NONDESTRUCTIVE EVALUATION OF CERAMIC MATRIX COMPOSITE COMBUSTOR COMPONENTS JG.Sun,M.JVerrilli, R. Stephan, T R. Barmett'and G. Ojare Argonne National Laboratory, Argonne, IL 60439 Glenn Resea Cleveland, OH 44135 birmingham, AL 35211 Technologies Research Center, East Hartford, CT 06108 ABSTRACT. Combustor liners fabricated from a SiC/SiC composite were nondestructively terrogatcd before and after combustion rig testing by X-ray, ultrasonic and thermographic echniqucs. In addition, mcchanical test results were obtained from witness coupons, nting the manufactured liners, and from coupons machined from the components after combustion exposur Thermography indications were found to correlate with reduced material properties obtained after rig testing. The thermography indications in the SiC/SiC liners were delaminations and damaged fiber tows,as determined through microstructural examinations. INTRODUCTION The pursuit of lower emissions and higher performance from gas turbine engines key engine components. One key engine component is the combustor, where liners fabricated from a ceramic matrix composite, silicon carbide fibers in a silicon carbide matrix(SiC/SiC), are under evaluation. SiC/SiC composite technology has progressed from fabrication of simple coupons to subelements, and now to actual components [1-4] This last area posed a significant challenge to nondestructive evaluation(NDE)techniques due to the complex geometries involved To evaluate SiC/SiC components in a combustion environment, the Rich-burn, Quick-quench, Lean-burn(RQL) sector rig was designed and fabricated. The purpose of he sector rig testing was to demonstrate the structural durability of the Sic/Sic liners in a combustion environment where stresses, temperatures, and pressures would as accurately as possible reflect the operating conditions found in a turbine engine X-ray, ultrasonic and thermographic techniques were used to inspect the as- manufactured SiC/SiC components and the liners after combustion testing. The purpose of this study was to characterize the post-exposure condition of one SiC/SiC combustor liner part through NDE evaluation via thermography and mechanical property measurement CP657, Review of Quantitative Nondestructive Evaluation Vol 22, ed. by D. O. Thompson and D. E. Chimenti @2003 American Institute of Physics 0-7354-0117-9/03/$20.00
EXPERIMENTAL Material and Component Combustor liners were manufactured by Honeywell Advanced Composites from a Sic/SiC composite developed under the NASA Enabling Propulsion Materials program A chemical vapor infiltrated, slurry-cast, and melt-infiltrated SiC matrix was reinforced with Sylramic SiC fibers. The fiber tows were woven into 5-harness satin weave cloth More details on the SiC/SiC material can be found in reference 5 The sector rig liner set consisted of 6 different component geometries and a total of 28 liners(Fig 1). For this study, one of the six components, the lean zone inner diameter liner (LZiD)was examined. The LZiDs are curved plates, each covering about a 30% arc of the total 60 sector, as shown in Fig. 1. A twelve ply thick leading edge reinforces the attachment hole region. The rest of the LZID is six plies thick. LZIDs containing 34 and 40% fiber volume were tested Flexure coupons were machined from the as-manufactured and exposed sector parts to obtain mechanical properties. The as-manufactured parts were supplied with extra material that was machined into test coupons. The exposed parts were machined to obtain est coupons of the same geometry as was used for the witness testing. Three point flexure testing was conducted at 815C. More details on the mechanical test procedures can be found in reference 6 Combustion Testing The Preheated Combustor and Materials Test Facility(PCMTF) at NASA Glenn Research Center was used to conduct the RQL sector rig testing. Air temperature(T3), pressure(P3), and flow rate(w3)were varied to approximate the anticipated service cycle of the High Speed Civil Transport(HSCT)engine combustor [7]. All aspects of an HSCT flight operation, such as take-off, climb, cruise, and decent, are inclue ded in the 230 minute The combustion testing was periodically interrupted in order to remove some of the SiC/SiC liners for post-exposure analysis. After completion of 260 hours of operation, the rig was disassembled and all Sic/Sic combustor liners were removed NDE Techniques As-manufactured and combustion-exposed SiC/SiC components were inspected using x-ray film radiography, ultrasonic, and thermography. Radiography is sensitive olumetric features such as large voids and gross porosity or thickness variations, but cannot detect planar (laminar) features such as delaminations in ceramic matrix composites. Ultrasonic techniques have high sensitivity for detection of delaminations, but quantitative interpretation of the data is difficult as they are affected by many experimental setup parameters. Thermography techniques are also very sensitive to delaminations, and when used in a through-transmission setup to determine thermal diffusivity, which is material property, the data can be quantitatively analyzed to relate with the detailed tructures of defects [8]. In this study, only the thermography results will be discussed 1012
FIGURE 1- Schematic of the RQL sector rig and images of all the SiC/Sic parts a transient through-transmission thermography method was used on the tested liners. Two high-energy flash lamps generate a heat pulse that travels through the thickness of the parts(Fig. 2). The raw data consists of multiple layered images acquiredat 120 Hz. A total of 40 frames were typically suitable to capture the temperature rise, The analysis technique was based on that given by Parker [9]. The front surface temperature rise is given by r(D2=Q|1+2∑(ye where L is the liner thickness, a is the thermal diffusivity, t is time elapsed, p is the material density, Q is the amount of heat absorbed by the liner from the flash lamps, and C is the specific heat of the material. Using the time to reach half of the maximum detected temperature((1/), the thermal diffusivity at each pixel can be calculated using At regions of planar defects, such as delaminations, the material surface temperature is lower than that of the undamaged composite and it will take a longer tin to reach the equilibrium part temperature (note that the part will eventually reach an equilibrium constant temperature if sufficient time is given). Therefore, from Eq. 2, the regions with damaged material will have lower thermal diffusivities The thermal diffusivity calculated from Eq. 2 is proportional to the thickness squared. Because it is impossible to provide thickness values for all pixels of an imaged part, it is common to use a representative thickness(typically the averaged thickness) for the calculation of the diffusivity image for the entire part. This method has some shortcomings when thickness variation is large. For example, in the case of a uniform material of constant diffusivity, Eq. 2 dictates that for a pixel with a thickness twice that of a second pixel, the half-rise-time ti at the first pixel will be 4 times longer than that of the second. However, if we choose to use the thickness of the second pixel for calculating diffusivities at all pixels, the calculated diffusivity at the first pixel would appear to be 4 times lower than its real diffusivity. Similarly, the diffusivity calculated from Eq. 2 appears to be higher for a thinner pixel. Therefore, care should be taken when interpreting measured diffusivity images. Note that this equation is exact when a test part has a niform thickness 1013
Lean Zone Bulkhead Heatshield Lean Transition Zone Liners Lean Zone OD Liners Lean Zone Rig Sidewall Lean Zone ID Liners Rich Zone Liner Lean Zone Bulkhead Heatshield Lean Transition Zone Liners Lean Zone OD Liners Lean Zone Rig Sidewall Lean Zone ID Liners Rich Zone Liner
ample R camera Lamps FIGURE 2- Schematic of the through-transmission( thru-thickness)thermography test configuration. RESULTS An example of thermography results obtained for three LZIDs after 115 hours of sector rig testing is shown in Figure 3. Part 216-1 contains 34% fiber volume fraction, and the other two contain 40 % Thickness of the middle sections of the liners was used to calculate the diffusivity values in the image. The apparent diffusivities are lower at the strips with fastener holes where the material is thicker and higher at the left edges where material is thinner. In addition to these diffusivity variations due to thickness change, regions with considerably lower diffusivities can be observed in all liners, especially for 216-1 To investigate the origin of local lower diffusivity values, a LZID was sectioned and polished. Figure 4 shows the microstructure of a cross-section taken from 216-1. The line on the thermography image in Fig 3 shows the location of the ection. Damaged SiC fiber tows and delaminations cracks can be seen. The diffusivity reduction shown in Fig. 3 is related to the damage(i.e, delamination) of the CMc material due to the sector testing. For a single delamination, Sun [8] showed that the diffusivity reduction could be used to determine the(air gap) thickness of the delamination Thermal diffusivity data for another LZId with 34% fiber volume, which was exposed in the combustion environment for 145 hours, is shown in Fig. 5. Thermal diffusivity FIGURE 3- Thermography images of three SiC/Sic liners after 115 hours of sector rig testing. Part 216-1 1014
IR Camera Temp Time Test Sample Lamps IR camera IR Camera Temp Time Test Sample Lamps IR camera
FIGURE 4- Microstructure of LZD 216-l after 115 hours of combustion exposure. The line in Fig 3 shows the location of this section. Arrows indicate delamination cracks and damaged tows are circled. .Tnner tegin Distance(pixels) FIGURE 5- Thermal diffusivity data for LZID 216-2 after 145 hrs of combustion exposure, a) line plots 7 part locations, b)average values for regions of coupons machined for strength measurement. values are plotted in Fig. 5a along 7 horizontal lines passing through several interesting regions in the thermal diffusivity image of 216-2. Although there is thickness variation 1015
Delamination