Protocol for selecting exemplary silicate deposit compositions for evaluating thermal and environmental barrier coatings

Abstract

A protocol for selecting representative silicate compositions for comparative testing of gas turbine coating materials is presented. It begins with a curated dataset of compositions of engine deposits and naturally occurring siliceous debris including volcanic ashes, sands, and dusts. The compositions are first reduced to the five major oxides—those of Ca, Mg, Fe, Al, and Si—and then distilled further using principal component analysis and k‐means clustering. The process ultimately yields four classes of possible deposits with common chemical characteristics. Each class is represented by a composition centroid and a range in Ca:Si ratios. Key thermophysical properties of the possible deposits are calculated and related to the glass network connectivity, characterized by the Si:O ratio. Finally, deposits from each of these classes are compared in terms of their reactions with prototypical thermal and environmental barrier oxides, with due consideration of the effects of composition variations within each deposit class. The protocol is, in principle, adaptable to datasets compiled by OEMs and researchers in gas turbine coatings.

Document Details

Document Type
Pub Defense Publication
Publication Date
Mar 05, 2022
Source ID
10.1111/jace.18413

Entities

People

  • Andrew R. Ericks
  • Carlos G. Levi
  • David L. Poerschke
  • Frank W. Zok

Organizations

  • Office of Naval Research
  • University of Minnesota

Tags

Readers

  • Neural Network Machine Learning.
  • Petroleum Engineering
  • Thin Film Deposition Science.