GEAR FAULT DETECTION UNDER SIMULTANEOUS VARIABLE LOAD AND VARIABLE SPEED USING A TIME-VARYING AR MODEL AND A TWO-SAMPLE KOLMOGOROV-SMIRNOV GOODNESS-OF-FIT TEST

摘要

Industrial gearbox usually has multiple gear pairs which have multiple different shaft speeds and works under variable load conditions. Condition detection of multiple shaft speed gearing system becomes a complicated topic because multiple dominant gear meshing frequencies are involved in the meshing motion signals, where the conventional spiky feature of initial gear fault conditions is usually invisible and hard to be numerically represented. This paper presents a combined technique, which unites a time-varying autoregressive model and a two sample Kolmogorov-Smirnov goodness-of-fit test, to detect the deterioration of gearing system with simultaneously variable shaft speeds and variable loads. A two-speed reduction gearbox was used to verify the proposed technique. Residuals of time-varying autoregressive model were analyzed. The time-varying autoregressive model residuals of both healthy gear condition and fault gear condition were compared with the original healthy time-synchronous average signals of the investigated gearbox by means of a two sample Kolmogorov-Smirnov goodness-of-fit test to detect the deviation from healthy gear condition. The results were also compared with the traditional kurtosis statistic and showed significant advantages in highlighting the presence of incipient gear fault in all different speed shafts involved in the meshing motion under variable load conditions.