Multiple incipient faults prognosis for CCBII braking system

Due to various operation modes, complex working environment, and various components with strong coupling in CCBII braking system, multiple incipient faults may happen simultaneously. It is difficult to carry out the multiple incipient faults prognosis for the CCBII braking system using the traditional faults prognosis techniques. In this article, a model-based fault prognosis approach is used for CCBII braking system with multiple incipient faults. Firstly, the models of the working modes are established, the latest concept of Augmented Global Analytical Redundancy Relations (AGARRs) is used to consider parametric and nonparametric incipient fault prediction. Before the faults are detected, mode tracking is realized by comparing Mode-change Signature Matrix (MCSM) with the coherence vector. Once the mode tracking fails, the fault hypothesis set, including suspected faults and suspected mode change, is established correspondingly. The degradation behavior of incipient fault is unknown in advanced and the unknown degradation characteristics of each incipient fault can be predicted with some matching dynamic models. Furthermore, the degradation models and incipient faults are identified using a mixed differential evolution (MDE) algorithm. The real-time online faults prognosis of CCBII braking system is realized. Finally, simulation tests validate the efficiency of the proposed scheme.