

is much more complicated and sensitive to minor changes in geometry which could result from the deposition of fly ash onto its surfaces Some of the side effects of deposition are increased heat transfer due to increases in surface roughness, and the blockage of essential film cooling holes These side effects, combined with constantly increasing turbine inlet gas temperatures, make the problem of deposition more and more prevalent

1.2 Literature Review

As previously mentioned, particulate entrained into a turbine can erode or corrode the turbine material, or it could deposit onto the turbine blade surface, causing severe damage and potential failure The effects of deposition on turbine hardware are well documented in the literature Kim et al [2] looked at the effects of large amounts of volcanic ash being ingested into the hot sections of two aero-engines It was found that deposition can cause clogging of critical film cooling holes, which can result in the deterioration of the turbine hardware As the hardware degrades it could potentially release in large portions and cause large amounts of damage to downstream components Figure 1 shows a possible outcome of blocked film cooling holes A significant portion

of the tip region is eroded away causing losses in aerodynamic performance and part integrity

Figure 1 - Leading edge of turbine nozzle guide vanes exposed to high volcanic ash concentrations Note the effects of increased surface heat transfer caused by

Another effect of deposition inside a turbine is made evident in a model proposed

by Wenglarz [3] He states that decreasing the throat area of nozzle guide vanes with deposition reduces the overall power generated by the turbine because of the reduction in mass flow through the turbine This fact is made apparent by examining Equation 1 Overall power output is directly proportional to total mass flow passing through the engine, therefore a decrease in total mass flow results in a decrease in overall power

is decreased when cooling holes are either partially or completely blocked Schlichting [6] presents the fundamental principles for boundary layer growth on rough surfaces He shows that boundary layer thickness increases substantially as the parameter known as equivalent sand roughness increases Schlichting developed the parameter equivalent sand roughness in an attempt to unify roughness measurements Bammert and Sandstede [7] quote rough surface momentum thicknesses at the trailing edge of a fifty percent reaction turbine blade three times as great as those on a smooth turbine blade The equivalent sand roughness used to create such a marked increase in momentum thickness was 3.3x10-3 µm The momentum thickness is directly related to the skin friction drag

along a surface, and with an increase in drag comes a loss in performance of turbomachinery Zhang, et al [8] show that the wake behind a symmetric airfoil, simulating the pressure surface of a turbine vane, increases in width with increasing surface roughness, and that non-dimensional total pressure losses also increase with increasing surface roughness They attribute the decreasing performance to thicker boundary layers at the airfoil trailing edge, and increased turbulent diffusion in the transverse direction within the wake as it travels downstream Boyle and Senyitko [9] show that as Reynolds number increases, at constant Mach number, a uniformly rough turbine vane will produce up to a 60% loss increase when compared to a smooth vane They also show a slight loss benefit at low Reynolds numbers due to the surface roughness, which was attributed to laminar separation being prevented by earlier boundary layer transition Zhang et al [10] demonstrates a 60% increase in total pressure loss with increased surface roughness, and a marked increase in the width of the vane wake for a scaled up version of a turbine vane They performed tests on uniformly rough vanes and vanes with variable roughness Zhang et al claim that the roughness elements cause earlier boundary layer transition from laminar to turbulent, thicker boundary layers, and increased turbulent diffusion, which combine to increase the losses for the rough vanes On top of performance degradation caused by deposition, cleaning and repairing turbine hardware is expensive and time consuming It is clear from current literature that particulate deposition poses significant problems for the operation of turbines using dirty fuels

1.2.1 Other Deposition Research

The studies conducted which show the negative effects of deposition on turbine hardware used either massive amounts of particulate in a fully operational jet engine or prefabricated, large scale, deposition patterns taken from in-service turbine hardware They do not allow the detailed study of the mechanisms behind particulate deposition or the physical characteristics of the deposits However, several other facilities are equipped

to examine the physical characteristics of deposits and the mechanisms behind deposit formation Jensen et al [11] describes a facility, called the Turbine Accelerated Deposition Facility (TADF), used to study the mechanisms of deposition on one-inch diameter cooled and un-cooled turbine material samples The TADF, shown in Figure 2, brings air into its base which is heated using natural gas to temperatures comparable to turbine inlet temperatures Particulate is injected into the flow prior to the flow being accelerated to turbine inlet Mach numbers The hot particulate laden flow is exhausted to atmosphere where the cooled and un-cooled turbine material samples are fixed The flow impinges on the samples and deposits are formed on the surface

Figure 2 - Turbine Accelerated Deposition Facility (TADF) at Brigham Young

University [11]

Crosby et al [12] used the TADF to study how gas temperature, particle size, and metal temperature affect the physical characteristics of the deposit They found that a deposition threshold exists at 960 °C, and that deposition increases with particle size, and deposition decreases with metal temperatures Ai et al [13] evaluated both impingement

and film cooling effects using the same facility They found that deposition could be reduced substantially with proper cooling

Due to the obvious complexities of high temperature deposition experiments, several other research groups have recently designed low temperature deposition simulators For example, Lawson and Thole [14] describe a low speed and low temperature facility used

to simulate deposition on a flat plate with a row of film cooling holes This facility uses low melting temperature wax to form deposition patterns on a flat plate with film cooling holes Adiabatic effectiveness is evaluated near the cooling holes using infrared thermography It was found that cooling effectiveness at low momentum flux ratios was decreased as deposition increased Another facility constructed by Vandsburger and Tafti [15] uses Teflon or PVC particles in an elevated temperature gas flow to simulate syngas ash particulate entrained in a power turbine The Teflon or PVC particles were chosen such that their momentum and thermal Stokes numbers match that of coal ash; this is to ensure they behave in a way similar to coal ash in turbine flow paths These facilities begin to inspect the mechanisms behind the deposition process and the effects of particulate deposition on turbine hardware and film cooling effectiveness However, there has not been a facility that is capable of simulating deposition on actual turbine hardware in a controlled flow environment similar to the turbine environment Testing actual turbine hardware is extremely useful in that the test results are direct indicators of what could happen in service A new facility is needed in order to conduct these tests in

a timely manner Kramer [16] has detailed the construction of such a facility called the Turbine Reacting Flow Rig (TuRFR) This thesis presents an overview of this facility,

some validation studies of the facility, and the initial deposit test results found by using the facility This thesis also presents the effects of deposits on the aerodynamic performance of the turbine hardware

CHAPTER 2

TuRFR DESCRIPTION AND VAILDATION STUDIES

2.1 Primary Flow Path Description

To study the effects of particulate deposition on turbine Nozzle Guide Vanes (NGV) a facility was constructed that is capable of attaining flow conditions representative of those found exiting the combustion section of a power generation turbine The facility, named the Turbine Reacting Flow Rig (TuRFR), utilizes current NGV technology for deposition tests and is capable of supplying film cooling air at flow rates and temperatures comparable to those of the specific NGV design One goal of the TuRFR is to reproduce deposition patterns found on turbine NGVs, which have been in service for thousands of hours, in a 1 to 2 hour accelerated test A brief description of the TuRFR and its sub-systems is provided below

Figure 3 shows a schematic of the TuRFR and the primary flow path The main air supply comes from two 16 MPa, 21 m3, storage tanks that are maintained at constant pressure by three Ingersol-Rand compressors Air from these tanks passes through a regulator and two parallel choked orifice flowmeters Pressure and temperature measurements at the orifice plate yield an uncertainty in mass flow rate of 3-5% over a

range of massflows from 0.5 to 1 kg/s Following the orifice meters, the airflow is split into four branches: main airflow, film cooling flow, fuel premix flow, and particulate flow The main airflow is brought into the base of the combustor at four evenly-spaced locations The air passes through a pebble bed and honeycomb system to evenly disperse and straighten the flow At this point, the air is seeded with particulate (a description of the particulate feed system is provided later on in Sec 2.2 Particulate Feed Sub-System) Next the air is heated, via the combustion of natural gas, to similar temperatures found in power generation turbines Natural gas is taken from a local low pressure supply line Natural gas predominantly consists of methane, but it can contain significant amounts of several other hydrocarbons The detailed chemical makeup and molecular weight of the natural gas was supplied by Columbia Gas Natural Gas used in this particular study has specific gravity of 0.592, a heating value of 38.6 MJ/m3, and consists of 95% methane, 2.35% ethane, 1.17% CO2, and 1.48% other hydrocarbons Fuel gas is compressed to 20 MPa, using two Fuelmaker reciprocating compressors, and stored in an array of 32 high pressure gas cylinders Fuel gas is passed through a choked orifice meter similar to the one used for the main air supply Temperature and pressure measurements at the orifice plate yield an uncertainty in fuel flow rate of 3% at flow rates

up to 0.03 kg/s Fuel is brought into the combustion section using eight fuel lines evenly dispersed around the circumference of the TuRFR At the end of each fuel line is a set of four flame-holders (32 flame-holders in all) The flame holders, shown in Figure 4, are necessary to insure flame stability and mitigate sooting The flameholders are constructed of stainless steel to increase oxidation resistance

Figure 3 - TuRFR schematic showing main flow path

After heating, the air is accelerated through a 60° axisymmetric cone with an inlet-to-exit area ratio of 68:1 The cone accelerates the heated air to Mach numbers in the range of 0.1 to 0.3, depending on the flow rate and temperature of the air Attached to the cone is the equilibration tube This tube allows the particulate to come to kinetic and

thermal equilibrium prior to entering the NGV cascade The length of the tube, 0.79 m, was prescribed so that a 40 micron diameter particle would be at thermal and kinetic equilibrium with the airflow upon exiting the tube

Figure 4 - Flameholder inside combustion section of TuRFR

The heated air leaves the equilibration tube and passes through a round to rectangular transition prior to entering the viewing section At its circular lower end, the round to rectangular transition incorporates a circumferential sliding seal which accommodates the thermal growth of the TuRFR The sliding seal allows for approximately 5cm of relative thermal growth between the upper TuRFR components

(round to rectangular transition, view section, and vane holding section) and the lower TuRFR components (the combustion section, cone, and equilibration tube) There are two viewing ports on the side of the TuRFR which allow access for various optical measurement techniques, such as infrared thermography and Laser Doppler Velocimetry (LDV), and for video recording of deposition on the vane leading edges as it is occurring (Figure 5) One of the optical cavities is equipped with a static pressure port and thermocouple to measure the gas properties just upstream of the final inlet contraction The uncertainties of these measurements are 3% for the pressure measurement and 2% for the temperature measurement

This same optical view section has two isolation cavities located on either side of the main gas path The cavities accept lower temperature air from a separate source and are separated from the main gas path with removable plates These plates can be perforated to allow cooler air to mix with the primary gas path This capability replicates pattern factors (spanwise temperature variations) created by dilution jets found in modern combustor liners For the present study, the removable plates were solid, allowing no dilution flow Zero dilution flow was selected for this study due to lack of information

on pattern factors seen by the NGV cascade in use

Finally, the air enters a contraction which transitions from the rectangular viewing/dilution section to the annular inlet area of the NGV cascade The contraction ratio for this final transition is roughly 2:1 with the NGV inlet area being approximately 0.0132 m2 The NGV hardware used in this study is from a CFM56-5B production engine The two nozzle guide vane doublets are mounted at the exit of this rectangular to

annular transition piece Figure 6 shows a top view of the NGV housing with the cover removed

Figure 5 - 3D cutaway view of upper section of TuRFR showing measurement

locations

Film cooling air is supplied to the NGV cascade from an auxiliary flow path The total film cooling flowrate is measured using an inline pneumatic volumetric flow meter with an uncertainty of 7% of reading at flow rates equal to 12% of the cascade inlet flow rate The outer casing and inner hub film cooling cavities are fed by separate lines The casing film cooling cavity is metered using an in-line pneumatic flow meter with an uncertainty of 7% at flow rates up to 8% of the cascade inlet flow A backflow valve is used to balance the temperatures for the two film cooling reservoirs The secondary flow path for the film cooling air is kept separate from the main flow path to prevent

density ratio was calculated as the NGV cascade inlet to film cooling reservoir temperature ratio This assumes the pressure of the film cooling air is equal to the free stream pressure, which is not exactly the case Thus density ratios calculated represent conservative estimates of the actual density ratio Film cooling air from the hub and casing cavities enters the hollow vanes through cooling cavities that are equipped with impingement inserts, similar to industrial practice

Figure 6 - Top view of NGV cascade and film cooling reservoir showing film cooling

temperature measurement locations

The NGV cascade exhausts to atmosphere from the test cell through a retractable ceiling and large bay doors Due to the high temperatures required of the TuRFR all of the components downstream of, and including, the combustion section are made from Inconel 600 series alloys These exotic metals are capable of retaining strength at extremely high temperatures which makes them aptly suited for this type of application

The strength of the alloys used in the TuRFR design allow it to operate at temperatures

up to 1200 °C for three to four hour tests Based on deposition threshold temperatures cited by previous researchers [1,2,12] the TuRFR is capable of producing gas temperatures and particulate impact velocities within the limits required to produce deposition

2.2 Particulate Feed Sub-System

One of the main goals of the TuRFR is the duplication of particulate deposition seen

on in-service hardware, inside a laboratory setting Without the ability to supply particulate to the main air flow passing through the turbine hardware, the TuRFR would not be able to accomplish its goal The particulate feed sub-system, shown in Figure 7, consists of a particulate hopper, an auger (with agitator), a driving motor, and a gearbox The particulate hopper is a 6 foot section of clear two inch diameter PVC pipe The PVC pipe is clear to allow for gauging of the particulate level during a test The hopper was designed to hold more than enough particulate for a 2.5 hour test at particulate concentration of 80 ppmw in a 1.18 kg/s main air flow If the particulate flows out of the hopper through the auger the pressure in the hopper will decrease due to the increase in volume Therefore the space above the particulate inside the hopper must be kept at the same static pressure as the air being bled from the main flow Pressure equalization is accomplished via a small piece of tubing connecting the hopper to the main air bleed line (not indicated in Figure 7)

Figure 7 - 3D CAD schematic of particulate feeder Pressure equalization tube is

not shown in the diagram

Directly below the hopper is the auger The auger is powered by a small DC motor connected to a speed reducing gearbox A speed reducing gearbox is used to maintain the low particulate flow rates required to match the net particle loading for a one to two hour test The small DC motor is powered by a variable power supply so that the particulate feed rate can be controlled The auger forces particulate into the air stream that has been

bled from the main air flow The particulate is picked up by the secondary air stream passing over the auger and is injected into the TuRFR in the direction of the flow directly beneath the flameholders To aid in particulate pickup a small contraction in this secondary flow path reduces the static pressure in the vicinity of the auger exit This static pressure decrease helps to pull the particulate from the auger and reduce clogging

in the feeder

A common problem with any type of particulate injection system is a phenomenon known as bridging This is when the particulate in the hopper, in essence, clogs by forming a bridge strong enough to support the weight of the particulate above When this occurs the particulate below the bridge flows normally while the particulate above the bridge is prevented from ever reaching the auger To mitigate this effect, the auger was fitted with agitators, welded to its end, which stir the particulate as it approaches

2.3 TuRFR Validation Studies

Prior to conducting deposition studies using the TuRFR, several validation studies were required The validation studies included a survey of the temperature entering the inlet of the NGV cascade, a temperature survey at the exit plane of the cascade for a non-heated main flow with heated film cooling air, and a total pressure survey at the exit plane of the cascade for a non-heated flow The purpose of these tests are to characterize and validate the inlet conditions entering the NGV cascade, to ensure the absence of flow anomalies, and to prove that film cooling air is in fact exiting the film cooling holes and

is affecting the free stream flow in an expected manner For the remainder of this thesis

the term “film heating” will refer to heated film cooling air being supplied through the film cooling holes

A traverse mechanism was constructed at the exit of the NGV cascade LabView software was configured to control the traverse system, acquire the pressure and temperature data, and monitor other important parameters of the TuRFR during operation The actual grid used for each survey is presented in the respective sections All of the grids were rectangular None of the measurement grids were able to traverse the entire inlet or exit flow field due to spatial constraints of the measurement probes Test conditions are indicated in Table 1, again with non-combusting flow and film heating

Temperature Map Pressure Maps Inlet

Temperature Map

4% Film Heating

0% Film Heating

4% Film Heating Inlet Velocity

Table 1 - Test conditions for TuRFR validation studies

2.3.1 NGV Inlet Temperature Survey

For the studies presented herein a uniform inlet temperature distribution was required If the inlet temperature is not uniform it must be characterized so that any unexpected non-uniformities in deposition can potentially be explained For this validation study the NGV cascade was removed from the TuRFR and the main air flow exhausted to atmospheric conditions through the annular hole where the NGV cascade is fixed The traverse system was mounted away from the annular exit of the TuRFR in order to protect it from the hot main air flow A metal arm was mounted to the traverse which held a 1.58 mm diameter, Type K, thermocouple inside the main air flow A grid

of 9 cm in the pitch dimension, and 5.4 cm in the span direction with 3 mm spacing between each point was used An outline of the grid used in this test is shown by the rectangular box in Figure 8 It is important to note that the thermocouple tip started approximately 5 mm from the hub radius inside the flow path When the test was complete the thermocouple tip was about 3 mm from the hub radius outside the flow path This offset is represented in the test data by a very cold region, which has a similar shape to the hub radius of the annular hole in the TuRFR The thermocouple probe was traversed in a plane that was approximately 3 mm above the cascade inlet plane Figure 8 shows the result of the inlet temperature survey non-dimensionalized by the average temperature of the inlet area The average temperature was calculated using the region indicated by the line and arrows in Figure 8

Figure 8 - NGV cascade inlet temperature survey

From these results it is evident that there is a spanwise temperature variation in the inlet temperature This variation is at most 10% of the average inlet temperature An inlet temperature variation may cause differences in the formation of the deposits along the vane span because deposit formation is highly dependent upon temperature

2.3.2 Film Cooling Validation Survey

The film cooling validation survey was conducted by supplying heated film cooling air to an unheated main flow Conducting the validation study in this way allows the air exiting the film cooling passages to stand out from the main flow without the need for combustion Also, the danger to the traverse mechanism incurred by a heated flow required that the flow be unheated

Film heating was supplied to the NGV cascade at a rate of 4% of the inlet mass flow The inlet mass flow for this test was set to 1.27 kg/s, which required a film cooling