4 ms·
Absolute Vibration Level and Damage Factors for Machinery Health Monitoring
- camtarn 8y ago"The raw data from a vibration transducer mounted on a test structure is obtained in time domain. The vibration signal in time domain is useful to the extent of finding out the overall vibration level. The overall vibration level may not exactly indicate the impending defect that is growing in the system. The frequency that is responsible for a particular defect is to be identified rather than the overall vibratory level. For this the vibratory signal in time domain is to be converted to frequency domain using Fast Fourier Transforms and the vibration analyzers (FFT Analyzers) do this job [2]. Special analysis techniques like envelope spectrum analysis, cepstrum analysis, spike energy method, shock pulse method, waterfall diagram etc., are used when the spectrum analysis does not give much information about the defect, or when the technique suits the system to be monitored than spectrum analysis." Might be interesting for anybody working with e.g. server or equipment logs. If you have such a volume of logs that identifying recurring events is impossible and counting absolute numbers of errors is not useful, could throwing the log into an FFT and examining the frequency spectrum of errors help?