A novel tilting pad journal bearing model including pivot flexibility as well as temporal fluid inertia effects on the thin film fluid flow aims to accurately predict the bearing forced performance. The predictive model also accounts for the thermal energy transport effects in a TPJB. A Fortran program with an Excel GUI models TPJBs and delivers predictions of the bearing static and dynamic forced performance. The calculation algorithm uses a Newton-Raphson procedure for successful iterations on the equilibrium pad radial and transverse displacements and journal center displacements, even for bearings pads with very soft pivots. The predictive model accounts for the effect of film temperature on the operating bearing and pad clearances by calculating the thermal expansion of the journal and pad surfaces. The pad inlet thermal mixing coefficient (lambda) influences moderately the predicted fluid film temperature field. Pad pivot flexibility decreases significantly and dominates the bearing stiffness and damping coefficients when the pivot stiffness is lower than 10% of the fluid film stiffness coefficients (with rigid pivots). Pivot flexibility has a more pronounced effect on reducing the bearing damping coefficients than the stiffness coefficients. Pad pivot flexibility may still affect the bearing behavior at a light load condition for a bearing with a large pad preload. Pad pivot flexibility, as well as the fluid inertia and the pads’ mass and mass moment of inertia, could influence the bearing impedance coefficients, in particular at high whirl frequencies. The stiffness and damping coefficients of a TPJB increase with a reduction in the operating bearing and pad clearances. The work delivers a predictive tool benchmarked against a number of experimental results for test bearings available in the recent literature. The static and dynamic forced performance characteristics of actual TPJBs can not be accurately predicted unless their pad flexibility and pivot flexibility, fluid film temperature, pad inlet thermal mixing coefficient, operating bearing and pad clearances, among others are well known in advance. However, the extensive archival literature showcasing test procedures and experimental results for TPJBs does not report the above parameters. Thus, reasonable assumptions on the magnitude of certain elusive parameters for use in the predictive TPJB model are necessary.

Rolling-element bearings are widely used in various mechanical and electrical applications. Accordingly, a reliable bearing health condition monitoring system is very useful in industries to detect incipient defects in bearings, so as to prevent machinery performance degradation and malfunction. Although several techniques have been reported in the literature for bearing fault detection and diagnosis, it is still challenging to implement a bearing condition monitoring system for real-world industrial applications because of the complexity of bearing structures and noisy operating conditions. The objective of this thesis is to develop a novel intelligent system for more reliable bearing fault diagnostics. This system involves two sequential processes: feature extraction and decision-making. The proposed strategy is to develop advanced and robust techniques at each processing stage so as to improve the reliability of bearing condition monitoring. First, a novel wavelet spectrum analysis technique is proposed for the representative feature extraction. This technique applies the wavelet transform to demodulate the resonance signatures that are related to bearing health conditions. A weighted Shannon function is proposed to synthesize the wavelet coefficient functions to enhance feature characteristics. The viability of this technique is verified by experimental tests corresponding to various bearing health conditions. Secondly, an enhanced diagnostic scheme is developed for automatic decision-making. This scheme consists of modules of classification and prediction: a novel neuro-fuzzy classifier is developed to effectively integrate the strengths of the selected fault detection techniques (i.e., the resulting representative features) for a more accurate assessment of bearing health conditions; a novel multi-step predictor is proposed to forecast the future states of bearing conditions, which will be used to further enhance the diagnostic reliability. The investigation results have demonstrated that the developed intelligent diagnostic system outperforms other related bearing fault diagnostic schemes.

Rolling-element bearings are widely used in various mechanical and electrical applications. Accordingly, a reliable bearing health condition monitoring system is very useful in industries to detect incipient defects in bearings, so as to prevent machinery performance degradation and malfunction. Although several techniques have been reported in the literature for bearing fault detection and diagnosis, it is still challenging to implement a bearing condition monitoring system for real-world industrial applications because of the complexity of bearing structures and noisy operating conditions. The objective of this thesis is to develop a novel intelligent system for more reliable bearing fault diagnostics. This system involves two sequential processes: feature extraction and decision-making. The proposed strategy is to develop advanced and robust techniques at each processing stage so as to improve the reliability of bearing condition monitoring. First, a novel wavelet spectrum analysis technique is proposed for the representative feature extraction. This technique applies the wavelet transform to demodulate the resonance signatures that are related to bearing health conditions. A weighted Shannon function is proposed to synthesize the wavelet coefficient functions to enhance feature characteristics. The viability of this technique is verified by experimental tests corresponding to various bearing health conditions. Secondly, an enhanced diagnostic scheme is developed for automatic decision-making. This scheme consists of modules of classification and prediction: a novel neuro-fuzzy classifier is developed to effectively integrate the strengths of the selected fault detection techniques (i.e., the resulting representative features) for a more accurate assessment of bearing health conditions; a novel multi-step predictor is proposed to forecast the future states of bearing conditions, which will be used to further enhance the diagnostic reliability. The investigation results have demonstrated that the developed intelligent diagnostic system outperforms other related bearing fault diagnostic schemes.

Synchronous generators play a vital role in power systems. One of the major mechanical faults in synchronous generators is related to bearings. The popular vibration analysis method has been utilized to detect bearing faults for years. However, bearing health monitoring based on vibration analysis is expensive. One of the reasons is because vibration analysis requires costly vibration sensors and the extra costs associated with its proper installation and maintenance. This limitation prevents continuous bearing condition monitoring, which gives better performance for rolling element bearing fault detection, compared to the periodic monitoring method that is a typical practice for bearing maintenance in industry. Therefore, a cost effective alternative is necessary. In this study, a sensorless bearing fault detection method for synchronous generators is proposed based on the analysis of electrical signatures, and its bearing fault detection capability is demonstrated. Experiments with staged bearing faults are conducted to validate the effectiveness of the proposed fault detection method. First, a generator test bed with an in- situ bearing damage device is designed and built. Next, multiple bearing damage experiments are carried out in two vastly different operating conditions in order to obtain statistically significant results. During each experiment, artificially induced bearing current causes accelerated damage to the front bearing of the generator. This in-situ bearing damage process entirely eliminates the necessity of disassembly and reassembly of the experimental setup that causes armature spectral distortions. The electrical fault indicator is computed based on stator voltage signatures without the knowledge of machine and bearing specific parameters. Experimental results are compared using the electrical indicator and a vibration indicator that is calculated based on measured vibration data. The results indicate that the electrical indicator can be used to analyze health degradation of rolling element bearings in synchronous generators in most instances. Though the vibration indicator enables early bearing fault detection, it is found that the electrical fault indicator is also capable of detecting bearing faults well before catastrophic bearing failure.

A fixed geometry hybrid thrust bearing is investigated with three different supply orifice diameters. The test rig uses a face-to-face thrust bearing design, with the test bearing acting as the rotor loading mechanism. A hydraulic shaker applies the static axial load, which is reacted by a second thrust bearing. The rotor is supported radially by two water-lubricated fluid film journal bearings and is attached to a 30,600 rpm motor via a high speed coupling with very low axial stiffness. Thrust bearings with three different orifice diameters (1.63, 1.80, and 1.93 mm) are tested for a range of supply pressures, fluid film thicknesses, and rotational speeds. The water-lubricated test bearings have eight pockets, with feed orifices located centrally in each pocket. Experimental results are comparted to predictions found using bulk flow model HYDROTHRUST. Analysis of the data reveals generally good agreements between predictions and measurements. Thrust-bearing inlet supply and inner radius flow rates all decreased with decreasing orifice diameters and bearing axial clearances. In most cases, the bearings with larger orifice diameters exhibit higher recess pressure ratios, operating clearances, and flow rates. The largest orifice diameter configuration does not display higher recess pressure ratios or operating clearances at high speeds for some supply pressures, but it does continue to require additional lubricant flow rate compared to the smaller orifice bearings. In these cases, the results are not reflected in predictions, which otherwise correlate very well with experimental measurements. Estimations of static loading axial stiffness are obtained using experimental results. An optimum hybrid thrust bearing orifice diameter will depend on the conditions of individual applications. Larger orifices generally provide larger operating clearances and higher stiffnesses, but also require higher flow rates. For most applications, a compromise of bearing performance parameters will be desired. The test results and comparisons presented will aid in sizing orifice diameters for future hybrid thrust bearing designs.

none

Gas bearings enable successful applications in high speed oil-free microturbomachinery. This thesis presents analysis and experiments of the dynamic performance of a rotor supported on Rayleigh step gas bearings. Comprehensive experiments demonstrate that Rayleigh step hybrid gas bearings exhibit adequate stiffness and damping capability in a narrow range of shaft speeds, up to ~ 20 krpm. Rotor coastdown responses were performed for two test bearing sets with nominal radial clearance of 25.4 ?m and 38.1 ?m, respectively. A near-frictionless carbon (NFC) coating was applied on the rotor to reduce friction against its bearings at liftoff and touchdown. However, the rotor still experienced dry friction at low shaft speeds (below ~ 4,000 rpm). Experiments show that the supply pressure raises the rotor critical speed and decreases the system damping ratio. The geometry of the Rayleigh steps distributed on the rotor surface generates a time varying pressure field and results in a sizable 4X super synchronous component of bearing transmitted load. The external supply gas pressure affects slightly the onset speed of instability of the rotor-bearing system. The unstable whirl frequencies are nearly fixed at the system natural frequency (~ 120 Hz). Analysis with a finite element model predicts the stiffness and damping force coefficients for the bearing accounting for a purely hydrodynamic operation condition. Predictions show the synchronous stiffness and damping coefficients decrease with shaft speed. Predicted threshold speeds of instability are lower, ~ 50% or less than the measurement due to the analytical model limitations assuming a grooved stator. The predicted synchronous responses to imbalance correlate well with the measurements, however. The Rayleigh step gas bearing shows similar characteristics to the flexure pivot tilting pad bearing (FPTPB) tested in 2003. However, the test Rayleigh step gas bearings exhibit a much reduced stable operating speed range, below 20 krpm. The maximum speed achieved is much lower to that determined with an identical rotor supported on FPTPBs, i.e. rotor dynamically stable up to 100 krpm. The FPTPB is more reliable in high speed oil-free applications due to its excellent stability characteristics.

Magnetic bearings use variable attractive forces generated by electromagnetic control coils to support rotating shafts with low friction and no material wear while providing variable stiffness and damping. Rotor deflections are stabilized by position feedback control along two axes using non-contacting displacement sensors. These sensor signals contain sensor runout error which can be represented by a Fourier series composed of harmonics of the spin frequency. While many methods have been proposed to compensate for these runout harmonics, most are computationally intensive and can destabilize the feedback loop. One attractive alternative is to increase the number of displacement sensors and map individual probe voltages to the two independent control signals. This approach is implemented using a circular sensor array and single weighting gain matrix in the present work. Analysis and simulations show that this method eliminates runout harmonics from 2 to n-2 when all sensors in an ideal n-sensor array are operational. Sensor failures result in reduced synchronous amplitude and increased harmonic amplitudes after failure. These amplitudes are predicted using derived expressions and synchronous measurement error can be corrected using an adjustment factor for single failures. A prototype 8-sensor array shows substantial runout reduction and bandwidth and sensitivity comparable to commercial systems. Nonlinear behavior in homopolar magnetic bearings is caused primarily by the quadratic relationship between coil currents and magnetic support forces. Governing equations for a permanent magnet biased homopolar magnetic bearing are derived using magnetic circuit equations and linearized using voltage and position stiffness terms. Nonlinear hardening and softening spring behavior is achieved by varying proportional control gain and frequency response is determined for one case using numerical integration and a shooting algorithm. Maximum amplitudes and phase reversal for this nonlinear system occur at lower frequencies than the linearized system. Rotor oscillations exhibit amplitude jumps by cyclic fold bifurcations, creating a region of hysteresis where multiple stable equilibrium states exist. One of these equilibrium states contains subharmonic frequency components resulting in quasiperiodic rotor motion. This nonlinear analysis shows how nonlinear rotor oscillations can be avoided for a wide range of operation by careful selection of design parameters and operating conditions.

Thesis (D. Min.)--Southwestern Baptist Theological Seminary, 2001. / Includes bibliographical references (leaves 125-134, 68-72).

Thesis (D. Min.)--Austin Presbyterian Theological Seminary, 2000. / Includes bibliographical references (leaves 177-188).