Self-adaptive health condition prediction considering dynamic transfer of degradation mode

This paper investigates the data characteristics of engineering equipment under different health conditions in order to realize more accurate health state prediction during its degradation process. An approach to classify the degradation modes is introduced to overcome the shortcomings of the single-prediction-method existing in current applications. Based on an analysis on the relationship between different prediction models and degradation modes, an association table "Mode-Model" is established, thereby the self-adaptive health condition prediction and residual life estimation can be achieved by optimal selection of prediction models, which considers the dynamic transfer of degradation mode. Finally, the efficiency of the proposed method is verified by a rolling bearing experiment.