Presenting a new structure based on reasoning in uncertainty conditions for determining Remaining Useful Life of rotary equipment using vibration data

Determining the Remaining Useful Life (RUL) for the devices is one of the important accomplishments of the Prognostic and Health Management (PHM) systems. Knowing the Remaining Useful Life, even if it may contain some acceptable prognostication errors, can be of great help when deciding the best maintenance strategy. In this paper anew structure is presented based on the combination of the device decay and failure status data which is obtained from classification of new feature extracted from vibration data. In order to prognosticate, estimation and determination of Confidence Interval for Remaining Useful Life is done using Auto-Regression Marcov Regime Switching. Forevaluation, the proposed model is implemented on bearing vibration data, and the results demonstrates that the proposed method has a greater stability compared to the similar methods.

Keywords:

Remaining Useful Life (RUL) , Prognostics & Health Management (PH1/I) , Autoregressive Markov Regime Switching (AMRS) , Wavelet Decomposition , Evidence Theory , Fuzzy clustering , Fuzzy C-Means , Kurtosis-Entropy , feature , Degradation