Operational joint identification using transfer path analysis and frequency-based substructuring

Large machines like engines or turbines are subject to operational forces that may be hard to model or measure directly. These machines are often connected to the environment via joints that can degrade over time, necessitating condition monitoring to detect and quantify changes in their dynamics. This paper presents a novel approach combining transfer path analysis (TPA) and frequency-based substructuring (FBS) to identify joint dynamics during operation. Unlike conventional TPA applications, we use the characterization of operational forces together with measured operational responses to estimate frequency response functions (FRFs) of the system with dynamic substructures. The estimated FRFs are then used in an FBS procedure to isolate the joint dynamics. The methodology is demonstrated on a magnetic bearing test rig, where both synthetic operational forces and run-up tests are conducted. Results show good agreement for synthetic force cases, while highlighting challenges with correlated excitation in run-up scenarios. The proposed method enables condition monitoring of joints in operational conditions.