Abstract

The efficiency and suitability of on-line vibration measurements as a diagnostic signal for condition monitoring and diagnostics of gear systems, are investigated. To this end, representation of a gear system as a single-degree-of-freedom dynamic system is considered. Mathematical models of faults or malfunctions of the gear system for use in the above dynamic system as its system and excitation parameters are synthesized. Further, the sensitivities of vibration signal parameters including time-domain, frequency domain and statistical parameters, to various different faults or malfunctions are identified and explained. Based on this information, the rational basis of using on-line vibration signal parameters as the diagnostic indices of a gear system is established. Experimental results are obtained from real-life gear testers that have different types of faults or malfunctions. On-line vibration signals obtained from these gears tests are analyzed for time-domain, frequency domain, quefrency domain, statistical and time series model parameters. The efficiency and performance levels of each of these indices in diagnosing a particular type of fault are demonstrated. Reliable signal parameters corresponding to various types of faults or malfunctions in the gear system, that serve as best indications of damage development, are identified from the test results. Theoretical predictions and experimental observations of various types of gear system failures are compared in order to establish the confidence levels about the on-line vibration signal parameters.

This content is only available via PDF.
You do not currently have access to this content.