Scenario Reduction Method based on Output Performance for Condition-based Maintenance Optimization
Condition-based maintenance (CBM) has been widely applied to maintenance policy. Due to the stochastic forecasting degradation, scenario reduction method has been developed to improve the efficiency of CBM. However, most existing scenario reduction methods focus mainly on the input performance of the forecasting degradation without considering the significant output performance characteristic based on the model.In order to warrant the CBM optimization precision while reducing stochastic degradation scenarios efficiently, a new scenario reduction method is formulated that the scenarios with same objective function values can be reduced to one representative scenario. As a result, the reduced scenarios by the proposed method can maintain the probability distributions of objective values, while keeping optimal thresholds close to that of initial scenarios. Finally, the method is applied to select the representative degradation scenarios for CBM optimization model by utilizing vibration-based degradation signals from a rotating machinery application. Compared to the traditional scenario reduction method, the proposed method further improves accuracy and reduction efficiency of CBM optimization.