The inter-shaft bearing could simplify the supporting structure of a dual-rotor system, so it is widely used in rotating machinery like gas turbine engine. When monitoring the vibration of an inter-shaft bearing, sensors usually could not be installed inside the engine because of the compact structure, then will be installed on the casing. The vibration of the inter-shaft bearing will be transmitted though both the two rotors to the supporting bearings then to the casing, and both the inner and outer ring rotate, making the condition monitoring much more difficult than normal bearings. In this paper, a dual-rotor system dynamic model with the rotor-bearing-casing structure is built to simulate the vibration of an inter-shaft bearing. The rotors and the casing are modeled using finite element method and the bearings are modeled using a nonlinear multi-body dynamic model. The normal state of the inter-shaft bearing is simulated, then the bearing fault states with localized defects on the inner race, the outer race and one of the rolling elements are simulated. Signal processing methods are utilized to extract bearing fault features from the casing vibration acceleration signal. Besides the fixed rotating speed condition, rotors under variable speeds are also simulated, then the vibration signal and fault feature become more complex and the signal analysis becomes a bigger challenge.
