REVIEW OF SCIENTIFIC INSTRUMENTS 85, 124902 (2014)
Aircraft engine-mounted camera system for long wavelength infrared imaging of in-service thermal barrier coated turbine blades James Markham,1,a) Joseph Cosgrove,1 James Scire, Jr.,2,b) Charles Haldeman,3 and Ian Agoos3 1
Advanced Fuel Research, Inc., 87 Church Street, East Hartford, Connecticut 06108, USA New York Institute of Technology, Northern Blvd, Old Westbury, New York 11568, USA 3 Pratt & Whitney, 400 Main Street, East Hartford, Connecticut 06108, USA 2
(Received 19 September 2014; accepted 20 November 2014; published online 5 December 2014) This paper announces the implementation of a long wavelength infrared camera to obtain high-speed thermal images of an aircraft engine’s in-service thermal barrier coated turbine blades. Long wavelength thermal images were captured of first-stage blades. The achieved temporal and spatial resolutions allowed for the identification of cooling-hole locations. The software and synchronization components of the system allowed for the selection of any blade on the turbine wheel, with tuning capability to image from leading edge to trailing edge. Its first application delivered calibrated thermal images as a function of turbine rotational speed at both steady state conditions and during engine transients. In advance of presenting these data for the purpose of understanding engine operation, this paper focuses on the components of the system, verification of high-speed synchronized operation, and the integration of the system with the commercial jet engine test bed. © 2014 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4903266] I. INTRODUCTION
Ceramic thermal barrier coatings (TBCs) continue to be developed to allow gas turbine engines to operate at higher turbine inlet gas temperatures.1, 2 When applied to the surface of metallic parts used in the hottest sections of modern propulsion and power systems, modern TBCs enable operation at higher energy efficiency than their predecessor technologies.3 However, engine designers only take into account about half of the possible temperature increase afforded by the thermal properties of current TBCs due to lack of processing reproducibility.3 As TBC materials and processing methods advance, engine development testing will push toward even higher turbine inlet gas temperatures. Confidence in TBC performance will allow the maximum energy efficiency to be reached. Such confidence will be gained only if accurate surface temperature mapping and condition monitoring of in-service components are obtainable in this extreme environment. For decades, turbine engines undergoing development testing have included ports for insertion of pointmeasurement radiation thermometry probes. With a singleelement high-speed detector, several individual points of surface temperature are measured on each metal super-alloy turbine blade as the blade passes through the field of view of the detector. Traditional instruments operate at near-infrared or visible wavelengths of the electromagnetic spectrum to take advantage of the transmissive properties of widely used solid silica and sapphire fiber optical materials.4 Inaccurate knowledge of the target emissivity is a well-known source of a) Author to whom correspondence should be addressed. Electronic mail:
[email protected]
b) This research performed while author was at Advanced Fuel Research, Inc.,
87 Church Street, East Hartford, Connecticut 06108, USA.
0034-6748/2014/85(12)/124902/7/$30.00
error during optical-based temperature measurements of turbine blades,5 but metal blades have been well characterized at these short wavelengths. During the history of TBC technology developments, researchers have shown that the ceramic materials used in turbine engines have low, variable emissivity and are partially transparent at the traditional wavelengths and into midinfrared wavelengths, but have quite high and stable emissivity in the long wavelength infrared (LWIR).6–10 However, challenges have been presented with regard to LWIR radiation thermometry of TBC turbine blades.11, 12 Accurate knowledge of LWIR emissivity is critical since small errors in emissivity lead to larger errors in the surface temperature measurement.11 Other considerations include surface emissivity changes during engine operation, optical penetration depth, and interference from combustion products and other radiating components.12 With these challenges in mind, pointmeasurement LWIR radiation thermometers have been constructed and installed on engines with TBC components.12–14 Troubling is that point measurements on turbine blades made with radiation thermometers are inherently difficult to assign to an absolute location on the target. In addition to potential errors in the design and manufacture of optics and related hardware, thermal expansion during engine warm up disturbs the on-engine mounted instrument components, the engine case and the target. Inaccuracies and thermal effects can shift the aim of the optical view from the design plan. Although camera systems are similarly prone to such errors, imaging enables identification of features on the target surface and thus aids in identifying precise locations. This advantage of imaging was shown on an industrial turbine engine for electric power generation with high-speed infrared cameras sensitive to near- and mid-infrared wavelengths.15 Turbine blades in industrial engines can rotate at more than 3600 revolutions
85, 124902-1
© 2014 AIP Publishing LLC
124902-2
Markham et al.
per minute (rpm).15 Very small design features were captured using these cameras at wavelengths where TBC is partially transparent with a relatively low and temperature-dependent emissivity. The LWIR thermal camera system presented herein was built to map in-service TBC surface temperatures by capturing quantitative images of TBC blades carried on the first stage turbine of a commercial aero propulsion engine. Typical turbine rotational speeds would be well above 10 000 rpm. The system operates at the narrow wavelength region where TBC emissivity is high and stable.12–14 The system provides thermal images of the in-service TBC turbine blades at temporal and spatial resolutions that are sufficient to enable the identification of blade cooling-hole locations. The synchronization component allows selection of any individual blade on the turbine wheel for examination, with tuning capability to image from leading to trailing edge. Its first application on an actual engine test delivered calibrated thermal images as a function of turbine rotational speed at both steady state conditions and during engine transients. In advance of presenting these data for the purpose of understanding engine operation, this paper describes: (1) the experimental setup including the LWIR camera, the optical system, the trigger synchronization, and the calibration of the system; (2) the utilities employed for image collection and processing; and (3) the integrated application on the commercial jet engine including review of the overall operation of the assembly. II. EXPERIMENTAL SETUP
A block diagram of the installed LWIR imaging radiation thermometer is shown in Figure 1. In the engine test cell are (1) the on-engine enclosure securing the LWIR camera and optics, and (2) the junction enclosure containing the triggering synchronization electronics and other instrumentation relays. The computer and monitor are in the engine control room. A. LWIR camera
A Sofradir-EC model IRE320VL camera based on a 320 × 256 pixel Stirling-cooled mercury cadmium telluride (MCT) focal plane array (FPA) was developed into the LWIR imaging radiation thermometer. MCT FPAs are well-known for high LWIR sensitivity and fast response. This camera provides user-selectable integration times down to less than a
FIG. 1. Block diagram of installed LWIR imaging radiation thermometer system.
Rev. Sci. Instrum. 85, 124902 (2014)
FIG. 2. Photograph of the LWIR camera (within red, cube-like enclosure) mounted on the engine. The engine nacelle is open for the photograph and closes completely for engine operation.
microsecond. As purchased, the camera engine (FPA, cooler, electronics) is within a metal-panel enclosure of dimensions about 25.4 × 18 × 16.5 cm (10 × 7 × 6.5 in.), weighs less than 6 kg (13 lb), and is rated for operational temperature from −30 ◦ C to 55 ◦ C (−22 ◦ F to 131 ◦ F). The camera was mounted at the top of the engine as shown in the photograph of Figure 2, within the cube-like enclosure (red). At this location the air temperature between the engine surface and the inner wall of its nacelle was expected to exceed the camera’s upper operational temperature limit. To account for this, three adjacent panels of the camera’s original enclosure were replaced with a trifold sheet of copper that included a single length of copper tubing brazed to its surface for cooling by flowing water. The cooling tube was tailored to the trifold sheet with several u-bends to provide a contact length of about 254 cm (100 in.), and was brazed to the in-facing surface so as to maintain the outside dimensions of the original enclosure. A fourth metal panel was modified by adding a three-input manifold to provide a copious flow of nitrogen gas into and through the camera enclosure, also for the purpose of cooling. The modified enclosure was then insulated with 12.7 mm (0.5 in.) thick temperatureresistant silicone sponge rubber of firm density (red). Cables and wires extending from the camera were protected in a silicone coated fiberglass bellows (rectangular-to-round cuff) that transitioned to a flexible conduit. The conduit, which is visible near the top of Figure 2, extended to the outside of the nacelle. The bellows and conduit were also utilized as an exit path for a portion of the nitrogen gas forced into the camera enclosure, thus maintaining temperatures within the bellows and conduit to within the allowable limits for the cables. The camera enclosure was connected to the engine by replacing a fifth metal panel, the front panel where light enters the camera. The replacement panel included four broad temperature range vibration mounts (Lord Corporation). Through these mounts, the enclosure was bolted to a custom on-camera mating flange designed with internal channels for watercooling. The water-cooled flange bolted onto an on-engine
124902-3
Markham et al.
Rev. Sci. Instrum. 85, 124902 (2014)
mating flange that aligned the camera with the engine probe. A narrow space was incorporated between the custom front panel and the water-cooled flange to allow for the exit of a portion of the nitrogen gas that was forced into the camera enclosure. The exiting cooling gas flowed in the direction of the engine’s exterior surface, blanketing certain optical components that allowed light in from the engine probe (discussed below). Numerous sensors were installed for monitoring the conditions experienced by the camera. Four thermocouples within the enclosure monitored points on the electronics and optics. One thermocouple at the exit of the enclosure measured the gas temperature entering the conduit. A pair of thermocouples measured the inlet and outlet temperatures of the cooling water flowing through the brazed tubing. A second pair of thermocouples monitored the on-camera mating flange water. Two accelerometers within the enclosure, on the front panel and on an optic-holding block, monitored vibrations. Finally, pressure sensors monitored the nitrogen gas supply to the enclosure and the city-water supply to the water-cooled parts. B. Optics and engine probe
A lens system and on-engine probe were designed to transfer the LWIR image of the target surface within the en-
gine to the FPA of the camera. A schematic is shown in Figure 3, which also includes structures described above. For maximum throughput, the prism, lenses and pressure window were manufactured from zinc selenide (ZnSe) with an antireflection coating. The prism and adjacent lenses up to the pressure window were secured by a metal sleeve that was secured within a housing. When installed on engine, only the tip of the housing where the prism was located protruded through the stator wall into the hot turbine inlet gas path. The tip of the housing was therefore protected with TBC. The probe entered the engine’s gas path upstream of the first stage TBC blades, viewing aft from the probe aperture. The internal reflection angle of the prism was slightly more than 90◦ . For this particular installation, the working distance from the blade to the probe aperture was about 5 cm (2 in.). The optical system achieved a nominal spatial resolution of 0.51 mm (0.020 in.) at the typical working distance when viewing nearnormal regions of a blade’s pressure surface. Within the housing and around the metal sleeve a flow of clean nitrogen gas was maintained by pressurizing the housing. The pressure window forced the flow of the clean nitrogen down the outside length of the metal sleeve and around the prism to then exit at the view-aperture into the turbine inlet flow. The clean gas worked to keep the hardware cool and also keep combustion gases and particles from entering the probe and fouling the prism (the only exposed optic in
FIG. 3. Cutaway schematic of the probe-mounted optical system.
124902-4
Markham et al.
Rev. Sci. Instrum. 85, 124902 (2014)
FIG. 4. Left: metal blades of turbo molecular pump. Right: LWIR images of one leading edge at several wheel speeds.
the system).16 Also notable is that the pair of lenses between the pressure window and FPA were mounted on an electronic positioning device for fine-tuning the focus of the image. C. Synchronization
The synchronization hardware provides a logic-level pulse to the camera when the turbine blade is in the desired position relative to the probe. To generate this pulse, the hardware utilizes a tachometer (once-per-revolution) signal provided by the engine. One could simply trigger the camera directly, using the tachometer signal, but then all of the images would be of the same position on the spool. The synchronization hardware is therefore required to provide varying phase angles between the tachometer signal and the signal sent to the camera. To generate the signal for the camera, the synchronization hardware utilizes a phase-locked loop circuit to generate a signal at 2200 times the frequency of the engine rotation. As the two signals are locked in phase, the high-frequency signal indicates 2200 locations around the spool, with an average step of 0.16◦ between locations. This provides for fine positioning of the selected blade in the field-of-view. The signal for the camera is generated by a microcontroller whose hardware counter is dividing the frequency within the phase-locked loop. Hardware features allow one of the digital pins of the microcontroller to transition at a software-selectable count. This signal is gated in hardware using another digital pin to maintain a nearly constant frame rate for the camera. That is, the gating signal reduces the camera trigger frequency from the once per revolution signal generated by the counter to a rate at which the camera can collect images. The microcontroller also serves to control the remainder of the synchronization electronics and to communicate via a serial link with the computer in the test cell control room. Commands sent from the computer select the trigger count and the phase detector used. The microcontroller in turn sends information back to the control room, including the measured engine speed and the lock state of the loop. The synchronization component was tested by viewing the turbine wheel of a turbo molecular pump. The pump head was fitted with an infrared viewport for optical access to the blades. A photograph of the pump’s turbine blades is shown
in Figure 4. The leading edge of each blade is the flat surface facing the camera, each being about 1.5 mm (0.06 in.) wide. Figure 4 also shows a sequence of near-room temperature LWIR images for different pump speeds, starting at 60 000 rpm (at left) and going down by 10 000 rpm increments to 10 000 rpm. Each image is a 1 s (100-frame, 1.08 μs integration each frame) average; taken in the order shown. The dark vertical stripe is the leading edge, and it appears slightly cooler than the rest of the blade and its surrounding background because the take-off angle is approximately normal to the surface. Because the blade is fairly reflective (low emissivity) the camera is also staring back into itself and “seeing” the cooled FPA. The leading edge has a mottled appearance because it is not a mirror-like surface. To be sure that the LWIR system was synchronized on a particular blade, a spot of ceramic paint (high emissivity in the LWIR) was placed in the center of one leading edge. The spot of about 0.75 mm (0.030 in.) diameter is clearly distinguishable on the leading edge at all speeds. Some pump warm up is noticeable, with the 10 000 rpm background being slightly warmer than the 60 000 rpm background (the temperature scales shown are in ◦ C; the scale to the left is for the series of moving blades). For comparison, the rightmost image is an averaged image of the same blade while stationary. The integration time blur, which is about 10 pixels wide at 60 000 rpm drops to 1.7 pixels in the 10 000 rpm image. Taking into account the slight differences in temperature (and temperature scale), it is clear that the 10 000 rpm image looks about the same as the stationary image and that the blur is not excessive even at the highest speed. D. Calibration
The LWIR imaging radiation thermometer was calibrated over the range of 300 ◦ C (572 ◦ F) to 1420 ◦ C (2588 ◦ F) with a custom calibration source to satisfy high temperature and field-of-view requirements. A cylindro-conical cavity17 of 5.08 cm (2 in.) inside diameter was designed, fabricated, and installed within a commercially available high-temperature tube furnace (Carbolite STF/450). The cavity was made of silicon carbide (SiC) for the benefit of good thermal conductivity (i.e., good temperature uniformity when heated within the tube furnace). For a flat surface of SiC, the normal spectral emittance in the narrow band pass of the LWIR imaging
124902-5
Markham et al.
radiation thermometer is typically about 0.9. The effective spectral emittance measured at the entrance aperture of the fabricated cavity was greater than 0.99 for the critical band pass. Calibrated thermal images from the cavity were obtained at several temperatures as a performance-check of the LWIR imaging radiation thermometer. Images from the cavity were generally uniform at ±1 ◦ C (±2 ◦ F) for the field-of-view and within ±5 ◦ C (±9 ◦ F) of the tube furnace temperature setting. III. IMAGE COLLECTION AND PROCESSING
Three separate utilities were employed for image collection and processing: the collect, calibration, and view programs. The collect program performs data collection from the camera and provides real-time viewing of the thermal images. Calibration calculations are performed by the calibration program using data stored by the collect program. Finally, the view program is used to review and/or re-process the saved data. Both the collect and view programs utilize the calibration data to generate temperatures from the raw images. The division of features among the three programs greatly simplifies the individual user interfaces and enhances the performance of each section of code. The operation of each program is now described. The collect program is used to acquire data from the focal plane array camera, save the data to disk, display raw or thermal data in real time, and to control synchronization with the engine under test. The program establishes a continuous “grab” from the camera to a circular buffer in memory. The status of the circular buffer is monitored by a dedicated thread that copies the images to a large save buffer as they become available. This thread also copies the images to a display buffer when the display code indicates that it is ready to display the next frame. The program displays the collected data in real time as long as the grab is active. A radio button is used to select between the display of raw image data and temperature data. Temperature data are calculated using a calibration file loaded by the user. For both the raw data and the temperature data, the results are shown in false “thermal” color. The display range can be adjusted manually or automatically scaled to the frame currently being displayed. Two save modes are available, burst and continuous. In burst mode, a single file is saved with a specified number of frames. Burst mode is intended to capture short lengths of video that can be averaged for calibration or to capture short transients that are controlled by the user. In continuous mode, all of the data collected are streamed to disk. A user-specified number of frames is stored in each file. In both save modes, raw data are stored, rather than the derived temperature data. Temperature data can be regenerated using the view program. The calibration program is used to generate a calibration file for use by the collect and view programs. The calibration file contains data on the response of each pixel and allows the other programs to convert the raw pixel values to temperature values. The program allows the user to select a number of files (usually collected in burst mode) and assign a temperature to each file. The frames in each file are averaged and the pixel response is fitted with a polynomial over the known radi-
Rev. Sci. Instrum. 85, 124902 (2014)
ance values for each file. These radiance values are computed from the assigned file temperatures, the specified pyrometer filter wavelength, and the specified emissivity of the calibration source. The view program is used to recall, process, and display saved data. It utilizes the same display controls as the collect program, allowing the user to view the raw data or temperature data in false color with an adjustable display range. The program can be used to display single frames or continuous video. Background loading of the image files allows for uninterrupted video display. The program also reports the current file and auxiliary data stored with it, and it allows the user to seek a particular frame number rather than scrolling through the video to find that particular frame. IV. APPLICATION AND DISCUSSION
Installation and the first collection of real-time LWIR images of in-service TBC blades took place in February 2013. Data were collected over a 3 h run period for operating conditions from idle to take-off. Thousands of calibrated LWIR thermal images of TBC blades were obtained as a function of turbine rotational speed at both steady state conditions and during engine transients, with each image assignable to a blade number on the turbine wheel. Real-time images identified blades that had lost TBC from areas of their leading edge. An example of data near the trailing edge of a TBC blade is shown in Figure 5. For this area of the blade the entire view is well within the range of take-off angles about normal where LWIR emissivity is high and stable for the TBC. It is noted that complex blade curvature, e.g., around the leading edge, can present LWIR image areas with view angles that approach grazing where TBC emissivity decreases (reflectivity increases).18 The thermal image on the left is an average of 80 images of the targeted blade; a row of cooling holes is prominent. The color scale of ±80 ◦ F is the relative change from the average of the image. The isotherm contours on the right expound a hotter surface in front of cooling holes and in the upper right of the image, with cooler regions at the lower right and close to the cooling holes. The range of false thermal colors demonstrates the ability of the system to monitor TBC response in the combustion environment and to monitor the effect of internal passages and film cooling on the surface temperatures. The measurement technique described in this paper provides an ability to monitor Bill of Material (BOM) blades in their actual operating environment in order to measure engine transient response and to evaluate different cooling designs, manufacturing methods, etc. Data from this three-hour run period will be the subject of companion publications which will focus on using the data for understanding engine operation. As detailed in earlier publications,12, 13 the combustion products at high temperatures and pressures in the fieldof-view influence the thermal intensity detected by the FPA. The approximately 1-μm wide LWIR band pass of the system was chosen not only to exploit the near-unity emissivity of TBC but to also minimize radiance contributions from the hot gases. The HITEMP gas spectra19 for H2 O and CO2 have previously been used to calculate the error contribution to an
124902-6
Markham et al.
Rev. Sci. Instrum. 85, 124902 (2014)
FIG. 5. LWIR data from an in-service TBC blade rotating at typical engine speeds. Data are presented as difference from average in ◦ F. At left a row of cooling holes is apparent in the circular thermal color image. At right an isothermal contour plot of the same data is presented where the white lines are the average temperature contour, the blue lines are lower in temp, and the red and orange lines are higher.
LWIR radiation thermometer.12, 14 The calculation considers the emissivity of the surface of interest, the attenuation of surface radiance by the hot gases, and the radiance contribution by the hot gases. Gas temperature, pressure, concentration, and path length are input parameters. Results of the calculation were presented for operating parameters appropriate to an industrial turbine (e.g., inlet pressure of 15 bars) for path lengths from 5 to 15 cm (2 to 6 in.)12 and 100 cm (39.4 in.).14 Similar calculations were performed for the present work, with estimations of the aircraft engine’s proprietary operating parameters as input. The results showed that for the measurement geometry describe above, the cooling holes and surface temperature patterns would be readily apparent in the LWIR images and therefore these features would not be obscured by interferences in the intervening combustion flow. After the fact, it is notable that the intensity variation in pixel location for the numerous images that were averaged to generate Figure 5 show that the information in the figure is not due to random effect, but is a statistically significant resolution of TBC surface temperature. Associated research and development is in progress to non-intrusively measure combustion gas temperatures at the turbine inlet with an infrared-based spectroscopic technique.20 For future work, the in situ spectral information from the turbine inlet gas temperature determination will be investigated as a means for real-time accounting of the combustion gas radiance contribution in the bandpass of the LWIR imaging radiation thermometer. The LWIR system described in this article has applications beyond the one just described. The typical thermal resolution of modern LWIR cameras is in the tens of milliKelvin. For instance, the thermal resolution of the camera used here is less than 20 mK. Since peak emissions of room temperature objects fall in the LWIR band, the LWIR imaging radiation thermometer is also suitable for monitoring blades coated with ceramic (or instead, coated with a low cost, high emissivity paint) in low temperature turbine rigs used for studying aerodynamics and heat transfer. Many turbine rigs are set-up
so that blade cooling gas comes in below room temperature, and turbine inlet gas comes in above so that the blade is about room temperature.21 V. CONCLUSION
A long wavelength infrared imaging radiation thermometer was developed and implemented to monitor the surface temperature of in-service TBC turbine blades. Its first application on an actual engine test delivered calibrated thermal images as a function of turbine rotational speed at both steady state conditions and during engine transients. Temporal and spatial resolutions achieved provided cooling-hole locations in the images at aero propulsion engine rotational speeds. The instrument provides capability to monitor TBC blades in their actual operating environment in order to measure engine transient response and to evaluate different cooling designs, manufacturing methods, processing reproducibility, and ultimately to gain confidence in TBC performance at maximum energy efficiency of engines. ACKNOWLEDGMENTS
Support to Advanced Fuel Research, Inc. for technology development includes Air Force SBIR Contract Nos. FA9101-09-C-0019 and FA9101-12-C-0025. For numerous beneficial technical interactions, gratitude goes to: Robert Howard and Bradley Winkleman of the Arnold Engineering Development Complex; Ruth Sikorski of the Air Force Research Laboratory; Bryan Engi of Infotech Enterprises America, Inc.; and Christopher Lehane, Andrew Consiglio, Keith Nachilly, Bruce Hockaday, Darren Wind, Allison Nicklous, Doug Thomesen, Martin Shenkle and Joel Wagner of Pratt & Whitney. 1 A.
U. Munawar, U. Shulz, and G. Cerri, “Microstructural evolution of GdZ and DySZ based EB-PVD TBC systems after thermal cycling at high temperature,” J. Eng. Gas Turbines Power 135(10), 102101 (2013).
124902-7 2 M.
Markham et al.
van Roode and A. K. Bhattacharya, “Durability of oxide/oxide ceramic matrix composites in gas turbine combustors,”J. Eng. Gas Turbines Power 135(5), 051301 (2013). 3 D. R. Clarke, M. Oechsner, and N. P. Padture, “Thermal barrier coatings for more efficient gas-turbine engines,” MRS Bull. 37, 891–897 (2012). 4 R. Vanzetti and A. C. Traub, in Theory and Practice of Radiation Thermometry, edited by D. P. DeWitt and G. P. Nutter (Wiley, New York, 1988), Chap. 11. 5 C. Kerr and P. Ivey, “An overview of the measurement errors associated with gas turbine aeroengine pyrometer systems,” Meas. Sci. Technol. 13, 873–881 (2002). 6 W. H. Atkinson and M. E. Cyr, “Evaluation of sensors for temperature measurement of ceramic materials,” in Proceedings of the 3rd Annual HITEMP Review (NASA Conference Publication 10051, 1990). 7 J. R. Markham and K. Kinsella, “Thermal radiative properties and temperature measurement from turbine coatings,” Int. J. Thermophys. 19(2), 537–545 (1998). 8 H. Latvakoski, J. Markham, M. Borden, T. Hawkins, and M. Cybulski, “Measurement of advanced ceramic coated superalloys with a long wavelength pyrometer,” AIAA Paper 2000-2212, 2000. 9 F. E. Pfefferkorn, F. P. Incropera, and Y. C. Shin, “Surface temperature measurement of semi-transparent ceramics by long-wavelength pyrometry,” J. Heat Transfer 125, 48–56 (2003). 10 J. I. Eldridge and C. M. Spuckler, “Determination of scattering and absorption coefficients for plasma-sprayed yttria-stabilized zirconia thermal barrier coatings,” J. Am. Ceram. Soc. 91(5), 1603–1611 (2008). 11 W. H. Atkinson, M. A. Cyr, and R. R. Strange, “Development of sensors for ceramic components in advanced propulsion systems. Phase II: Temperature sensor systems evaluation,” Report No. NASA-CR-195283, 1994.
Rev. Sci. Instrum. 85, 124902 (2014) 12 J.
Markham, H. Latvakowski, D. Marran,J. Neira, P. Kenny, and P. Best, “Challenges of radiation thermometry for advanced gas turbine engines,” in Proceedings of the 46th International Instrumentation Symposium, Bellevue, WA (ISA, 2000), Vol. 397, pp. 299–313. 13 J. R. Markham, H. M. Latvakoski, S. L. F. Frank, and M. Ludtke, “Simultaneous short and long wavelength pyrometer measurements in a heavyduty gas turbine,”J. Eng. Gas Turbines Power (Trans. ASME) 124, 528– 533 (2002). 14 Pyrometry of Combustion Turbine Blades with Thermal Barrier Coating (EPRI, Palo Alto, CA, and KEMA Nederland B.V., Arnhem, Netherlands, 2002) (Publicly available effective December 6, 2006). 15 H.-G. Brummel, D. LeMieux, M. Voigt, and P. Zombo, “On-line turbine blade monitoring,” Power 149(7), 11 (2005). 16 C. Kerr and P. Ivey, “A review of purge air designs for aeroengine-based optical pyrometers,” J. Turbomach. 124, 227–234 (2002). 17 R. E. Bedford, in Theory and Practice of Radiation Thermometry, edited by D. P. DeWitt and G. P. Nutter (Wiley, New York, 1988), Chap. 12. 18 R. J. Chandos and R. E. Chandos, “Radiometric properties of isothermal, diffuse wall cavity sources,” Appl. Opt. 13(9), 2142–2152 (1974). 19 L. S. Rothman, I. E. Gordon, A. Barbe et al., “The HITRAN 2008 molecular spectroscopic database,” J. Quant. Spectrosc. Radiat. Transfer 110, 533– 572 (2009). 20 J. Scire, J. Cosgrove, and J. Markham, “Turbine inlet gas temperature measurement system,” in Proceedings of the 57th International Instrumentation Symposium, St. Louis, MO (ISA, 2011), Vol. 488, pp. 295–307. 21 R. M. Mathison, C. W. Haldeman, and M. G. Dunn, “Aerodynamics and heat transfer for a cooled one and one-half stage high-pressure turbine – Part I: Vane inlet temperature profile generation and migration,” in Proceedings of the ASME Turbo Expo, Glasgow, Scotland, ASME GT201022716 (ASME, 2010), pp. 299–312.
Review of Scientific Instruments is copyrighted by the American Institute of Physics (AIP). Redistribution of journal material is subject to the AIP online journal license and/or AIP copyright. For more information, see http://ojps.aip.org/rsio/rsicr.jsp
Aircraft engine-mounted camera system for long wavelength infrared imaging of in-service thermal barrier coated turbine blades James Markham,1,a) Joseph Cosgrove,1 James Scire, Jr.,2,b) Charles Haldeman,3 and Ian Agoos3 1
Advanced Fuel Research, Inc., 87 Church Street, East Hartford, Connecticut 06108, USA New York Institute of Technology, Northern Blvd, Old Westbury, New York 11568, USA 3 Pratt & Whitney, 400 Main Street, East Hartford, Connecticut 06108, USA 2
(Received 19 September 2014; accepted 20 November 2014; published online 5 December 2014) This paper announces the implementation of a long wavelength infrared camera to obtain high-speed thermal images of an aircraft engine’s in-service thermal barrier coated turbine blades. Long wavelength thermal images were captured of first-stage blades. The achieved temporal and spatial resolutions allowed for the identification of cooling-hole locations. The software and synchronization components of the system allowed for the selection of any blade on the turbine wheel, with tuning capability to image from leading edge to trailing edge. Its first application delivered calibrated thermal images as a function of turbine rotational speed at both steady state conditions and during engine transients. In advance of presenting these data for the purpose of understanding engine operation, this paper focuses on the components of the system, verification of high-speed synchronized operation, and the integration of the system with the commercial jet engine test bed. © 2014 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4903266] I. INTRODUCTION
Ceramic thermal barrier coatings (TBCs) continue to be developed to allow gas turbine engines to operate at higher turbine inlet gas temperatures.1, 2 When applied to the surface of metallic parts used in the hottest sections of modern propulsion and power systems, modern TBCs enable operation at higher energy efficiency than their predecessor technologies.3 However, engine designers only take into account about half of the possible temperature increase afforded by the thermal properties of current TBCs due to lack of processing reproducibility.3 As TBC materials and processing methods advance, engine development testing will push toward even higher turbine inlet gas temperatures. Confidence in TBC performance will allow the maximum energy efficiency to be reached. Such confidence will be gained only if accurate surface temperature mapping and condition monitoring of in-service components are obtainable in this extreme environment. For decades, turbine engines undergoing development testing have included ports for insertion of pointmeasurement radiation thermometry probes. With a singleelement high-speed detector, several individual points of surface temperature are measured on each metal super-alloy turbine blade as the blade passes through the field of view of the detector. Traditional instruments operate at near-infrared or visible wavelengths of the electromagnetic spectrum to take advantage of the transmissive properties of widely used solid silica and sapphire fiber optical materials.4 Inaccurate knowledge of the target emissivity is a well-known source of a) Author to whom correspondence should be addressed. Electronic mail:
[email protected]
b) This research performed while author was at Advanced Fuel Research, Inc.,
87 Church Street, East Hartford, Connecticut 06108, USA.
0034-6748/2014/85(12)/124902/7/$30.00
error during optical-based temperature measurements of turbine blades,5 but metal blades have been well characterized at these short wavelengths. During the history of TBC technology developments, researchers have shown that the ceramic materials used in turbine engines have low, variable emissivity and are partially transparent at the traditional wavelengths and into midinfrared wavelengths, but have quite high and stable emissivity in the long wavelength infrared (LWIR).6–10 However, challenges have been presented with regard to LWIR radiation thermometry of TBC turbine blades.11, 12 Accurate knowledge of LWIR emissivity is critical since small errors in emissivity lead to larger errors in the surface temperature measurement.11 Other considerations include surface emissivity changes during engine operation, optical penetration depth, and interference from combustion products and other radiating components.12 With these challenges in mind, pointmeasurement LWIR radiation thermometers have been constructed and installed on engines with TBC components.12–14 Troubling is that point measurements on turbine blades made with radiation thermometers are inherently difficult to assign to an absolute location on the target. In addition to potential errors in the design and manufacture of optics and related hardware, thermal expansion during engine warm up disturbs the on-engine mounted instrument components, the engine case and the target. Inaccuracies and thermal effects can shift the aim of the optical view from the design plan. Although camera systems are similarly prone to such errors, imaging enables identification of features on the target surface and thus aids in identifying precise locations. This advantage of imaging was shown on an industrial turbine engine for electric power generation with high-speed infrared cameras sensitive to near- and mid-infrared wavelengths.15 Turbine blades in industrial engines can rotate at more than 3600 revolutions
85, 124902-1
© 2014 AIP Publishing LLC
124902-2
Markham et al.
per minute (rpm).15 Very small design features were captured using these cameras at wavelengths where TBC is partially transparent with a relatively low and temperature-dependent emissivity. The LWIR thermal camera system presented herein was built to map in-service TBC surface temperatures by capturing quantitative images of TBC blades carried on the first stage turbine of a commercial aero propulsion engine. Typical turbine rotational speeds would be well above 10 000 rpm. The system operates at the narrow wavelength region where TBC emissivity is high and stable.12–14 The system provides thermal images of the in-service TBC turbine blades at temporal and spatial resolutions that are sufficient to enable the identification of blade cooling-hole locations. The synchronization component allows selection of any individual blade on the turbine wheel for examination, with tuning capability to image from leading to trailing edge. Its first application on an actual engine test delivered calibrated thermal images as a function of turbine rotational speed at both steady state conditions and during engine transients. In advance of presenting these data for the purpose of understanding engine operation, this paper describes: (1) the experimental setup including the LWIR camera, the optical system, the trigger synchronization, and the calibration of the system; (2) the utilities employed for image collection and processing; and (3) the integrated application on the commercial jet engine including review of the overall operation of the assembly. II. EXPERIMENTAL SETUP
A block diagram of the installed LWIR imaging radiation thermometer is shown in Figure 1. In the engine test cell are (1) the on-engine enclosure securing the LWIR camera and optics, and (2) the junction enclosure containing the triggering synchronization electronics and other instrumentation relays. The computer and monitor are in the engine control room. A. LWIR camera
A Sofradir-EC model IRE320VL camera based on a 320 × 256 pixel Stirling-cooled mercury cadmium telluride (MCT) focal plane array (FPA) was developed into the LWIR imaging radiation thermometer. MCT FPAs are well-known for high LWIR sensitivity and fast response. This camera provides user-selectable integration times down to less than a
FIG. 1. Block diagram of installed LWIR imaging radiation thermometer system.
Rev. Sci. Instrum. 85, 124902 (2014)
FIG. 2. Photograph of the LWIR camera (within red, cube-like enclosure) mounted on the engine. The engine nacelle is open for the photograph and closes completely for engine operation.
microsecond. As purchased, the camera engine (FPA, cooler, electronics) is within a metal-panel enclosure of dimensions about 25.4 × 18 × 16.5 cm (10 × 7 × 6.5 in.), weighs less than 6 kg (13 lb), and is rated for operational temperature from −30 ◦ C to 55 ◦ C (−22 ◦ F to 131 ◦ F). The camera was mounted at the top of the engine as shown in the photograph of Figure 2, within the cube-like enclosure (red). At this location the air temperature between the engine surface and the inner wall of its nacelle was expected to exceed the camera’s upper operational temperature limit. To account for this, three adjacent panels of the camera’s original enclosure were replaced with a trifold sheet of copper that included a single length of copper tubing brazed to its surface for cooling by flowing water. The cooling tube was tailored to the trifold sheet with several u-bends to provide a contact length of about 254 cm (100 in.), and was brazed to the in-facing surface so as to maintain the outside dimensions of the original enclosure. A fourth metal panel was modified by adding a three-input manifold to provide a copious flow of nitrogen gas into and through the camera enclosure, also for the purpose of cooling. The modified enclosure was then insulated with 12.7 mm (0.5 in.) thick temperatureresistant silicone sponge rubber of firm density (red). Cables and wires extending from the camera were protected in a silicone coated fiberglass bellows (rectangular-to-round cuff) that transitioned to a flexible conduit. The conduit, which is visible near the top of Figure 2, extended to the outside of the nacelle. The bellows and conduit were also utilized as an exit path for a portion of the nitrogen gas forced into the camera enclosure, thus maintaining temperatures within the bellows and conduit to within the allowable limits for the cables. The camera enclosure was connected to the engine by replacing a fifth metal panel, the front panel where light enters the camera. The replacement panel included four broad temperature range vibration mounts (Lord Corporation). Through these mounts, the enclosure was bolted to a custom on-camera mating flange designed with internal channels for watercooling. The water-cooled flange bolted onto an on-engine
124902-3
Markham et al.
Rev. Sci. Instrum. 85, 124902 (2014)
mating flange that aligned the camera with the engine probe. A narrow space was incorporated between the custom front panel and the water-cooled flange to allow for the exit of a portion of the nitrogen gas that was forced into the camera enclosure. The exiting cooling gas flowed in the direction of the engine’s exterior surface, blanketing certain optical components that allowed light in from the engine probe (discussed below). Numerous sensors were installed for monitoring the conditions experienced by the camera. Four thermocouples within the enclosure monitored points on the electronics and optics. One thermocouple at the exit of the enclosure measured the gas temperature entering the conduit. A pair of thermocouples measured the inlet and outlet temperatures of the cooling water flowing through the brazed tubing. A second pair of thermocouples monitored the on-camera mating flange water. Two accelerometers within the enclosure, on the front panel and on an optic-holding block, monitored vibrations. Finally, pressure sensors monitored the nitrogen gas supply to the enclosure and the city-water supply to the water-cooled parts. B. Optics and engine probe
A lens system and on-engine probe were designed to transfer the LWIR image of the target surface within the en-
gine to the FPA of the camera. A schematic is shown in Figure 3, which also includes structures described above. For maximum throughput, the prism, lenses and pressure window were manufactured from zinc selenide (ZnSe) with an antireflection coating. The prism and adjacent lenses up to the pressure window were secured by a metal sleeve that was secured within a housing. When installed on engine, only the tip of the housing where the prism was located protruded through the stator wall into the hot turbine inlet gas path. The tip of the housing was therefore protected with TBC. The probe entered the engine’s gas path upstream of the first stage TBC blades, viewing aft from the probe aperture. The internal reflection angle of the prism was slightly more than 90◦ . For this particular installation, the working distance from the blade to the probe aperture was about 5 cm (2 in.). The optical system achieved a nominal spatial resolution of 0.51 mm (0.020 in.) at the typical working distance when viewing nearnormal regions of a blade’s pressure surface. Within the housing and around the metal sleeve a flow of clean nitrogen gas was maintained by pressurizing the housing. The pressure window forced the flow of the clean nitrogen down the outside length of the metal sleeve and around the prism to then exit at the view-aperture into the turbine inlet flow. The clean gas worked to keep the hardware cool and also keep combustion gases and particles from entering the probe and fouling the prism (the only exposed optic in
FIG. 3. Cutaway schematic of the probe-mounted optical system.
124902-4
Markham et al.
Rev. Sci. Instrum. 85, 124902 (2014)
FIG. 4. Left: metal blades of turbo molecular pump. Right: LWIR images of one leading edge at several wheel speeds.
the system).16 Also notable is that the pair of lenses between the pressure window and FPA were mounted on an electronic positioning device for fine-tuning the focus of the image. C. Synchronization
The synchronization hardware provides a logic-level pulse to the camera when the turbine blade is in the desired position relative to the probe. To generate this pulse, the hardware utilizes a tachometer (once-per-revolution) signal provided by the engine. One could simply trigger the camera directly, using the tachometer signal, but then all of the images would be of the same position on the spool. The synchronization hardware is therefore required to provide varying phase angles between the tachometer signal and the signal sent to the camera. To generate the signal for the camera, the synchronization hardware utilizes a phase-locked loop circuit to generate a signal at 2200 times the frequency of the engine rotation. As the two signals are locked in phase, the high-frequency signal indicates 2200 locations around the spool, with an average step of 0.16◦ between locations. This provides for fine positioning of the selected blade in the field-of-view. The signal for the camera is generated by a microcontroller whose hardware counter is dividing the frequency within the phase-locked loop. Hardware features allow one of the digital pins of the microcontroller to transition at a software-selectable count. This signal is gated in hardware using another digital pin to maintain a nearly constant frame rate for the camera. That is, the gating signal reduces the camera trigger frequency from the once per revolution signal generated by the counter to a rate at which the camera can collect images. The microcontroller also serves to control the remainder of the synchronization electronics and to communicate via a serial link with the computer in the test cell control room. Commands sent from the computer select the trigger count and the phase detector used. The microcontroller in turn sends information back to the control room, including the measured engine speed and the lock state of the loop. The synchronization component was tested by viewing the turbine wheel of a turbo molecular pump. The pump head was fitted with an infrared viewport for optical access to the blades. A photograph of the pump’s turbine blades is shown
in Figure 4. The leading edge of each blade is the flat surface facing the camera, each being about 1.5 mm (0.06 in.) wide. Figure 4 also shows a sequence of near-room temperature LWIR images for different pump speeds, starting at 60 000 rpm (at left) and going down by 10 000 rpm increments to 10 000 rpm. Each image is a 1 s (100-frame, 1.08 μs integration each frame) average; taken in the order shown. The dark vertical stripe is the leading edge, and it appears slightly cooler than the rest of the blade and its surrounding background because the take-off angle is approximately normal to the surface. Because the blade is fairly reflective (low emissivity) the camera is also staring back into itself and “seeing” the cooled FPA. The leading edge has a mottled appearance because it is not a mirror-like surface. To be sure that the LWIR system was synchronized on a particular blade, a spot of ceramic paint (high emissivity in the LWIR) was placed in the center of one leading edge. The spot of about 0.75 mm (0.030 in.) diameter is clearly distinguishable on the leading edge at all speeds. Some pump warm up is noticeable, with the 10 000 rpm background being slightly warmer than the 60 000 rpm background (the temperature scales shown are in ◦ C; the scale to the left is for the series of moving blades). For comparison, the rightmost image is an averaged image of the same blade while stationary. The integration time blur, which is about 10 pixels wide at 60 000 rpm drops to 1.7 pixels in the 10 000 rpm image. Taking into account the slight differences in temperature (and temperature scale), it is clear that the 10 000 rpm image looks about the same as the stationary image and that the blur is not excessive even at the highest speed. D. Calibration
The LWIR imaging radiation thermometer was calibrated over the range of 300 ◦ C (572 ◦ F) to 1420 ◦ C (2588 ◦ F) with a custom calibration source to satisfy high temperature and field-of-view requirements. A cylindro-conical cavity17 of 5.08 cm (2 in.) inside diameter was designed, fabricated, and installed within a commercially available high-temperature tube furnace (Carbolite STF/450). The cavity was made of silicon carbide (SiC) for the benefit of good thermal conductivity (i.e., good temperature uniformity when heated within the tube furnace). For a flat surface of SiC, the normal spectral emittance in the narrow band pass of the LWIR imaging
124902-5
Markham et al.
radiation thermometer is typically about 0.9. The effective spectral emittance measured at the entrance aperture of the fabricated cavity was greater than 0.99 for the critical band pass. Calibrated thermal images from the cavity were obtained at several temperatures as a performance-check of the LWIR imaging radiation thermometer. Images from the cavity were generally uniform at ±1 ◦ C (±2 ◦ F) for the field-of-view and within ±5 ◦ C (±9 ◦ F) of the tube furnace temperature setting. III. IMAGE COLLECTION AND PROCESSING
Three separate utilities were employed for image collection and processing: the collect, calibration, and view programs. The collect program performs data collection from the camera and provides real-time viewing of the thermal images. Calibration calculations are performed by the calibration program using data stored by the collect program. Finally, the view program is used to review and/or re-process the saved data. Both the collect and view programs utilize the calibration data to generate temperatures from the raw images. The division of features among the three programs greatly simplifies the individual user interfaces and enhances the performance of each section of code. The operation of each program is now described. The collect program is used to acquire data from the focal plane array camera, save the data to disk, display raw or thermal data in real time, and to control synchronization with the engine under test. The program establishes a continuous “grab” from the camera to a circular buffer in memory. The status of the circular buffer is monitored by a dedicated thread that copies the images to a large save buffer as they become available. This thread also copies the images to a display buffer when the display code indicates that it is ready to display the next frame. The program displays the collected data in real time as long as the grab is active. A radio button is used to select between the display of raw image data and temperature data. Temperature data are calculated using a calibration file loaded by the user. For both the raw data and the temperature data, the results are shown in false “thermal” color. The display range can be adjusted manually or automatically scaled to the frame currently being displayed. Two save modes are available, burst and continuous. In burst mode, a single file is saved with a specified number of frames. Burst mode is intended to capture short lengths of video that can be averaged for calibration or to capture short transients that are controlled by the user. In continuous mode, all of the data collected are streamed to disk. A user-specified number of frames is stored in each file. In both save modes, raw data are stored, rather than the derived temperature data. Temperature data can be regenerated using the view program. The calibration program is used to generate a calibration file for use by the collect and view programs. The calibration file contains data on the response of each pixel and allows the other programs to convert the raw pixel values to temperature values. The program allows the user to select a number of files (usually collected in burst mode) and assign a temperature to each file. The frames in each file are averaged and the pixel response is fitted with a polynomial over the known radi-
Rev. Sci. Instrum. 85, 124902 (2014)
ance values for each file. These radiance values are computed from the assigned file temperatures, the specified pyrometer filter wavelength, and the specified emissivity of the calibration source. The view program is used to recall, process, and display saved data. It utilizes the same display controls as the collect program, allowing the user to view the raw data or temperature data in false color with an adjustable display range. The program can be used to display single frames or continuous video. Background loading of the image files allows for uninterrupted video display. The program also reports the current file and auxiliary data stored with it, and it allows the user to seek a particular frame number rather than scrolling through the video to find that particular frame. IV. APPLICATION AND DISCUSSION
Installation and the first collection of real-time LWIR images of in-service TBC blades took place in February 2013. Data were collected over a 3 h run period for operating conditions from idle to take-off. Thousands of calibrated LWIR thermal images of TBC blades were obtained as a function of turbine rotational speed at both steady state conditions and during engine transients, with each image assignable to a blade number on the turbine wheel. Real-time images identified blades that had lost TBC from areas of their leading edge. An example of data near the trailing edge of a TBC blade is shown in Figure 5. For this area of the blade the entire view is well within the range of take-off angles about normal where LWIR emissivity is high and stable for the TBC. It is noted that complex blade curvature, e.g., around the leading edge, can present LWIR image areas with view angles that approach grazing where TBC emissivity decreases (reflectivity increases).18 The thermal image on the left is an average of 80 images of the targeted blade; a row of cooling holes is prominent. The color scale of ±80 ◦ F is the relative change from the average of the image. The isotherm contours on the right expound a hotter surface in front of cooling holes and in the upper right of the image, with cooler regions at the lower right and close to the cooling holes. The range of false thermal colors demonstrates the ability of the system to monitor TBC response in the combustion environment and to monitor the effect of internal passages and film cooling on the surface temperatures. The measurement technique described in this paper provides an ability to monitor Bill of Material (BOM) blades in their actual operating environment in order to measure engine transient response and to evaluate different cooling designs, manufacturing methods, etc. Data from this three-hour run period will be the subject of companion publications which will focus on using the data for understanding engine operation. As detailed in earlier publications,12, 13 the combustion products at high temperatures and pressures in the fieldof-view influence the thermal intensity detected by the FPA. The approximately 1-μm wide LWIR band pass of the system was chosen not only to exploit the near-unity emissivity of TBC but to also minimize radiance contributions from the hot gases. The HITEMP gas spectra19 for H2 O and CO2 have previously been used to calculate the error contribution to an
124902-6
Markham et al.
Rev. Sci. Instrum. 85, 124902 (2014)
FIG. 5. LWIR data from an in-service TBC blade rotating at typical engine speeds. Data are presented as difference from average in ◦ F. At left a row of cooling holes is apparent in the circular thermal color image. At right an isothermal contour plot of the same data is presented where the white lines are the average temperature contour, the blue lines are lower in temp, and the red and orange lines are higher.
LWIR radiation thermometer.12, 14 The calculation considers the emissivity of the surface of interest, the attenuation of surface radiance by the hot gases, and the radiance contribution by the hot gases. Gas temperature, pressure, concentration, and path length are input parameters. Results of the calculation were presented for operating parameters appropriate to an industrial turbine (e.g., inlet pressure of 15 bars) for path lengths from 5 to 15 cm (2 to 6 in.)12 and 100 cm (39.4 in.).14 Similar calculations were performed for the present work, with estimations of the aircraft engine’s proprietary operating parameters as input. The results showed that for the measurement geometry describe above, the cooling holes and surface temperature patterns would be readily apparent in the LWIR images and therefore these features would not be obscured by interferences in the intervening combustion flow. After the fact, it is notable that the intensity variation in pixel location for the numerous images that were averaged to generate Figure 5 show that the information in the figure is not due to random effect, but is a statistically significant resolution of TBC surface temperature. Associated research and development is in progress to non-intrusively measure combustion gas temperatures at the turbine inlet with an infrared-based spectroscopic technique.20 For future work, the in situ spectral information from the turbine inlet gas temperature determination will be investigated as a means for real-time accounting of the combustion gas radiance contribution in the bandpass of the LWIR imaging radiation thermometer. The LWIR system described in this article has applications beyond the one just described. The typical thermal resolution of modern LWIR cameras is in the tens of milliKelvin. For instance, the thermal resolution of the camera used here is less than 20 mK. Since peak emissions of room temperature objects fall in the LWIR band, the LWIR imaging radiation thermometer is also suitable for monitoring blades coated with ceramic (or instead, coated with a low cost, high emissivity paint) in low temperature turbine rigs used for studying aerodynamics and heat transfer. Many turbine rigs are set-up
so that blade cooling gas comes in below room temperature, and turbine inlet gas comes in above so that the blade is about room temperature.21 V. CONCLUSION
A long wavelength infrared imaging radiation thermometer was developed and implemented to monitor the surface temperature of in-service TBC turbine blades. Its first application on an actual engine test delivered calibrated thermal images as a function of turbine rotational speed at both steady state conditions and during engine transients. Temporal and spatial resolutions achieved provided cooling-hole locations in the images at aero propulsion engine rotational speeds. The instrument provides capability to monitor TBC blades in their actual operating environment in order to measure engine transient response and to evaluate different cooling designs, manufacturing methods, processing reproducibility, and ultimately to gain confidence in TBC performance at maximum energy efficiency of engines. ACKNOWLEDGMENTS
Support to Advanced Fuel Research, Inc. for technology development includes Air Force SBIR Contract Nos. FA9101-09-C-0019 and FA9101-12-C-0025. For numerous beneficial technical interactions, gratitude goes to: Robert Howard and Bradley Winkleman of the Arnold Engineering Development Complex; Ruth Sikorski of the Air Force Research Laboratory; Bryan Engi of Infotech Enterprises America, Inc.; and Christopher Lehane, Andrew Consiglio, Keith Nachilly, Bruce Hockaday, Darren Wind, Allison Nicklous, Doug Thomesen, Martin Shenkle and Joel Wagner of Pratt & Whitney. 1 A.
U. Munawar, U. Shulz, and G. Cerri, “Microstructural evolution of GdZ and DySZ based EB-PVD TBC systems after thermal cycling at high temperature,” J. Eng. Gas Turbines Power 135(10), 102101 (2013).
124902-7 2 M.
Markham et al.
van Roode and A. K. Bhattacharya, “Durability of oxide/oxide ceramic matrix composites in gas turbine combustors,”J. Eng. Gas Turbines Power 135(5), 051301 (2013). 3 D. R. Clarke, M. Oechsner, and N. P. Padture, “Thermal barrier coatings for more efficient gas-turbine engines,” MRS Bull. 37, 891–897 (2012). 4 R. Vanzetti and A. C. Traub, in Theory and Practice of Radiation Thermometry, edited by D. P. DeWitt and G. P. Nutter (Wiley, New York, 1988), Chap. 11. 5 C. Kerr and P. Ivey, “An overview of the measurement errors associated with gas turbine aeroengine pyrometer systems,” Meas. Sci. Technol. 13, 873–881 (2002). 6 W. H. Atkinson and M. E. Cyr, “Evaluation of sensors for temperature measurement of ceramic materials,” in Proceedings of the 3rd Annual HITEMP Review (NASA Conference Publication 10051, 1990). 7 J. R. Markham and K. Kinsella, “Thermal radiative properties and temperature measurement from turbine coatings,” Int. J. Thermophys. 19(2), 537–545 (1998). 8 H. Latvakoski, J. Markham, M. Borden, T. Hawkins, and M. Cybulski, “Measurement of advanced ceramic coated superalloys with a long wavelength pyrometer,” AIAA Paper 2000-2212, 2000. 9 F. E. Pfefferkorn, F. P. Incropera, and Y. C. Shin, “Surface temperature measurement of semi-transparent ceramics by long-wavelength pyrometry,” J. Heat Transfer 125, 48–56 (2003). 10 J. I. Eldridge and C. M. Spuckler, “Determination of scattering and absorption coefficients for plasma-sprayed yttria-stabilized zirconia thermal barrier coatings,” J. Am. Ceram. Soc. 91(5), 1603–1611 (2008). 11 W. H. Atkinson, M. A. Cyr, and R. R. Strange, “Development of sensors for ceramic components in advanced propulsion systems. Phase II: Temperature sensor systems evaluation,” Report No. NASA-CR-195283, 1994.
Rev. Sci. Instrum. 85, 124902 (2014) 12 J.
Markham, H. Latvakowski, D. Marran,J. Neira, P. Kenny, and P. Best, “Challenges of radiation thermometry for advanced gas turbine engines,” in Proceedings of the 46th International Instrumentation Symposium, Bellevue, WA (ISA, 2000), Vol. 397, pp. 299–313. 13 J. R. Markham, H. M. Latvakoski, S. L. F. Frank, and M. Ludtke, “Simultaneous short and long wavelength pyrometer measurements in a heavyduty gas turbine,”J. Eng. Gas Turbines Power (Trans. ASME) 124, 528– 533 (2002). 14 Pyrometry of Combustion Turbine Blades with Thermal Barrier Coating (EPRI, Palo Alto, CA, and KEMA Nederland B.V., Arnhem, Netherlands, 2002) (Publicly available effective December 6, 2006). 15 H.-G. Brummel, D. LeMieux, M. Voigt, and P. Zombo, “On-line turbine blade monitoring,” Power 149(7), 11 (2005). 16 C. Kerr and P. Ivey, “A review of purge air designs for aeroengine-based optical pyrometers,” J. Turbomach. 124, 227–234 (2002). 17 R. E. Bedford, in Theory and Practice of Radiation Thermometry, edited by D. P. DeWitt and G. P. Nutter (Wiley, New York, 1988), Chap. 12. 18 R. J. Chandos and R. E. Chandos, “Radiometric properties of isothermal, diffuse wall cavity sources,” Appl. Opt. 13(9), 2142–2152 (1974). 19 L. S. Rothman, I. E. Gordon, A. Barbe et al., “The HITRAN 2008 molecular spectroscopic database,” J. Quant. Spectrosc. Radiat. Transfer 110, 533– 572 (2009). 20 J. Scire, J. Cosgrove, and J. Markham, “Turbine inlet gas temperature measurement system,” in Proceedings of the 57th International Instrumentation Symposium, St. Louis, MO (ISA, 2011), Vol. 488, pp. 295–307. 21 R. M. Mathison, C. W. Haldeman, and M. G. Dunn, “Aerodynamics and heat transfer for a cooled one and one-half stage high-pressure turbine – Part I: Vane inlet temperature profile generation and migration,” in Proceedings of the ASME Turbo Expo, Glasgow, Scotland, ASME GT201022716 (ASME, 2010), pp. 299–312.
Review of Scientific Instruments is copyrighted by the American Institute of Physics (AIP). Redistribution of journal material is subject to the AIP online journal license and/or AIP copyright. For more information, see http://ojps.aip.org/rsio/rsicr.jsp
