Diagnostic information on gearbox condition for mechatronic systems

Abstract

To maintain mechatronic systems, operational current diagnostic information on gearing condition is needed. Diagnostic information on gearbox condition can be gained from vibration signals. To extract suitable information from the signal we need to know the relation between factors influencing the diagnostic signal and the gearbox condition. There are several factors that have influence on diagnostic signals and these may be divided into four groups: design (DF), production technology (PTF) operational (OF) and change of condition (CCF). To gain information from the vibration signal, many experiments have to be carried out. Because many factors influence the vibration signal, the experiments on real objects are very expensive. To reduce the cost, mathematical modelling and computer simulation may be used. Papers devoted to mathematical modelling and computer simulation are concentrated on one-stage gearboxes. In this research work, a system that consists of a driving motor, coupling, double-stage gearbox and driven machine is considered. Using the model, it is possible to investigate the above-mentioned factors that influence the vibration signal. Faults occur during a gearbox operation. Gearing faults may be divided into single faults (a crack in a tooth or its breakage) and distributed faults caused by gearing wear (pitting, scuffing, abrasion and erosion). The methods of fault detection using signal analysis is given in this paper. The paper shows how mathematical modelling and computer simulations help signal interpretation, taken from real measurements. Frequency, time and time-frequency analysis are used for signal interpretation. As a result, we obtain spectra, cepstra and time-frequency spectrograms. The paper refers to the condition monitoring method, which used simple wide-band signal estimators. The method may be easily adopted for use in monitoring the online gearbox condition in mechatronics systems, whereas an advanced signal interpretation may be used for offline condition monitoring.