Abstract
Cracked rotors exhibit nonlinear spring characteristics of a piecewise linear (PWL) type due to the "breathing" mechanism—the periodic opening and closing of cracks during rotation. While extensive research has focused on the dynamic behavior of cracked rotors to develop fault diagnosis systems, most studies assume linear bearing supports. In practical applications, however, rotors often encounter nonlinearities from bearing supports and other mechanical elements. This paper investigates the vibrational behavior of a cracked rotor within a nonlinear system, specifically focusing on combined resonances. Through numerical simulations using a PWL model and theoretical analysis via a Power Series (PS) model, we examine harmonic and 1/2-order subharmonic resonances. Our findings reveal that the dynamic response in nonlinear systems differs fundamentally from linear ones, with the angular position of the rotor's unbalance playing a decisive role in the resulting vibration signatures.
Introduction
In high-speed rotating machinery, such as power plant turbines and industrial motors, a transverse fatigue crack is a critical failure mode. Detecting these cracks before they lead to catastrophic failure is a primary goal of structural health monitoring. Traditionally, engineers identify cracks by monitoring vibrational "signatures"—specific resonance peaks that emerge as the rotor spins.
However, a significant gap exists between theoretical models and industrial reality. Most diagnostic frameworks are built on the assumption that the rotor is supported by linear bearings. In practice, many systems are inherently nonlinear due to the physical properties of the bearings or the geometry of the supports. When the nonlinearity of a "breathing" crack interacts with the nonlinearity of the bearing supports, the resulting physics become significantly more complex, rendering traditional linear diagnostic tools ineffective.
Modeling the "Breathing" Mechanism
To capture the complex dynamics of a cracked rotor, this study utilizes two distinct mathematical approaches:
- Numerical Simulations (Piecewise Linear Model): The "breathing" of a crack—where the crack opens and closes as the shaft rotates—is modeled as a transition between two stiffness states. This PWL model allows for a precise simulation of how the system's stiffness changes instantaneously during a rotation cycle.
- Theoretical Solutions (Power Series Model): To complement the simulations, a Power Series (PS) model is employed to derive analytical solutions. This helps in understanding the underlying mathematical structure of the resonances and validates the numerical findings.
By focusing on harmonic resonance (occurring at the major critical speed) and 1/2-order subharmonic resonance, the research highlights how these specific vibration peaks shift in intensity and frequency under nonlinear conditions.
The Impact of Nonlinear Bearing Supports
The study demonstrates that the vibration of a cracked rotor in a nonlinear system is not merely the sum of its parts. The interaction between the crack’s piecewise linear stiffness and the bearing’s nonlinear supports creates "combined resonances." These behaviors are qualitatively different from those observed in rotors with linear supports.
One of the most significant findings is the sensitivity of the system to the angular position of the unbalance (mass eccentricity). In a linear system, the unbalance might simply increase the amplitude of vibration. In a nonlinear system, however, the physical orientation of the "heavy spot" relative to the crack location can completely alter the resonance pattern. Depending on this orientation, a crack might produce a clear, detectable signal in one configuration but remain virtually hidden in another.
Key Results and Diagnostic Implications
The research yields several critical insights for the future of machine diagnostics:
- Sensitivity to Unbalance: The vibrational behavior of both harmonic and 1/2-order subharmonic resonances is highly sensitive to the angular position of the unbalance. This suggests that unbalance is not just a noise factor but a primary variable in crack detection.
- Distinct Signatures: Because the dynamic response in nonlinear systems is fundamentally different, existing fault diagnosis rules derived from linear models are likely to fail or provide false negatives when applied to real-world nonlinear machinery.
- Subharmonic Indicators: The 1/2-order subharmonic resonance remains a key indicator for crack identification, but its appearance is heavily modulated by the system's nonlinearities, requiring more sophisticated analysis than simple peak-tracking.
Conclusion
This study underscores the necessity of accounting for system nonlinearities when developing crack detection algorithms. The interaction between crack breathing and nonlinear bearing supports creates a dynamic environment where the angular position of unbalance becomes a dominant factor. Future work should focus on expanding these models to include higher-order resonances and experimental validation to map these specific effects across a wider range of industrial rotor configurations. For engineers, the takeaway is clear: a one-size-fits-all approach to vibration analysis is insufficient for the complex, nonlinear realities of modern rotating machinery.
Summary updated