Hunting Tooth Gear Condition Indicator : For Epicycle Gear Fault Detection

This paper looks at the use of the hunting tooth (HT) time synchronous average (TSA) to improve fault detection on gears. HT analysis is based on the period at which two damaged teeth in a gear mesh. The HT frequency tends to be low, requiring a long acquisition to calculate the TSA. Because of the long acquisition, the HT TSA may improve the signal-to-noise of some gear fault pairs, such as seen in the epicyclic gearbox planet/ring gear mesh. This study develops two Condition Indicators (CI) based on the HT TSA from two different data sets. The first data set has been used to quantify the performance of gear algorithms using statistical separability. The second data set is from an OH-58 helicopter gearbox that was seeded with a damaged planet gear. Planet gears are typically difficult to analyze as if only one planet has damage, and the fault may be masked by the other good planets, making early fault detection difficult. The proposed CI showed good fault detection and trends for both data sets and was superior for the epicyclic gearbox. The performance of the HT TSA and new CIs is quantified relative to other gear CIs and validated using gear fractography crack length.