Date of Completion

8-24-2011

Embargo Period

8-19-2011

Advisors

Dr. Michael Renfro; Dr. Mark Aindow

Field of Study

Mechanical Engineering

Degree

Master of Science

Open Access

Campus Access

Abstract

Thermal barrier coating (TBC) systems are an integral part of today’s engine parts. The ceramic coatings offer protection from thermal loads imposed on components such as turbine blades. Coated engine parts are currently taken out of operation based on group data when statistical minimums are met. In order to confirm the TBC’s state, the parts must be sectioned and are therefore no longer useful. Due to their importance and inevitable failure, the inspection of TBCs and the eventual use of remaining lifetime models on individual components would save potentially functional parts from being removed from use before necessary. Using thermally sprayed sample coupons and simulated thermal loading in elevator furnaces, non-destructive evaluation (NDE) methods have successfully been under develop at UCONN for several years. In this study, coated blades that have undergone engine operation were used to further the lifetime prediction model, presenting new challenges. The NDE method used involves performing photoluminesent piezospectroscopy (PLPS) to collect R1 and R2 peaks from Cr+3 within the alumina of an underlying thermally grown oxidation (TGO) layer of the TBC system. Strains in this Cr+3 doped alumina shift the peaks, which can then be converted into a stress, and consequently, applied to a remaining lifetime model. The key challenge is detecting the signal though the over lying coating as one must look directly through the TBC ceramic in order to detect signals from the TGO. Engine-run blades have significant surface contamination, in which one of the primary constituents is alumina. The contamination, which collects on top of the TBC surface, luminesses an unstressed set of peaks which mask the desired signal emanating from the TGO. The results presented in this thesis are an integral part of the development and verification of a novel approach to clean the surface contaminate through laser ablation. Mark Majewski was primarily responsible for developing the ablation hardware and procedures. In this thesis, microstructure and geometry characterization of the ablated surfaces was done and in addition and most importantly, the suitability of the ablated areas for PLPS stress measurements were determined by making such measurements. A contribution of this thesis was developing a suitable way to censor the data to bring success to the overall effort. After ablation, blades were analyzed for chemical composition both optically and using an electron microscope to ensure minimal damage to the TBC itself. PLPS was performed to extract the underlying TGO spectra. Results showed a contrast with the spectra from uncleaned blades and cleaned blades. While redeposition and columnar infiltration caused contaminate spectra to sometimes be present after cleaning, the contrast in both signal strength and resultant stress values offered a way to censor the “false” signals out of data sets taken over ablated regions. With the engine run blades used, stress values followed the general trend of previously developed lifetime models. While the results are promising, more study is needed to further confirm the results. Specifically, engine parts with high enough hours such that decisive shifts in stress are expected must be analyzed to confirm preliminary findings. Finally, an examination of some unexpected peak shifts found after cleaning should be carried out to understand the origin of this phenomena. It should be noted that these shifts do not disallow the use of the developed procedures for NDE of engine parts.

Major Advisor

Dr. Eric Jordan

Share

COinS