Approach to rough set stratified fault diagnosis of turbogenerator vibration

As the structure and vibration of turbo-generators are complicated, their faults are generally multi-level and random, and meanwhile, it also has other characteristics, such as fault information incompleteness, etc. Starting from general problems, this paper constructs a rough set stratified fault diagnosis model with rough set information measurement from data table decomposition. Rule sets obtained from this model have high supportiveness and practicality, and its stratified diagnosis approach is similar to human reasoning approaches, and so is easy to understand. Comparison is performed with real cases between the rough set stratified fault diagnosis model and common rough set fault diagnosis models, and proves the effectiveness of this approach.