Improved Failure Detection Techniques Based on Spectrometric Oil Analysis Data

This paper discusses a set of general analysis techniques which have been developed to improve the capability of spectrometric oil analysis to detect failures of oil-wetted components in turbojet engines. Present knowledge of the failure processes involved is used to estimate the properties of the colloidal suspension of the wear particles in the oil and to show how these properties affect the spectrometric concentration readings obtained. Many of the anomalies occasionally observed in the spectrometric data are explained in terms of the effects of the random distribution of the wear debris throughout the oil and the size distribution of the individual wear particles. Combining these results, it is shown that the rate of wear metal liberation is a better indication of the mechanical condition of a component than is the absolute concentration of wear metal in the oil. The method is made operationally effective through the use of a general purpose digital computer to perform time-consuming tasks such as calculation of wear metal liberation rates, correction for oil consumption effects, and quantitative correlation analysis for diagnostic purposes.