Methodology of evaluation of in-service loads applied to the output shafts of automatic transmissions.

摘要

This work presents a novel methodology for evaluation of in service loads applied to the output shafts of automatic transmissions. It also presents a novel methodology of data reduction for shaft load signals as an alternative to the cycle counting methods.Current durability testing of automatic transmission output shafts uses 50 000 stall torque cycles from zero to wide open throttle. In the majority of cases, these requirements lead to an over design that can result in an unnecessarily bulky transmission system. As a solution to this problem a novel methodology for evaluationof loads applied to the output shafts of automatic transmissions was developed. The methodology is based on real world loading conditions and therefore leads to a more realistic estimation of the fatigue life of shafts. The methodology can beused as a tool for shaft optimisation in different drive conditions. Using the developed methodology the effects of different road conditions on the fatigue life of a transmission output shaft were compared. Four routes having differing drivingconditions were investigated and of those routes, the route with most stop-start events resulted in the greatest reduction in fatigue life.A novel methodology of data reduction for shaft load signals was also developed. The methodology is based on knowledge of the bandwidth and dynamicrange of the expected in-service load signal. This novel methodology allows significant reduction of the volume of data to be acquired. It preserves the time sequence of peaks and valleys of the signal, which is vital in the case of fatigue analysis. This is in contrast to current methods based on cycle counting. Cycle counting methods achieve high data reduction but do not preserve the time sequence of the signal. The developed novel methodology has been validated on the newly developed data acquisition system capable of real time data acquisition and compression of shaft torque signal. The performed tests show that the proposed one-channel low cost system equipped with 1 GB compact flash card can store well over 10 000 hrs of load history.