Parametric modeling methods: theory and a case study

It has been proven that as a machine condition changes, the vibration characteristics will also change [26]. Renwick and Bubson have proven that the vibrations that exist in rotating machinery can be used to indicate many defect modes before catastrophic failure occurs, whether the defect serves as a source for periodic energy or merely modifies the transmission path [20]. Jianfang, Yueming and Qi developed an effective method for decomposing bearing vibration signals and detecting faults using adaptive noise canceling and the kurtosis analysis technique [12]. Peczely used vibration analysis to characterize the condition of gas-engine units and detect different types of failure [17]. Piety, Piety, and Scheibel [18] applied predictive maintenance techniques to centrifugal fans and found that several faults were detectable through the application of spectral frequency analysis of vibration signals. These recognizable faults include imbalance, misalignment, bent or bowed shafts, bearing faults of all types, and structural degradation, to name a few.