An approach to detect the damage quantitatively in rotary machine was presented based on physical model and gray theory.For the purpose of detecting sun-gear tooth-chipped damage in 2K-H planetary gear sets

physics-based models for damage system with multi-stage severity were established

and then simulation data of damage system was obtained.Then

feature vector of simulation data extracted and computed was used as standard damage mode.Relational analysis in grey theory was carried out between standard damage mode and the feature vector of signal

and also checked.The quantitative detection result for damage was achieved from relational grade computed before.Tooth-chipped seeded tests were carried out and used to validate the approach.It is proved that the approach presented above is effective to detect the sun-gear tooth-chipped damage

and its potential profit in other rotary mechanical equipments will be researched in the future.