Health indicator metrics applicable to inductive wear debris sensors

Abstract

The Inductive Wear Debris Sensor is a relatively new invention that is increasingly being used for the detection of incipient machinery damage or failures by sensing metallic debris in lubrication systems. This type of sensor is typically used in-line and has a superior particle size detection range compared to traditional techniques such as the ubiquitous spectrometric oil analysis. There is, however, very little in the literature regarding the application and interpretation of data arising from this type of sensor. Unlike other condition monitoring sensors, no data will be generated by an Inductive Wear Debris Sensor in an ideal system; however, in real applications it is necessary to discriminate between occasional particles unrelated to a failure and incipient failure particles. Inductive Wear Debris Sensor data could be misinterpreted if a simple cumulative count limit was applied to the data. A short-term rate of particle generation is sometimes used as an alternative; however, it too can be misleading with short succession particles producing high instantaneous rates possibly causing false alarms. The purpose of this work was to develop a robust metric (or group of metrics) that when applied to Inductive Wear Debris Sensor data would reliably identify a failure event and exclude non-failure related particles. The Health Indicator described herein consists of three subordinate Condition Indices that collectively are shown to reliably detect the onset of rolling contact fatigue. The metrics have been applied to bearing test rig data (seeded fault) and data obtained from a non-seeded fault test of a complex helicopter gearbox.

Helicopter Planet Gear Rim Crack Diagnosis and Trending Using Cepstrum Editing Enhanced with Deconvolution

“…This is achieved by changing the values of the coefficients in the MED filter (g). The sequence of the process consists of the following:  Assume initial values of the filter, g = f (1) = [g0, . .…”

“…However, gears are often subjected to high torques and cyclic loads, which eventually lead to component failures due to fatigue. Gearbox health monitoring is conducted using a variety of methods including wear debris monitoring [ 1 , 2 ], thermography [ 3 ] and vibration analysis [ 4 , 5 , 6 ]. It is common to combine more than one method, e.g., wear debris and vibration trending, to provide a better indication of health through a hybrid model [ 7 ].…”

Helicopter Planet Gear Rim Crack Diagnosis and Trending Using Cepstrum Editing Enhanced with Deconvolution

“…However, gears are often subjected to high torques and cyclic loads which eventually lead to component failures due to fatigue. Gearbox health monitoring is conducted using a variety of methods including wear debris monitoring [1,2], thermography [3], and vibration analysis [4][5][6]. It is common to combine more than one method, e.g.…”

“…• Calculate vector f (1) : New filter coefficient defined as with the correlation between y 3 (t) and the data d(t). f (1) is found by solving the system Rg = f ; where R is a N×N autocorrelation matrix defined with the N first delay of the autocorrelation function of the data d(t).…”

Multichannel Inductive Sensor Based on Phase Division Multiplexing for Wear Debris Detection

“…Inductance detection could differentiate between ferrous and nonferrous debris [11,12], and it is insensitive to air bubbles and water droplets [13,14]. Some inductive sensors, such as MetalSCAN (an on-line oil debris sensor made by GasTOPs Ltd. ), have been utilized in the wind turbine gearbox and aviation machinery [15,16,17,18]. Similar to other sensors, the sensitivity is an important feature for the inductive sensor.…”