Development of a Comprehensive Method for Time-Frequency Analysis of Rotating Machinery
Abstract
This research was concerned with stochastic and statistical characterizations of quasi-periodic random processes, such as those associated with rotating machinery. The value of such characterizations includes enhanced abilities in condition monitoring for purposes of safety and performance. Major achievements of this research effort include: (1) characterization of the influence of period uncertainty on estimation of a periodic time/frequency spectrum associated with a nominally wide sense cyclostationary (wsc) process, (2) large sample distribution descriptions for the AR(p) and MV(p) spectral estimators for processes with mixed spectrum, (3) a time-to-angle transformation method to better accommodate period variability, combined with an improved method for tracking real sinusoids with slowly varying frequency, (4) greater insight into issues related to application of advanced spectral analysis methods for characterizing random processes associated with engines, compressors, and helicopter drivetrains, (5) a Matlab-based virtual signal analyzer which incorporates a number of our results in a user-friendly fashion, and (6) development of research collaborations and workshops which serve to bring signal processing problems associated with rotating machinery to a broader base of researchers in industry, defense and academic institutions.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 27, 1998
- Accession Number
- ADA337835
Entities
People
- Peter J. Sherman
Organizations
- Iowa State University