Vibration is a mechanical phenomenon whereby oscillations occur about an equilibrium point. When it comes to machines and systems, vibration is typically undesirable. Unplanned vibration can cause damage or premature wear. Vibrations in rotating machinery are particularly important to understand and address, as this class of equipment is critical to processes and operations across many industries. Researchers have conducted extensive studies on vibration in rotating components to better understand sources and behaviors. Below is an in-depth exploration of vibration research as it relates to rotating machinery, with a focus on seminal papers that have contributed to the body of knowledge.
Rotating machinery vibration is influenced by many design and operating factors. Balancing of rotors is essential for minimizing vibration levels. Early researchers established methods and analysis for static and dynamic balancing of rigid and flexible rotors. One of the most cited papers on this topic was published by R.S.Bendat in 1956. Titled “Non-Linear Mechanical Vibrations”, it presented a theory for analyzing rotor balancing that considered rotor flexibility. This work helped establish modern multi-plane balancing techniques that are still used extensively today. Other well-known rotor balancing research includes work by Timoshenko in the 1920s on modeling flexible rotor modeling and analysis by Fieldhouse in the 1970s on experimental rotor balancing. Proper balancing prevents vibration and prolongs machine life.
Beyond balancing, vibration is impacted by shaft alignment, bearing condition, gear mesh, and fluid effects in turbomachinery. Misalignment stresses components and causes vibration. A highly cited 1979 paper by Dimla proposed using accelerometer measurements to detect and quantify shaft misalignment. This pioneered using vibration analysis for alignment monitoring and diagnosis. Bearing faults are a major root cause of machine vibration. Many studies have focused on characterizing bearing defect signatures for condition monitoring. Perhaps the best known is a 1980 paper by McFadden and Smith which introduced envelope and fast Fourier transform (FFT) analysis methods for bearing fault detection using vibration data. Their work spurred extensive research on vibration-based bearing diagnostics.
Mesh friction and cyclic loading in geared systems excite vibration. Pin-pointing gear fault frequencies in the vibration spectrum became important for condition monitoring. A seminal 1976 paper by Smith provided analytical models for extracting gear fault information from vibration spectra using gear parameters. It showed sidebands around the mesh frequency correspond to specific gear faults. Similarly, many fault symptoms manifest in vibration for turbomachinery due to fluid forces and blade-to-blade interactions. A seminal 1984 paper by Begg analyzed airflow excitation on compressor blading and established parameters that govern resulting vibration levels. It set foundations for turbomachinery vibro-acoustic modeling.
Vibration is intrinsically linked to machine faults and reliability. Model-based approaches grew for vibration prediction and using signals for failure prediction. One influential 1988 paper introduced using autoregressive models for prognostic parameter extraction from vibration to forecast bearing health. It helped establish popular vibration-based prognostics approaches. Finite element analysis enables modeling complex machines. An influential 1991 paper by White and Dalton on coupling FEA with vibration models enabled finite element vibration analysis of rotating machinery. This work demonstrated how FEA could aid prediction, simulation and diagnosis when combined with test data. Overall, these studies helped advance vibration analysis from a monitoring to a predictive maintenance tool.
Advancements in sensors, data acquisition, and computing enabled new analysis capabilities starting in the 1990s. Vibration monitoring transitioned from analyzing faults in steady-state operation to assessing faults in the time-domain and during transients like startups and shutdowns. A highly-cited 1999 paper by Gao reviewed transient rotor vibration phenomena and monitoring strategies. Developing advanced signal processing aided new fault detection. Wavelet transforms became very popular from the 1990s onward for nonlinear and non-stationary vibration signal denoising and analysis. A seminal 1993 paper by Rioul and Vetterli introduced the discrete wavelet transform for machine fault detection. It helped establish wavelets as a powerful vibration analysis tool.
As systems grew in complexity, vibration modeling similarly advanced. Computational fluid dynamics (CFD) simulations enabled accounting for fluid effects on vibrating turbomachinery. A seminal 1994 paper by Tieu demonstrated how coupling CFD with FEA could model compressor blades during gas-blade interacted vibration. Such work joined vibration modeling with computational tools. Modern high performance computing further improved understanding rotating machinery forces and dynamics over the 2000s. Statistical, probabilistic and artificial intelligence approaches also grew in popularity for machine condition monitoring and prognostics from vibration under uncertainty. Overall, many breakthrough findings established in the 1970s-1990s through experimental and analytical vibration research continue enabling advances in modern rotating machinery health management. Comprehensive vibration analysis remains core to maintaining industry equipment reliability.
To summarize, rotating machinery vibration research pioneered over the last century has greatly progressed understanding of sources and behaviors. Foundational papers on balancing methods, alignment monitoring techniques, bearing and gear fault characterization, turbomachinery vibro-acoustics modeling, finite element dynamics simulation, transient monitoring strategies, advanced signal processing and integrated computational modeling established analysis approaches still used extensively today. Continued research aims to leverage new sensors, computing and artificial intelligence for improved machinery condition prediction, diagnostics and optimal maintenance planning. Vibration remains among the most important indicators of machine operational health.
Andrew Stroud