A Review of Tilting Pad Bearing Theory PDF

Hindawi Publishing Corporation
International Journal of Rotating Machinery

Volume 2011, Article ID 908469, 23 pages
doi:10.1155/2011/908469
Review Article
A Review of Tilting Pad Bearing Theory
Timothy Dimond, Amir Younan, and Paul Allaire

Rotating Machinery and Controls Laboratory, Department of Mechanical and Aerospace Engineering, University of Virginia,
122 Engineer’s Way, Charlottesville, VA 22911, USA
Correspondence should be addressed to Timothy Dimond, [email protected]
Received 24 January 2011; Accepted 5 May 2011
Academic Editor: R. Kirk
Copyright © 2011 Timothy Dimond et al. This is an open access article distributed under the Creative Commons Attribution
License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly
cited.
A theoretical basis for static and dynamic operation of tilting pad journal bearings (TPJBs) has evolved over the last 50 years.
Originally demonstrated by Lund using the pad assembly method and a classic Reynolds equation solution, the current state
of the art includes full thermoelastohydrodynamic solutions of the generalized Reynolds equation that include fluid convective
inertia effects, pad motions; and thermal and mechanical deformations of the pads and shaft. The development of TPJB theory is
reviewed, emphasizing dynamic modeling. The paper begins with the early analyses of fixed geometry bearings and continues
to modern analyses that include pad motion and stiffness and damping effects. The development of thermohydrodynamic,
thermoelastohydrodynamic, and bulk-flow analyses is reviewed. The theories of TPJB dynamics, including synchronous and
nonsynchronous models, are reviewed. A discussion of temporal inertia effects in tilting pad bearing is considered. Future trends
are discussed, and a path for experimental verification is proposed.
1. Introduction The hydrodynamic action in fluid film bearings is fun-

damentally a fluid-structure interaction effect. When these
Rotating machinery such as pumps, compressors, fans, effects are linearized and perturbed in the two orthogonal
turbines, and generators are ubiquitous in industrial set- radial directions relative to the shaft, they result in equivalent
tings. Fluid film journal bearings are suited to applications lateral stiffness and damping coefficients which can then be
experiencing higher speeds and loads, often in excess of
used in lateral vibration analysis of rotors.
90 m/s surface velocity and with typical bearing specific
loads of 700 kPa–3.5 MPa. In these bearings, hydrodynamic Tilting pad journal bearings are a source of both static
action is used to support the rotor on a thin lubricating support and dynamic stiffness and damping. Tilting pad
film. Typical film thicknesses between the rotor and the journal bearings have a number of pads, typically four or
bearing surface are on the order of 100 μm. This compares five. A four pad bearing is shown in Figure 1. Each pad
to shaft diameters on the order of 25 mm to 1 m, leading in the bearing is free to rotate about a pivot and cannot
to typical diametral clearance ratios for fluid film bearings support a moment. As a result, the destabilizing forces are
of 1-2 μm/mm. With the dimensions and rotational speeds greatly reduced or eliminated, and the bearings are no longer
for high-speed applications, lubricating flows are at the low a potential source of rotordynamic instability. This feature
end of the turbulent regime for oil lubricants. There is not a has made tilting pad journal bearings the standard fluid-film
single accepted Reynolds number corresponding to the onset bearing for most high-speed applications.
of turbulence in bearing lubrication. The prior literature High-speed rotordynamic applications often have rotors
indicates a range, with turbulence onset taken to occur over that pass through one or two bending critical speeds as
a range of Re from 1000 to 1500 [1–5]. Some applications the machines are accelerated to the operating speed. The
use water or other low-viscosity process fluid as the bearing damping from the fluid film bearings is required to safely
lubricant, which results in a highly turbulent lubricating pass through these bending critical speeds as the rotating
flow. element is accelerated. The damping also helps suppress
2 International Journal of Rotating Machinery
Many of the investigators of bearing dynamics were

concerned with the onset of rotordynamic instability, so
stability assessments in the literature are common. Less
common are predictions of critical speeds and unbalance
response. The previous work considered in this paper is
not comprehensive, but is representative of developments
of bearing models. The works cited in this paper and their
references do give a comprehensive treatment of bearing
modeling developments.
Section 5 summarizes the current state of the art for
the thermoelastohydrodynamic (TEHD) lubrication theory
for bearings. The effects of the hydrodynamics due to
fluid flow, the energy equation for heat transfer within
the bearing, mechanical deflections, and thermal growth
are summarized. Section 6 summarizes the averaged flow
approaches, including the mixing length theory originally
proposed by Constantinescu, and the bulk flow approach
proposed by Hirs. In these approaches, the properties of the
Figure 1: Tilting pad journal bearing. lubricant film are averaged. The TEHD analysis described
in Section 5 is based on a differential equation approach
to characterization of the lubricating film and the bearing
potentially destabilizing forces from sources such as radial components. As a result, the flow characteristics are modeled
seals, balance pistons, impeller eye seals, internal friction fits, locally, including across the lubricant film. The averaged flow
and unbalanced electromagnetic forces [6]. approaches described in Section 6, which can be extended to
Characterization of the dynamic response of tilting pad THD analyses, rely on averaging of flow properties across the
bearings is vital to successful design of high-speed rotating lubricating film.
machinery. Theoretical models exist for prediction of the Section 7 summarizes the current state of the art in
dynamic response. These are particularly important at the tilting pad bearing theory for bearing dynamics. Tilting
design stage of modern high-speed rotating machinery. pad bearings are the only practical bearings used in high-
These models have evolved from analytical solutions of the speed flexible rotor industrial machines. The three tilting
lubricating film of fixed geometry bearings to full finite ele- pad bearing dynamic models that appear recently in the
ment and finite difference numerical solutions that include literature are also considered. These dynamic models, (1)
analysis of the lubricating flow; the energy balance between the full stiffness-damping model (full KC); (2) the reduced
the lubricant, the bearing, and the rotor; and mechanical and order stiffness-damping-mass (KCM) model; (3) the syn-
thermal deformations of the shaft and bearing pads. chronously reduced model, are described in Sections 5.4–
Additionally, there is still some controversy within the 7.4. The frequency-dependent stiffness damping (frequency-
rotordynamic community over the proper dynamic model- dependent KC), an implicit version of the full KC model, is
ing method for tilting pad journal bearings. Some researchers also considered.
question whether consideration of excitation frequencies Section 8 discusses two approaches to fluid temporal
other than rotor operating speed is necessary. There is also inertia effects in fluid film bearings and the relative impor-
continued discussion on the number of degrees of freedom to tance and limitations to their analysis.
retain within the bearing, either implicitly or explicitly, when Discussion and conclusions are provided in Section 9.
nonsynchronous dynamic models are considered. The issue This summarizes the evolution of bearing modeling and
of the presence of fluid temporal inertia effects also arises as discusses future trends and opportunities for experimental
part of these discussions. This paper will review and discuss confirmation.
those issues.
This paper is organized into nine sections. In Section 2, 2. Tilting Pad Bearing Modeling Development
the history of the development of tilting pad bearing dynamic
theory is briefly discussed, including a stiffness-damping 2.1. Early Lubrication Theory and Nondimensionalization.
bearing model. This discussion considers the history of lubri- The fundamental lubrication equation was originally for-
cation theory with an emphasis on development of bearing mulated by Reynolds in 1886 [7]. He developed the theory
dynamic models. Section 3 reviews initial developments of that explained the experimental results of Tower and Petroff.
bearing dynamic models, including key developments in Reynolds assumed that the flow could be treated as isoviscous
fixed geometry bearings and synchronously reduced tilting and laminar. These assumptions were well justified due to
pad bearing dynamic models. Section 4 reviews later devel- typical bearing operating conditions in the late 1800s. He
opments, including nonsynchronous tilting pad bearing also assumed that since the lubricating film was thin in the
dynamic models, thermohydrodynamic (THD) lubrication radial direction compared to the circumferential and axial
analysis, and thermoelastohydrodynamic (TEHD) lubrica- directions that there was no pressure gradient across the film
tion analysis. radially. The dominant flow characteristic was then due to
International Journal of Rotating Machinery 3
V1 clearance cd as the characteristic length, and ωD as the

characteristic velocity:
ξ U1
ρωDcd
Re = , (3)
μ
where ρ is the density of the lubricant, ω is the shaft
η h(ξ) rotational speed, and D is the diameter of the shaft. However,
the Reynolds number in this form overemphasizes the effect
of the shaft diameter, so an alternative reduced Reynolds
number is also typically defined:
U0

ρωDcd cd ρωcd2
Figure 2: Plain slider. Re∗ = = . (4)
μ D μ
Both Re and Re∗ consider the effect of the fluid density but
shearing effects. By simplification of the fluid Navier-Stokes do not include the load on the bearing. However, Re and Re∗
and continuity equations consistent with the assumptions, a are the currently accepted methods of nondimensionalizing
single equation describing lubricating flows was derived: the lubricating flow turbulence characteristics in modern
bearing treatments. Recent papers often report the bearing
d dp d 3 dp dh
h3 + h = 6μ (Uo + U1 ) + 2V1 . specific load W/(LD) in conjunction with Re or Re∗ when
dη dη dz dz dη describing results instead of using the Sommerfeld number.
(1) The analyses of Reynolds and Sommerfeld focused solely
on the solution to the hydrodynamic flow field problem,
Equation (1) is the classic Reynolds equation, expressed
described by (1). Later research included simultaneous
in terms of a single pad local coordinate system. In (1), η is
solutions of the energy equation, resulting in a thermo-
the local coordinate direction along the slider, h is the film
hydrodynamic (THD) analysis. Another refinement to the
thickness, p is the developed pressure in the lubricating film,
problem was to include mechanical deformation effects due
μ is the lubricant dynamic viscosity, z is the axial direction,
to applied pressures and thermal growth, resulting in the
Uo and U1 represent the relative velocity between slider and
thermoelastohydrodynamic (TEHD) theory. The historical
pad in the η direction, and V1 represents the squeeze velocity
development of these theories is discussed briefly in Sections
between the slider and the pad. The geometry and velocity
3-4.
vectors for a plain slider is shown in Figure 2. Reynolds noted
that since the lubricating film thickness is small compared to
the radius of curvature of typical bearings, (1) can be written 2.2. Geometric Considerations. Other bearing design geo-
in a Cartesian coordinate system without significant loss of metric properties will enter the discussion. The dynamic
accuracy. Reynolds was able to provide some series solutions properties of bearings are affected by the bearing design
to (1) to find the pressure field within the bearing, but the geometry. The geometric properties are as follows.
first closed-form solution to Reynolds equation was found (i) Bearing preload, m = 1 − cb /c p . The difference
by Sommerfeld [8]. Sommerfeld also developed a group of between the bearing clearance cb and the pad clear-
bearing parameters for nondimensionalization of the results. ance c p form a converging hydrodynamic wedge
The Sommerfeld number is purely through geometry, even for a centered rotor.
2
μNLD D (ii) Pivot location: pivot location relative to the leading
So = . (2)
W cd edge of the pad expressed as a percentage of pad arc
length
The Sommerfeld number recognizes the effect of the net
force applied to the shaft at the bearing, W, the projected area (iii) Load orientation: napplied load relative to the
of the bearing LD, the clearance ratio cd /D, the rotational bearing pads. Load on pad and load between pad
speed in rev/s N, and the dynamic viscosity of the lubricating configurations are typical.
fluid μ as parameters that affect bearing operation in the
laminar regime. Many of the theoretical results in the 3. Fixed Geometry and Synchronously Reduced
literature use (2) for nondimensionalization. Bearing Dynamics
One of the assumptions in arriving at the Sommerfeld
number is that the lubricating flow is laminar. While The solutions by Reynolds and Sommerfeld were for the
appropriate for lower surface speed bearings where the pressure field and the net forces of the lubricating oil
laminar flow assumption can be justified, the assumption is film. Most analysts of the era considered the rotor to be
increasingly violated due to the high rotational speeds typical simply supported at the bearings. As the understanding
for many modern bearings. An additional dimensionless of the linearized bearing response improved, researchers
group for evaluating bearings operating in the turbulent recognized the equivalent stiffness and damping effects
regime is the Reynolds number using the bearing diametral provided by the lubricating film.
The first attempts to quantify the dynamic response of

the lubricating film itself were made by Stodola [9] and
by Stodola’s student Hummel [10]. Stodola and Hummel
were able to obtain a solution for the oil film stiffness
and correctly obtained analytical linearized direct and cross-
coupled stiffness terms based on the Sommerfeld closed
form solution to (1). However, they did not recognize
the damping effect of the oil film, and their predictions
indicated that an unstable rotor would have vibration levels
that increased without bound. This is a limitation of linear
analysis that does not consider the nonlinear behavior of the
oil film under large excursions or practical considerations
such as contact between the rotor and the stator. Hummel
acknowledged in his thesis that the rotor vibration remained
finite but did not provide a specific mechanism.
Another early analysis that recognized the effect of Figure 3: Michell Combined Tilting Pad Journal and Thrust
bearing flexibility on critical speeds was reported by Linn and Bearing [12].
Prohl [11]. While not referring to oil film flexibility explicitly,
a general bearing flexibility was assumed and the resulting
lowering of the critical speeds compared to a rigid support noncircular bearing bores were discovered to enhance the
assumption was demonstrated. stability margin and were a lower cost option to a tilting pad
Fixed geometry radial bearings were standard in the arrangement.
first half of the 20th century, and tilting pad bearings only Sternlicht [16] presented a finite difference solution to
saw significant adoption begin during the 1960s. However, the Reynolds equation based on an isoviscous lubricant. The
the tilting pad thrust bearing was invented independently finite difference solution was used to calculate the developed
by Kingsbury and Michell. Michell also invented the tilting pressure field, which was then integrated to calculate forces.
pad journal bearing and installations of the tilting pad The force solution was then perturbed to determine eight
journal bearing appear as early as 1916 [12]. An installation stiffness and damping coefficients based on rotor motion at
of a combined tilting pad journal and thrust bearing on the journal. These eight coefficients are widely accepted as
the H.M.S. Mackay of the British Royal Navy, which was a proper model for fixed geometry journal bearings where
launched in 1918, is shown in Figure 3. Fixed geometry temporal inertia is not important.
bearings remained the standard in the first half of the 20th Solutions to the perturbed Reynolds equation also began
century due to reduced cost for fixed geometry installations, to appear in textbooks, including the ones by Smith [17],
the higher parasitic losses associated with tilting pad bearings Pinkus and Sternlicht [18], and Tondl [19]. Smith’s treatment
compared to fixed geometry bearings, lower load capacity of the subject was brief, but did include the eight stiffness
of tilting pad bearings [13], and lower operating speeds and damping coefficients. Pinkus and Sternlicht investi-
that could tolerate the destabilizing effects of fixed geometry gated the stability of rotors supported in plain journals,
bearings. Boyd and Raimondi in particular stated [14]: but the solutions were performed in polar coordinates.
Modern rotordynamics analyses are performed in cartesien
“[T]he plain journal bearing compares favorably coordinates for simplicity and for direct comparison to
with the pivoted-pad bearing and by many vibration measurements. Tondl’s text was an investigation of
criteria is somewhat superior to the latter.” various sources of rotordynamic instability, of which fixed
geometry bearings were a significant contributor. Tondl’s
Because of these factors, the perceived drawbacks to treatment accounted for direct stiffness, direct damping, and
tilting pad bearings outweighed the benefits. cross-coupled stiffness terms. His investigation included a
The advantages of tilting pad bearings in removing the treatment of the perturbed Reynolds equation for both linear
bearings as a source of self-excited vibrations was originally and nonlinear rotor vibrations. Tondl also considered the
recognized by Hagg in 1946 [15]. Hagg presented experi- benefits of noncircular fixed geometry bearing stator profiles,
mental results for several tilting pad bearings, including 3- including lobed bearings.
pad, 4-pad, 5-pad, and 6-pad bearings. However, the fluids Even with the improvements to fixed geometry designs to
model that Hagg employed to explain the stabilizing features enhance stability, there is still a limit where the destabilizing
of these bearings was fundamentally incorrect. A linear flow forces are high enough to overcome the damping and drive
profile was assumed that did not account for the Reynolds the rotor unstable. Typically, the limit is reached when the
equation, (1). operating frequency is greater than twice the first bending
While the development reported by Hagg was significant, natural frequency [15]. These limits began to be reached on
many analysts continued to work with plain journal bearings. a more consistent basis in the 1960s. The advantage of the
The benefit of improved stability margin was still not stabilizing effect of the tilting pad bearing was then seen
significant enough to designers of the era to overcome to overcome the drawbacks of decreased load capacity and
the perceived drawbacks discussed previously. Additionally, higher parasitic losses.
Adoption of tilting pad bearings was also made easier coefficients were synchronously reduced. A finite element
by analytical solutions that allowed designers to understand formulation of (1) was used, so it was an isoviscous, laminar
the dynamic properties. One of the major advances in analysis. The finite element method was used to produce
understanding the dynamics of tilting pad bearings came single pad solutions with pad assembly similar to Lund [20].
from Lund’s landmark paper in 1964 [20]. Based on analyses The reduced stiffness and damping coefficients were plotted
of fixed pads, which are essentially partial arc bearings, versus Sommerfeld number. The pad inertia effects were
stiffness, and damping coefficients were calculated. The neglected in the analysis.
equations of motion for the pads were then considered based Nicholas and Kirk [26] examined several fixed and
on the calculated fixed pad stiffness and damping values, tilting pad bearings, including four-pad and five-pad tilting
which were then summed vectorially to calculate the full pad bearings, for application to axial compressors. The
bearing coefficients. This pad assembly method was not a synchronously reduced stiffness and damping coefficients
simultaneous solution of the lubrication problem for all were used for unbalance and stability analyses. This paper
the pads. The pad assembly method is less computationally explored the effect of manufacturing tolerances on the per-
expensive and more approximate. However, this approach formance of the several bearing types. For a four-pad tilting
was suitable for the computers available in the 1960s. The pad bearing the synchronously reduced stiffness was shown
dynamic coefficients were then reduced synchronously, or to vary by 50 percent and synchronously reduced damping
using the shaft rotational frequency as the excitation fre- by half an order of magnitude due to typical manufacturing
quency of interest, to obtain the eight stiffness and damping tolerances, though that was an extreme case. Variations of 10
coefficients related to rotor motion. Results were presented percent in the dynamic coefficients were more typical. The
for four-pad, five-pad, and six-pad tilting pad bearings. Lund effect of bearing preload on the compressor stability margin
recast rotations of the pads as equivalent translations. Later was investigated, and lower preload was determined to
solutions treated pad rotations as rotational motion. enhance stability because the bearing stiffness was reduced,
The landmark work by Lund led to a significant allowing more motion for the damping to be effective. The
research effort to extend the analyses of tilting pad bearings. four-pad bearing configuration for axial compressors was
Thermohydrodynamic and TEHD solutions and turbulence explored further in [27]. Nicholas and Kirk again used the
corrections for high rotational speeds also begin to appear synchronously reduced bearing dynamic coefficients for both
for tilting pad bearings. The use of synchronously reduced unbalance response and stability margin.
coefficients to described the linearized dynamics became the Jones and Martin [28] performed another geometric
norm, following the results reported by Lund [20]. study of tilting pad bearing characteristics, considering
Orcutt [21] followed the same basic approach as Lund different preloads; bearing L/D ratios; 3, 5, and 7 tilting
[20] by developing a partial arc bearing solution. He pads; and load orientation. The analysis was used to calculate
accounted for turbulence effects in the lubricating film using minimum oil film thickness, average pad temperatures,
the analysis of Ng and Pan [22]. Similar to Lund, Orcutt bearing parasitic power losses, and the synchronously
solved the lubrication problem for each pad individually and reduced stiffness and damping coefficients. The analysis was
then performed a synchronously reduced assembly method described as quasi-THD, since average pad temperature was
similar to Lund. While not a comprehensive formulation, used to calculate the average oil viscosity for each pad. The
modifications to (1) were included to account for turbulence modeling included turbulence effects and was performed
effects in the lubricating film without resorting to a full using finite difference methods.
Navier-Stokes solution. Orcutt considered different lubri- Ettles [13] developed a TEHD analysis of tilting pad
cants, different numbers of pads, and different pad preloads. bearings. The analysis included a generalized Reynolds equa-
Orcutt’s analysis indicated that symmetry in the tilting pad tion solution using the turbulence model of Constantinescu
bearing leads to isotropic bearing dynamic properties. This [29] and the local calculated Reynolds number to obtain an
is not strictly correct, since a simultaneous TEHD solution effective viscosity. The energy equation was simplified into a
and operating experience indicates differential heating of 1D solution, and the relative error compared to a 2D solution
the tilting pads, but symmetric tilting pad bearings such was calculated to have a maximum value of 3.52 percent for
as four-pad bearings in a load-between-pad configuration an L/D ratio of 9.9. Elastic deformation of the pads due to
are nearly isotropic. Orcutt’s results were plotted against applied loads and thermal expansion were also considered.
the Sommerfeld number. He also suggested that the results Ettles’ solution was a simultaneous, iterative solution for all
showed that operation above the first bending critical was the bearing pads. The nondimensional dynamic coefficients
possible, which was generally avoided by designers of the era were reduced synchronously and were plotted as a function
to avoid stability problems. Including pad preload was also of Sommerfeld number. The results were compared to
claimed to improve dynamic characteristics. It was shown dynamic experiments by Malcher [30] for four pad bearings.
later [23, 24] that low preload leads to more stable systems. The theoretical results were within 10 percent of the reported
Nicholas et al. [25] developed stiffness and damping measured values. The reduction of effective film stiffness and
coefficients for the five pad tilting pad journal bearing. damping due to pad flexibility was noted. A stability analysis
Several bearing configurations were considered, including of the pad motion as a check for pad flutter was also included.
load on pad and load between pad, different pivot offsets Hashimoto et al. [31] also developed a TEHD analysis
ranging from 0.5 to 0.55 and different bearing preloads suitable to large scale tilting pad bearings with two pads.
from 0 to 0.5. The reported effective stiffness and damping The large generator application could be supported on two
lower bearing pads in a load between pad configuration, of 1100–1400. Turbulence was accounted for by lowering
so the top pads were eliminated. The generalized Reynolds the inlet oil temperature into the pad, which was justified
equation with a turbulence model relating effective viscosity by measurements showing the improved heat transfer in
to local Reynolds number was solved simultaneously with a the bearing due to the onset of turbulence. As a result, the
1D energy equation and a deformation model for the bearing treatment of turbulence was purely empirical with no formal
pads. Pad preloads of 0, 0.1, and 0.2 were considered, and the turbulence model. The model also accounted for thermal and
pads were centrally pivoted for all cases. The results assuming mechanical deformations of the bearing pads and showed
a laminar fluid were compared to the result obtained when the resulting drop in effective stiffness and damping due
the turbulence model was included. The bearing stiffness for to deformation. A finite difference solution was employed.
the turbulent case was up to 20 percent lower for Sommerfeld Fillon et al. [36] also demonstrated the need to consider
number up to about 0.4. For higher Sommerfeld numbers, bearing element deformation to obtain accurate predictions
the stiffness for the turbulent case was higher than the
of bearing behavior.
laminar case by as much as 100 percent. A similar trend
Brockwell et al. [37] developed a THD solution similar
was observed in the damping coefficients, but the crossover
point was at a Sommerfeld number of 0.1-0.2. The dynamic to that of Ettles [34]. The two-dimensional THD model
coefficient reduction method was not explicitly stated but included pad thermal expansion and elastic deformations,
the reported coefficients are consistent with synchronous along with pivot flexibility. Beam theory was used to model
reduction. pad flexibility. Nondimensional forms of the hydrodynamic,
Knight and Barrett [32] presented a THD analysis of four energy, and heat transfer equations were presented. The
pad tilting pad bearings with central pivots in a load on viscosity terms were averaged across the film thickness,
pad configuration. No turbulence model was considered in resulting in a bulk-flow approach to the lubrication prob-
calculating the results. The solution was based on a finite lem. Dynamic coefficients were calculated for synchronous
element solution to the classic Reynolds equation combined excitations based on a perturbed Reynolds solution. Static
with a finite difference solution to the 2D energy equation. and dynamic results were compared to a test of a five-pad
A simultaneous solution for the pads was developed. The tilting pad bearing in a load-between-pad configuration. The
effective viscosity for each bearing pad was based on the running speeds tested were 900, 1800, 2700, and 3600 rpm.
cross-film average temperature for each bearing pad. This Qualitative trends were matched by the predictions, with
average viscosity was then treated as constant in the Reynolds differences of 10–15 percent between theory and experiment
solution. Heat transfer through the pads was treated as radial for equivalent bearing stiffness and damping. The agreement
conduction and shaft surface temperatures were based on the was improved by including the effect of shaft flexibility in the
overall average film temperature. The full bearing coefficients overall identification and analysis procedures.
were calculated, which were reduced synchronously for Hopf and Schüeler [38] performed investigations of the
presentation of the results. When compared to an isother- transition from laminar to turbulent flow in large turbine
mal calculation, a difference of 10–35 percent in dynamic bearings. The THD analysis was not described in detail, but
properties was reported. This demonstrated the effect of was able to predict measured temperatures in the bearing
temperature and viscosity on the stiffness and damping tilting pads. The sudden drop in bearing local temperature
coefficients. due to the onset of turbulence or Taylor vortices and the
Brugier and Pascal [33] also investigated tilting pad bear- resulting enhanced heat transfer was documented.
ings for large turbogenerator sets. The bearings considered Hyun et al. [39] also developed a THD solution to the
had three tilting pads. The TEHD model accounted for ther- tilting pad bearing performance. The generalized Reynolds
mal effects in the lubricant as well as mechanical and thermal equation was solved, using Reichardt’s wall formula and
deformations of the bearing pads. The model included a the turbulence model presented by Ng and Pan [22] to
generalized Reynolds equation and energy equation similar model turbulence. A three-dimensional model was used for
to that described in detail in Section 5. The solution for the the energy equation, with correction factors for cavitation.
dynamic coefficients was based on numerical differentiation Predictions were compared to experimental measurements
and simulated shaft perturbations within the code. The of film pressure, film temperature, pad temperature, and load
description of coefficient extraction method was unclear, but capacity, with agreement within 5 percent for a four-pad
the plots in the paper are consistent with synchronously tilting pad bearing in a load-between-pad configuration.
reduced coefficients. The solution was obtained using finite Nicholas and Wygant [40] presented results and design
difference techniques with overrelaxation. There was no considerations for highly loaded bearings, with specific
indication that a turbulence model was used. The paper also loads of up to 3.45 MPa (500 psi). The paper focused on
showed a drop in effective stiffness and damping due to pad pivot design, specifically steel pivots with bronze pads and
deformation, which acts like a spring in series with the oil steel pivots with steel pads. The synchronously reduced
film. stiffness and damping coefficients were used to facilitate the
Ettles [34] presented another THD analysis of tilting pad discussion on the effect of pivot stiffness on the dynamic
bearings. The synchronously reduced coefficients were cal- coefficients. The effective stiffness and damping were lowered
culated and compared to results reported by Brockwell and since the pivot acts as an additional spring in series with the
Dmochowski [35]. Ettles considered the transition region oil film. Hertzian contact theory was used to calculate the
for turbulence in the lubricating flow to be in the range effective pivot stiffness.
Several researchers have investigated transient effects in They addressed the fact that synchronous reduction was
tilting pad bearings. These studies were influenced in part typical at the time but indicated that it would lead to erro-
by a review of bearing failures presented by Conway-Jones neous stability predictions. This allowed for the possibility of
and Leopard [41]. In their review of failure modes, Conway- nonsynchronous dynamic reduction of bearing coefficients.
Jones and Leopard determined that thermal transients as However, the analysis only considered the onset of instability,
a function of oil inlet temperature and rotational speed where the real part of the rotor-bearing system eigenvalues
led to bearing seizure failures. A theoretical study of the is zero. Only the imaginary part of the eigenvalue was
phenomena was performed by Monmousseau and Fillon considered in the reduction, which is not the most general
[42]. Their TEHD analysis was based on the generalized linear solution. The most general solution includes the real
Reynolds equation described in Section 5, and transient part of the eigenvalue. Shapiro and Colsher presented the
forms of the energy and heat transfer equations. Thermal full stiffness and damping matrices for a five-pad tilting pad
growth was modeled using a plane stress assumption and bearing, resulting in 28 stiffness and 28 damping coefficients.
free boundary conditions on expansion. Finite difference Allaire et al. [46] presented a pad assembly method for
techniques were used in the solution. The pivot flexibility tilting pad bearings that explicitly included the motion of
model developed by Kirk and Reedy [43] was also used in the pads and the resulting stiffness and damping coefficients.
the model. The theoretical study predicted seizing after 53 The solution was based on perturbing the shaft or pad as
seconds of operation for a rotor was accelerated from 0 to appropriate for several applied loads and static eccentricities
10,000 rpm over 5 s, with oil lubricant inlet temperature of to develop a table of data. The overall equilibrium point
20◦ C. The time to seizure was extended or eliminated by and resulting dynamic coefficients were then found through
reducing the rate of shaft acceleration and increasing the oil linear interpolation. This method is valid for isoviscous,
inlet temperature in the analysis. laminar analyses but needs modification to account for
Transient effects were also considered by Monmousseau thermal and turbulence effects. The analysis did not account
et al. [44] in the analysis of tilting pad bearings. The TEHD for hot oil carryover, where oil exiting one pad affects the oil
analysis again included the generalized Reynolds equation, entering the next pad. The full coefficient matrix for tilting
and transient forms of the energy and heat transfer equa- pad bearings was developed which was independent of the
tions, with thermal growth included in the overall model. pad inertia and excitation frequency. Results were presented
Finite difference techniques were used in the solution. A step for a five-pad bearing with load on pad and zero-pad preload.
change in bearing specific load was modeled. Predictions Most of the plots were of the full stiffness and damping
were compared to pad babbitt temperature measurements, coefficients as a function of Sommerfeld number, but one set
with 10–15 percent difference using the TEHD model. The of synchronously reduced plots was presented.
thermal transient period was on the order of 60 s, while Parsell et al. [47] further explored the frequency effects
the mechanical transient period was on the order of one in tilting pad bearings. The authors postulated that syn-
shaft rotation at a running speed of 4,000 rpm. The authors chronously reduced bearing coefficients may be an accept-
concluded that thermal and elastic effects should be modeled able engineering approximation for rotordynamic stability,
to accurately capture bearing transient behavior. though it was mathematically incorrect. The synchronous
coefficients have since been shown to give a nonconservative
4. Nonsynchronous Bearing estimate of stability, for example, [48]. The main purpose
Model Development of the paper was to plot reduced bearing stiffness and
damping coefficients as a function of whirl frequency ratio,
The work by Lund in 1964 [20] reported synchronously which is the ratio of excitation frequency to running speed.
reduced bearing coefficients for the tilting pad journal Sommerfeld numbers of 0.1, 1, and 10 were considered for
bearing. While appropriate for unbalance response analysis five-pad load between pad configurations. Preloads of 0 and
since the unbalance forcing frequency is driven by shaft 0.3 were also considered. The frequency dependence was
rotation, the synchronous coefficients are not appropriate reduced for low Sommerfeld number and for high preload.
in general. The proper reduction method is based on For high Sommerfeld number and zero preload, the reduced
overall system excitation frequency which is in general coefficients were highly frequency dependent, and effective
nonsynchronous with shaft rotation. This distinction was stiffness and damping approached zero for whirl frequency
made clear during the presentation of Nicholas et al. [23], ratios from 0.3 to 0.5. This effect was postulated later [49]
which reported results for the stability analysis of an 11-stage to be due to light bearing load compared to running speed.
compressor. Nicholas et al. used the synchronously reduced In that case, the shaft runs nearly centered in the bearing, so
bearing coefficients to estimate the compressor stability negligible pressure forces act on the shaft.
margin. During the discussion of the paper, Lund told the Rouch [50] presented a method for modeling pivot
presenter that the use of synchronously reduced coefficients flexibility as a spring in series with the effective oil film
for stability margin was mathematically incorrect and that dynamic coefficients. The pad assembly method was used to
reduction at the rotor natural frequency was correct [24]. determine the oil film dynamic characteristics. A full matrix
This comment spurred much research into the development including the effect of pad rotations and pad translations
of nonsynchronous bearing dynamic models. was presented. Dynamic reduction was performed nonsyn-
Shapiro and Colsher [45] examined the effect of bearing chronously. Plots of effective reduced stiffness and damping
preload on the dynamic response to tilting pad bearings. as a function of excitation frequency were presented. The
pivot stiffness reduced the effective stiffness and damping by local coordinate system for each pad, with transformation to
up to 50 percent due to a spring in series with the oil film. The global coordinates as the final step. An excitation frequency
effect was most pronounced when the effective pivot stiffness dependence in reduced stiffness and damping coefficients
was the same order of magnitude as the oil film stiffness. A was demonstrated. The bearings considered had effective
stability analysis of a flexible rotor was also performed using stiffness that decreased by about 20 percent from 0.5X to 1X,
the nonsynchronous and synchronous bearing coefficients. where X is the excitation frequency corresponding to shaft
It was shown that the nonsynchronous coefficients gave a rotational speed. A difference in effective bearing coefficients
more conservative estimate of rotor stability compared to depending on the distance from the stability margin was
synchronously reduced coefficients. also demonstrated. This demonstrated that consideration of
Lund and Pedersen [51] presented an approximate both the real and imaginary parts of the eigenvalues give a
method for including pad deformations and pivot flexibility more general treatment in stability analysis, since the only
in the overall dynamic response of tilting pad bearings. An case that has a purely imaginary eigenvalue is the neutrally
isoviscous solution to (1) was obtained using a finite dif- stable solution. The neutrally stable solution is appropriate
ference method. The pad deformations were approximated for the onset of instability, but is not appropriate for stable
as the deformation of a beam under a distributed pressure or unstable systems.
load, and the pivot flexibility was obtained using Hertzian Earles et al. [55] developed a finite element solution
contact theory. The model incorporated the effect of pad for a single tilting pad including the lubricating film and
deformation by an increased effective clearance between the pad deformation effects. The lubricating flow was treated
pad and the shaft. The effective clearance was also treated as as laminar, isoviscous, and incompressible. Thermal effects
dynamic and harmonic, using the vibrational motion of the were not considered. The pad was modeled using plane strain
pad deflections to calculate the effective clearance. The pivot isoparametric finite elements, and the pressure solution for
flexibility was treated as a spring in series, and the overall the lubricant was solved simultaneously to determine the
bearing coefficients were dynamically reduced. The authors flow field and the final pad dimensions. The pad degrees of
advocated nonsynchronous reduction of bearing coefficients freedom were reduced to a single coordinate using Guyan
in general, but synchronously reduced bearing coefficients techniques; the single coordinate represented the final pad
were presented in the results. The reduction in stiffness and radius of curvature. The stiffness and damping coefficients
damping due to flexibility effects by up to 50 percent was were found through numerical perturbation of the shaft
demonstrated. and pad positions and deformations. The result for the
Branagan [52] presented a TEHD finite element solution single pad was then transformed to local coordinates, which
for fixed geometry and tilting pad journal bearings. Poly- implied usage of the results for pad assembly solutions.
nomial profiles for the thermal and viscosity solutions were The mechanical deformations were shown to lower effective
assumed in the axial and cross-film directions, leaving a 1D stiffness and damping coefficients and were within 5 percent
solution in the circumferential direction. A simultaneous, of the coefficients reported by Lund and Pedersen. A
iterative solution procedure for the Reynolds equation, the damped eigenvalue solution similar to Barrett et al. [54]
energy equation, and the deformation model was performed was employed. The single pad model was extended to a
to obtain accurate estimates of the changes in boundary full tilting pad bearing model with lubricating film and
conditions due to hot oil carryover. The full bearing coef- pivot flexibility effects [56]. The full stiffness and damping
ficients were calculated, and a damped eigenvalue analysis matrices with flexibility effects were presented. Additional
was used to reduce the full bearing coefficients to the terms to account for pivot flexibility and pad deformation
eight coefficients related to the shaft degrees of freedom. effects were incorporated into the global bearing stiffness and
The dynamic reduction was performed to improve the run damping matrices. The full coefficients were used to perform
times for subsequent rotordynamic analyses. Reduction of a damped eigenvalue analysis.
the computational expense for rotordynamic models in this White and Chan [57] presented a finite element THD
fashion is not important with modern computers, where bearing analysis with turbulence correction. Turbulence
a full eigenvalue analysis of a rotor beam model with full correction factors were based on bulk flow theory proposed
bearing coefficients has a run time of less than 10 s [53]. It by Hirs [58]. The full stiffness and damping coefficients
was demonstrated that inclusion of pad and pivot flexibility were found from the perturbed Reynold’s equation and the
effects could reduce the calculated stiffness and damping reduction method of Parsell et al. [47] was used for reduction
coefficients by up to 50 percent. to the eight stiffness and damping coefficients. The Parsell et
Barrett et al. [54] provided an extension to reduction of al. method only uses the imaginary part of the eigenvalue
full tilting pad bearing coefficients to the eight frequency- to perform the dynamic reduction. The method is correct
dependent stiffness and coefficients. In their analysis, Barrett for forced response analyses and analyses to determine the
et al. included the real part of the eigenvalue to allow for onset of instability, but is not correct for a general free
general damped analyses. The analysis considered bearings response analysis with damped eigenvalues. White and Chan
in a load between pad and treated pad inertia as negligible. compared the effective stiffness and damping coefficients
Single pad solutions similar to Lund [20] were developed, for synchronous and half-whirl reduction frequencies and
and the pad assembly method was applied. Pad rotations showed a reduction in effective damping for half-frequency
were treated as equivalent translations due to small angle per- whirl of up to 20 percent. They also showed that the effective
turbations. The dynamic reduction was performed in the pad damping was reduced for off-center pivots. Nondimensional
dynamic coefficients were plotted as a function of Sommer- The following discussion does not consider mechanical
feld number. deformations, which are geometry specific. A very brief
Brockett and Barrett [59] presented a tilting pad bearing discussion of the key equations follows. A comprehensive
dynamic reduction method suitable for transfer matrix derivation of the equations is provided in [63]. While
analyses. The reduction resulted in a second-order transfer more recent work has further refined the TEHD analysis,
function with a fourth-order residual frequency dependent particularly with the inclusion of a 3D energy equation [67],
stiffness. The reduction admitted damped eigenvalue solu- these refinements have only been applied to fixed geometry
tions. Results of the bearing representation in a transfer bearing analyses.
matrix model were compared to a finite element solution Reynolds’ equation, (1), is the fundamental equation for
for a flexible rotor with the full bearing coefficients modeled lubricating flows assuming a laminar, isoviscous lubricant.
explicitly. Agreement within 1 percent of the system eigen- The generalized Reynolds equation results from the fluid
values was obtained, but the transfer matrix method missed continuity and momentum equations, with the assumption
highly damped modes. that the pressure profile is constant across the lubricating
Kim et al. [60] presented a nonsynchronous reduction of film. The generalized form includes convective inertia effects
tilting pad bearing coefficients, including pad deformation through an eddy-viscosity model and allows for cross-film
effects. The pad deformation model reduction was achieved variations in viscosity. The formulation of the generalized
through modal representation of pad movement, with modal Reynolds equation in terms of pad local coordinates is [63–
truncation. Further dynamic reduction of the oil film 66]
and mechanical flexibility effects was performed nonsyn-
chronously to obtain the eight frequency-dependent stiffness ∂
∂p ∂ 3
∂p
and damping coefficients. The analysis accounted for thermal h3 Γ η, z, Re∗ + h Γ η, z, Re∗
∂η ∂η ∂z ∂z
effects on lubricant viscosity as well as pivot flexibility, pad (5)
rotations, and pad deformations. A turbulence model was ∗

∂h
= −UG η, z, Re .
not included. Variable viscosity due to temperature changes ∂η
was accounted for. The theoretical results were compared to
results reported by Brockwell et al. [37], and the analysis In (5), Γ and G represent generalized local effective viscosity
predicted the measured drop in stiffness and damping due functions with turbulence effects included, and U represents
to mechanical deformations. The theory agreed with the the motion of the journal relative to the bearing pad.
Brockwell et al. data within 10 percent. Theoretical results Equation (5) has been modified from [63] to include the
were also compared to experimental results reported by effect of reduced Reynolds number Re∗ = (ρωh2 )/μ, where ρ
Fillon et al. [61]. The effects of the TEHD model on the is the lubricant density and ω is the rotational speed. This is
coefficients was shown in separate plots with synchronously to account for turbulence effects, especially those due to the
reduced coefficients. A stability estimate for an eight-stage low viscosity of the lubricating fluid for some process fluid
gas reinjection compressor was performed and compared lubricated bearings. For example, the viscosity of water is two
to results presented by Wilson and Barrett [62]. It was orders of magnitude less than the viscosity of oil [68].
shown that use of the frequency-dependent stiffness and The effective cross-film viscosities Γ and G, as a function
damping coefficients in the rotor bearing model resulted in a of the journal radial and axial positions, are given by [63]
lower stability margin compared to a synchronously reduced
bearing model, which agreed with the Wilson and Barrett 1

results. Γ η, z, Re∗ = χ2 η, ξ, z, Re∗

0
5. Thermoelastohydrodynamic Tilting Pad

χ2 η, 1, z, Re∗
Bearing Lubrication Theory −

χ1 η, ξ, z, Re∗ dξ,
χ1 η, 1, z, Re∗
Tilting pad bearing lubrication theory has evolved, from 1

1
fixed geometry isoviscous analytical solutions, to advanced G η, z, Re∗ =

χ1 η, ξ, z, Re∗ dξ,
χ1 η, 1, z, Re∗ 0
finite element solutions including hydrodynamic, energy,
and deformation effects. The modifications to (1) and ξ

1
additional equations to model the energy balance and χ1 η, ξ, z =
dξ ,
0 μe η, ξ , z, Re∗
turbulence modeling are summarized in Sections 5.1–5.3.
ξ
∗

ξ
χ2 η, ξ, z, Re =
dξ ,
5.1. Generalized Reynolds Equation. Modern tilting pad 0 μe η, ξ , z, Re∗
bearing lubrication theory is based on thermoelastohy-
drodynamic models that include equations describing the (6)
hydrodynamic flows, heat transfer and shear heating, and
mechanical deformations [63–66]. Whiles these solutions where χi represent intermediate viscosity functions, μe
evaluate temperature effects in the lubricant film and pad represents the effective turbulent viscosity, ξ represents the
deformations, they do not involve the advanced elasto- local pad squeeze direction, and ξ represents the dummy
hydrodynamic solutions found for ball or roller bearings. variable of integration.
The flow profile in the bearing is treated as a combination developed by Elrod and Ng [69], with a modification by
of Couette and Poiseuille flow, which is expressed as: Suganami and Szeri [1] to account for transition flow. The
turbulence is modeled using an eddy viscosity law. The

∂p
u η, ξ, z = χ2 η, ξ, z, Re∗ effective viscosity is found by.

∂η

m
μe η, ξ, z = μ 1 + , (10)
U ν
+
χ1 η, h, z, Re∗ (7) where ν is the kinematic viscosity. The eddy viscosity m is

calculated from
∂p χ2 η, h, z, Re∗
− ∗

m ξ+
∂η χ1 η, h, z, Re =κ ξ +
− δl+ tanh . (11)
ν δl+
∗

× χ1 η, ξ, z, Re , In (11), based on turbulent boundary layer theory, κ = 0.4

∂p
and δl+ = 10.7. These constants were found empirically and
w η, ξ, z, Re∗ = χ2 η, ξ, z, Re∗ are reported in Elrod and Ng [69]. The nondimensional
∂z

distance from the wall ξ + is defined as:
∂p χ2 η, h, z, Re∗

− ∗

χ1 η, ξ, z, Re∗ , ξ |τ |
∂z χ1 η, h, z, Re ξ+ = . (12)
ν ρ
(8)
And the local shear stress in the lubricant in (12) is found
where u represents local velocity in the sliding direction, from:
w represents local velocity in the axial direction, and U
2 2
represents the surface velocity of the shaft relative to the pad. ∂u ∂u

The solutions to (5)–(8) result in the pressure field τ = μe + . (13)
∂ξ ∂ξ
developed in the lubricating film as a function of radial
and axial position. The pressure across the film is assumed The inclusion of the effective viscosity μe is implicit in
constant since it is small compared to the radial and axial (10)–(13), so the shear stress distribution and the effective
length scales. Integrating the pressure field over the area of local viscosity in the turbulent regime is typically found
the bearing surfaces yields the net forces in the bearing. iteratively, such as in the finite element code developed by
He [63]. By including the eddy-viscosity model in the TEHD
code, convective inertia effects are approximated in the
5.2. Energy Equation. The viscosity of many lubricants is a
solution. However, temporal inertia effects are not currently
strong function of temperature. The developed pressures in
considered in the analysis.
hydrodynamic bearings are not large enough to significantly
Equations (10)–(13) are valid for fully developed tur-
affect the viscosity. To account for temperature effects, the 2D
bulent flows [69]. To account for low levels of turbulence
energy equation, including shear heating terms, is considered
and transitional flows, Suganami and Szeri proposed an
in the model presented in [63–66].
additional factor γ, which modifies (10) as [1].

∂T ∂T
m
ρC p u +v μe η, ξ, z = μ 1 + γ . (14)
∂η ∂ξ ν
⎡ 2 ⎤ The factor γ is dependent on the maximum Reynolds’
2
∂ ∂T ∂ ∂T ⎣ ∂u ∂w ⎦ number in the lubricating flow and is defined as.
= k + ke + μe + ,
∂η ∂η ∂ξ ∂η ∂ξ ∂ξ ⎧
⎪
⎪0, Remax < ReL ,
(9) ⎪
⎪
⎪
⎨ 1/8
where v is the fluid velocity in the squeeze direction, C p is the γ = ⎪1 − ReH − Remax , ReL ≤ Remax ≤ ReH ,
⎪
⎪ ReH − ReL
specific heat of the lubricant, T is the lubricant temperature, ⎪
⎪
⎩ 1, ReH < Remax ,
k is the heat conductivity of the lubricant, and ke is the
effective heat conductivity of the lubricant corrected for (15)
turbulence. The relative importance of the shear heating where Remax = (ρuh)/μ is the maximum local Reynolds’
term, μe [(∂u/∂ξ)2 + (∂w/∂ξ)2 ], in (9) is dependent on the number in the lubricant and u is the local fluid velocity, ReL is
lubricant considered. For oil-lubricated bearings operated at the critical Reynolds’ number for the onset of transition flow
high speed, shear heating effects are significant and viscosity from laminar flow and ReH is the critical Reynolds’ number
variation due to temperature must be considered in the for the onset of turbulence. There is not a consensus in the
analysis. For many process fluid lubricants such as water, the literature on the critical Reynolds’ numbers ReL and ReH ,
shear heating effects can often be neglected due to the low corresponding to the onset of transition flow and turbulent
lubricant viscosity, allowing for isoviscous analyses. flow, respectively [63]. Proposed values of Reynolds’ number
for the onset of transition flow range from about 500–1000
5.3. TEHD Solution Turbulence Modeling. The turbulence and range from about 800–1500 for the onset of turbulent
model implemented by He [63] was based on models flow [1–5].
5.4. Perturbed Reynolds Equation. Once the pressure profile contribution of turbulent stresses. In nondimensional form,
solution is found, the generalized Reynolds equation is per- the proposed model was
turbed. The first-order perturbation results in the equivalent 2
∂ 3 ∂ p R ∂ 3 ∂ p
stiffness and damping coefficients, ki j and ci j , respectively, h Gη + h Gz
which have the general form [63–66]. ∂η ∂η L ∂z ∂z
(18)
1 ∂h cr 2 ∗
∂f ∂f = − Re ∇ hI .
ki j = − , ci j = − . (16) 2 ∂η R
∂ui j ∂u̇i j
Implicit in (18) is the shear stress at the boundaries. The
The specific stiffness and damping coefficients are defined by term multiplied by Re∗ in (18) is the contribution of fluid
He [63]. inertia in terms of turbulent stresses. In (18), Gη and Gz are
When (5)–(16) are considered, it is apparent that the parameters dependent on the average Reynolds number of
turbulence model chosen significantly affects the predicted the flow Re, R is the radius of the bearing, L is the axial length
dynamic coefficients for the tilting pad bearing. of the bearing, and cr is the radial bearing clearance. The
gradient operator ∇ is defined as
6. Averaged Flow Approaches =i

∂ R ∂
∇ +k (19)
∂η L ∂z
Another approach to the lubrication problem is the averaged
and the averaged flow profiles I have the following defini-
flow method, where the properties of the lubricant across
tions:
the film are averaged. First proposed by Constantinescu [70]
in terms of turbulent mixing length theory, the model was I∗ = iGη Iη + kGz Iz ,
refined in several follow-on papers [29, 71–73].
1 ∂Iηη ∂Iηz
Other authors have used predominantly empirical bulk Iη = 2
+ ,
flow approaches, including Hirs [58] and San Andrés [74– V cr ∂η ∂z
76]. In all of the averaged-flow formulations, the temporal
1 ∂Iηz ∂Izz
inertia term ρ∂u/∂t is retained from the Navier-Stokes Iz = + ,
equations and is incorporated into the analysis. V 2 cr ∂η ∂z
The averaged flow approaches rely on averaging of the h (20)
2
fluid properties across the film, including average velocity Iηη = u dξ = αUm2 h + βV 2 h − γUm V h,
0
and viscosity. For example, the formulations typically con-
h
sider an average fluid velocity of the form:
Iηz = uwdξ = α Um Wm h − γ Wm V h,
0
h
1 h
ud y = uavg . (17)
h 0 Izz = w2 dξ = α Wm2 h.
0
Section 6.1 considers the mixing length theory approach. In (20), Um represents the mean fluid velocity in the slider
Section 6.2 considers the empirical approaches. direction, Wm represents the mean fluid velocity in the axial
direction, u, v, and w represent local fluid velocities, and the
6.1. Mixing Length Theory. Constantinescu published a coefficients α, β, γ, and δ and their primes are dependent on
series of papers [29, 70–73] detailing the application of the Reynolds number of the flow. The assumed flow profile
mixing length theory to the turbulent lubrication problem. used in the definitions of the I terms is a parabolic profile
His focus was on high Reynolds number flows, with Re ≥ which is interpreted as a Poiseuille mean velocity.
1000. The Reynolds number is defined as Re = (ρωDh)/μ, For laminar flows, Gη = Gz = 1/12, α = α = α = 6/5,
where h is the local film thickness. The turbulence was β = δ = 2/15, and γ = 1/5. For turbulent flows (Re > 5000)
modeled by treating all variables associated with the flow as [73], the coefficients are functions of the mean Reynolds
a mean value, defined by (17) plus a fluctuation. Turbulent number of the flow:
stress terms resulted from the simplification of the Navier- 1
= 12 + 0.0136 Re0.9 ,
Stokes equations. To model these stresses, Prandtl’s mixing Gη
length theory was employed, which produced results that 1
agreed with experimental data on Poiseuille flow. Analytical = 12 + 0.0043 Re0.96 ,
Gz
solutions for average film properties were possible with this
assumption. This can be extended for other average film α = α = α = 1,
(21)
properties, such as viscosity, in THD analyses similar to those 0.885
proposed by San Andrés [74–76]. β= ,
Re0.367
The series of papers culminated in a journal bear-
γ = γ = 0,
ing lubrication theory published by Constantinescu and
Galetuse in 1982 [73]. In this paper, a modification to 1.95
δ= .
the Reynolds equation was proposed that included the Re0.43
The Constantinescu approach allows for rapid solutions to transition flow regime because experimental data is fitted to
the lubrication problem, but there are some drawbacks. The the model, assuming such data is available for that bearing
system of equations relies on a minimum of six empirical configuration and flow condition. The transition region is
coefficients to characterize turbulence, which requires exten- where many modern oil-lubricated bearings operate. The
sive experimental data to validate. The approach also relies key drawback is that the method totally relies on empirical
on an assumption that the average of the product of flow data. The types of experiments that would be required to
profiles is the product of the averages, that is, obtain a complete set of empirical coefficients were alluded
h to by Hirs [78]. However, it is not clear how extensive the
u2 d y = hU 2 , experimental support would have to be, especially when
0
factors such as shaft eccentricity ratio, pivot offset, thermal
h
effects and bearing preload are considered. Thermal effects
uwd y = hUW, (22)
make the estimation of an average Reynolds number difficult.
0
h Pivot offset and nonzero bearing preload will alter the
w2 d y = hW 2 shape of the converging wedge. Hirs indicated that high-
0 eccentricity bearings gave less accurate results [58]. The
which cannot be justified as noted by Szeri [77]. Constanti- combination of moderate operating eccentricity and high
nescu acknowledged that (22) was at best only approximately bearing preload may mimic a zero-preload, high-eccentricity
correct [71]. bearing in terms of shape of the converging wedge.
6.2. Empirical Approaches. Hirs [58] proposed a model 6.3. Comparison of Approaches. Taylor and Dowson [79]
that was predominantly based on experimentally measured directly compared the mixing length approach of Constan-
bulk-flow properties relative to a surface or wall and the tinescu, the eddy-viscosity model of Ng, Pan, and Elrod, and
corresponding shear stresses at the boundaries based on a set the empirical bulk-flow approach of Hirs. When comparing
of flow conditions. The method does not consider the shape the predicted factors Gη , all three methods gave close
of internal flow profiles or fluctuations within the lubricating agreement for Reynolds number greater than 2000. The
film. Constantinescu model overpredicted Gη by up to 50 percent
By solely considering the average flow properties and when compared to Elrod and Ng and Hirs. All three models
the boundary conditions, Hirs developed a set of pressure deviated from each other in for 1000 ≤ Re ≤ 2000. Based on
equations for sliding surfaces as these results, Taylor and Dowson concluded that the eddy-
1+m0 viscosity model proposed by Elrod and Ng [69] was more
h2 ∂p μ accurate than the model proposed by Constantinescu [72].
−
μU ∂η ρUh The transition region from laminar flow to turbulent
(1+m0 )/2 flow presents challenges to the both the eddy-viscosity model
1
= n0 Uη Uη2 + Uz2 and the bulk flow approaches. Suganami and Szeri [1] were
2
able to address this challenge in part with an additional
2 (1+m0 )/2
scaling factor in the effective viscosity obtained from the
+ Uη − 1 Uη − 1 + Uz2 ,
eddy-viscosity model, represented by (11). The Suganami
1+m0 and Szeri scaling factor was able to model temperature rise
h2 ∂p μ in a bearing that was run in the transition flow region more
−
μU ∂z ρUh accurately than either the laminar or turbulent models. This
(1+m0 )/2 scaling factor, although empirical, reduces the discrepancy
1
= n0 Uz Uη2 + Uz2 between the Elrod and Ng model and the Hirs model in the
2 transition region.
2 (1+m0 )/2 Bouard et al. [80] compared three turbulence models
+Uz Uη − 1 + Uz2 , using a finite difference solution to the generalized Reynolds
(23) equation, the energy equation, and the heat transfer equa-
tion. The three models compared were the Ng and Pan
where Uη and Uz represent dimensionless mean flow veloc- model, [22], the Elrod and Ng model [69], and the Constan-
ities, and the constants n0 and m0 are found empirically tinescu model [29]. The comparison was performed within a
from representative flows. In a subsequent paper, Hirs [78] common finite difference framework, which is distinct from
delineated the various flow regimes requiring experimental the bulk-flow approach used by Constantinescu. The three
data to determine these constants and summarized results models were compared to experiments reported by Taniguchi
for experiments that were already available in the literature. et al. [4]. All three models overpredicted temperatures in
The minimum Reynolds number for any of the empirical the laminar flow regime, which was attributed to poor
coefficients was 1000. characterization of the experimental boundary conditions.
In terms of friction factors, the Hirs approach can All three models matched the experiment within 2 percent
generally produce more accurate results for the lubrication at operating speeds above 3,600 rpm. All three models gave
θ
cξθ kξθ
cξξ ψ cθθ ψ
cθξ kθξ
cξξ
y kξξ
η y kθθ
kξξ
cηθ η
x cθη
x
ξ kηθ Jp
ξ kθη
Ms
ω cηη ω cηη
kηη kηη
Figure 4: Free body diagram, dhaft translational degrees of Figure 5: Free body diagram, pad rotational degrees of freedom,
freedom, and rigid pivots. and rigid pivots.
similar predictions of power loss. The Constantinescu turbu- When a force balance is considered on the free body
lence model gave predictions closest to measurement for film diagrams, the resulting equations of motion can be expressed
thickness and babbitt temperature. The authors concluded in matrix form as [46, 81]
that all three models gave similar predictions and that ⎡ ⎤ ⎡ ⎤
Constantinescu should be used because the computational Ms 0 0 ⎧ ⎪
⎫
⎪ cηη cηξ cηθ ⎧ ⎫
⎢ η̈
⎥⎨ ⎬ ⎢ ⎥⎪⎨η̇⎪
⎬
⎢
0 Ms 0 ⎥ ⎢ ⎥
run times were shorter. The TEHD analysis performed by
⎢ ⎥ ξ¨ + ⎢cξη cξξ cξθ ⎥ ξ˙
He gave better predictions in the laminar regime [63], using ⎣ ⎦⎪⎩ ⎪ ⎭ ⎣ ⎦⎪⎩ ⎭⎪
the Elrod and Ng model with the modification proposed by 0 0 J p θ̈ cθη cθξ cθθ θ̇
Suganami and Szeri [1] in the transition region. ⎡ ⎤ (24)
kηη kηξ kηθ ⎧ ⎫ ⎧ ⎫
⎢ ⎥⎪ ⎬ ⎪
⎨η ⎪ ⎨ fη ⎪
⎬
⎢ ⎥
+ ⎢kξη kξξ kξθ ⎥⎪ξ ⎪ = ⎪ fξ ⎪,
7. Review of Tilting Pad Bearing ⎣ ⎦⎩ ⎭ ⎩ ⎭
θ 0
Dynamic Models kθη kθξ kθθ
The tilting pad dynamics developed from various TEHD where ki j represents lubricant equivalent stiffness, ci j repre-
models are based on explicit modeling of the motion of sents lubricant equivalent damping, Ms represents the mass
the pads. The modeling procedure is summarized in the of the shaft, J p represents the mass moment of inertia of
following section. The development of the bearing model the pad about the pivot, η and ξ represent the orthogonal
reduced to the shaft degrees of freedom and an experimen- directions in the pad local coordinate system, x and y
tally identified two-degree-of-freedom bearing model are represent the global coordinate system, and θ represents
also summarized. rotations of the bearing pad about the pivot. For brevity, (24)
can be rewritten in matrix form as
7.1. Single Pad Bearing Dynamics. The lubricating film is Mp v̈p + Cp v̇p + Kp vp = fp . (25)
typically represented with stiffness and damping coefficients
in linear analyses. To illustrate this concept, two free-body Primed matrices refer to the pad local (η, ξ) coordinate
diagrams are provided. The first, Figure 4, shows rigid shaft systems, and unprimed matrices refer to the fixed (x, y)
interactions with a single tilting pad through the oil film. The coordinate system. The vector vp represents the rotor and
springs and dampers in Figure 4 are shown schematically for pad motion in pad local coordinates. The tilting pad bearing
clarity of the figure. The actual reactions are fluid structure dynamics are typically transformed to the global shaft
interaction forces between the shaft, the lubricating film, and coordinate (x, y) system for the purposes of rotordynamic
the pad [77]. analyses. The coordinate transformation from local to global
The second free-body diagram, Figure 5, shows the coordinates for the local pad degrees of freedom is given by
linearized fluid-structure interactions between the pad and [46]:
the shaft, and between the pad and ground. The free body
⎡ ⎤
diagrams are shown separately because the linearized stiff- ⎧ ⎫ − sin ψ − cos ψ 0 ⎧ ⎫
⎪
⎨x ⎪
⎬ ⎢ ⎥⎨⎪η ⎪
⎬
ness and damping coefficients are in general non-selfadjoint. ⎢ ⎥
y = ⎢ cos ψ − sin ψ 0⎥ ξ , (26)
In Figures 4 and 5, a single pad is shown for clarity of the ⎪
⎩θ ⎪
⎭ ⎣ ⎦⎪ ⎪
⎩θ ⎭
figures. A typical bearing would have four or five pads. 0 0 1
where ψ is the angle between the x-axis and the pivot and the coordinate transformations in Appendix A. Equation
location. In matrix form, (26) can be rewritten as up = QT vp , (28) can also be written in matrix notation as
or alternatively Qup = vp , where Q represents the single-
pad coordinate transformation matrix and up represents
the vector of single pad dynamics expressed in global Mü + Cu̇ + Ku = f. (29)
coordinates, or up = [x y θ]T . The total transformation
from local to global coordinates on a per-pad basis is then
given by 7.3. Tilting Pad Bearing Model Reduction. Equation (28) has
not been traditionally used to describe tilting pad journal
fp = QT fp = QT Mp Qüp + QT Cp Qu̇p + QT Kp Qup bearing behavior in rotordynamic analyses. It is typically
(27) reduced dynamically to the shaft degrees of freedom asso-
= Mp üp + Cp u̇p + Kp up . ciated with the bearing, and the pad degrees of freedom are
not explicitly considered. Historically, there are a few reasons
7.2. Assembled Tilting Pad Equation of Motion. Once the for this choice. Older rotordynamics analyses were based on
individual pad equations of motion are transformed to the transfer matrix method, originally described separately
global coordinates, the overall equations of motion can be by Myklestad [84] and Prohl [85]. The method, based on
assembled [46]. The fundamental equation of motion with beam theory, can be modified to include a discrete stiffness
no fluid temporal inertia effects, rigid pads, and rigid pivots from a bearing as long as it is related to the appropriate
expressed in terms of shaft degrees of freedom and pad beam degree of freedom. The transfer matrix method is not
rotations for an N p pad bearing is [46, 82, 83] capable of admitting a full-coefficient representation of a
⎡ ⎤ tilting pad bearing unless it is transformed into an equivalent
Ms 0 0 0 ··· 0 ⎧ ⎪ ẍ ⎪
⎫ transfer function. This exact transfer function was developed
⎢ ⎪ ⎪
⎢ 0 Ms 0 0 ··· 0 ⎥ ⎪
⎪
⎥⎪ ÿ ⎪⎪
⎪ by Brockett and Barrett [59]. However, it was published in
⎢ ⎥⎪
⎪ ⎪
⎪
⎢ ⎥⎪
⎪ ⎪
⎪ 1993, when computer power was reaching a point where
⎢ 0 0 J1 0 · · · 0 ⎥⎪ ⎨ θ̈1 ⎪ ⎬
⎢ ⎥ desktop finite element analyses of rotors were feasible. It was
⎢ .. .. ⎥⎪ θ̈ ⎪
⎢ 0 0 0 J2 ⎥
. . ⎥⎪⎪ 2⎪ ⎪ also published after the use of less sophisticated reduced-
⎢
⎢ ⎪ . ⎪
⎥⎪ ⎪
⎢ .. .. .. .. .. ⎥⎪
⎪ . ⎪
. ⎪ order bearing models had become the industry standard.
⎣ . . . . . 0 ⎦⎪⎪
⎪
⎩
⎪
⎪
⎪
⎭ Modern rotordynamic analysis packages using finite element
0 0 0 · · · 0 JN p θ̈N p
formulations such as [53] are capable of using the full
⎡ ⎤ coefficient representation.
cxx cxy cxθ1 cxθ2 · · · cxθN p ⎧ ⎪ ẋ ⎪
⎫
⎪
⎢ c ⎥⎪
⎪ ⎪ There were also some fundamental misunderstandings of
⎢ yx cy y c yθ1 c yθ2 · · · c yθN p ⎥⎪ ⎪
⎢ ⎪ ẏ ⎪
⎥⎪
⎪
⎪
⎪
⎪
⎪ the tilting pad journal bearing results originally presented by
⎢ ⎥⎪
⎢ cθ 1 x cθ 1 y cθ 1 θ 1 0 · · · 0 ⎥⎨ θ̇1 ⎪
⎪ ⎪
⎬ Lund in 1964 [20]. The design curves presented by Lund were
⎢ ⎥
+⎢
⎢ cθ x .. . ⎥
⎪ θ̇2 ⎪ reduced synchronously, or using the shaft running speed as
⎢ 2 cθ 2 y 0 cθ 2 θ 2 . . ⎥
.
⎥⎪
⎪ ⎪
⎪
⎢ . .. .. ⎥⎪
⎪
⎪
⎪ .. ⎪
⎪
⎪
⎪
the reduction frequency. While not intended to be used in
⎢ . .. .. ⎥⎪ ⎪
⎣ . . . . . 0 ⎦⎪ ⎪
⎩
. ⎪
⎪
⎭
general [51], the synchronous coefficients as an excitation-
cθN p x cθN p y 0 · · · 0 cθ N p θ N p θ̇Np frequency-independent representation of tilting pad bearing
dynamics became the industry standard. Manufacturers
⎡ ⎤
kxx kxy kxθ1 kxθ2 · · · kxθN p ⎧ ⎪ x ⎪
⎫ developed design tools based on the synchronously reduced
⎢ k ⎪ ⎪
⎢ yx ky y k yθ1 k yθ2 · · · k yθN p ⎥
⎥⎪⎪
⎪ y ⎪
⎪
⎪ coefficients for rotating machinery. This has been encoded
⎢ ⎥⎪⎪ ⎪
⎪
⎢ ⎥⎪⎪ ⎪
⎪ in industry standards such as API 617 for centrifugal
⎢ kθ1 x kθ1 y kθ1 θ1 0 · · · 0 ⎥⎪ ⎨ θ1 ⎪ ⎬
⎢ ⎥ compressors [86]. However, future editions of the API
+⎢ .. .. ⎥ θ
⎥⎪
⎢ kθ x 2 ⎪ standards will reflect nonsynchronous tilting pad bearing
⎢ 2 kθ2 y 0 kθ2 θ2 . . ⎥⎪⎪
⎪
⎪
⎪
⎪
⎪
⎢ .
⎢ . .. .. .. .. ⎥
⎥⎪⎪
⎪ .. ⎪
. ⎪
⎪
⎪
coefficients [87].
⎣ . . . . . 0 ⎦ ⎪
⎪ ⎪
⎪ For the purposes of this discussion, (28) will be reduced
⎩
kθN p x kθN p y 0 · · · 0 kθ θ θN p ⎭ to the shaft degrees of freedom. The shaft degree of freedom
Np Np
⎧ ⎫ approach was also heavily influenced by the performance

⎪
⎪ fx ⎪
⎪ of fixed geometry fluid film bearings, which do not have
⎪
⎪f ⎪ ⎪
⎪
⎪ ⎪
⎪
⎪
⎪
y ⎪ the additional degrees of freedom associated with the pads.
⎪ ⎪
⎪ ⎪
⎨0⎪ ⎬ Equation (28) is partitioned into shaft and pad degrees of
= . freedom. It can be rewritten in block matrix format as
⎪
⎪ 0⎪ ⎪
⎪
⎪ ⎪
⎪
⎪
⎪ .. ⎪
⎪
⎪
⎪ . ⎪
⎪
⎪
⎪ ⎪ ⎡ ⎤ ⎡ ⎤ ⎡ ⎤
⎩ ⎪ ⎭
0 Ms 0 ü Cuu Cuθ u̇s Kuu Kuθ us
⎣ ⎦ s
+⎣ ⎦ +⎣ ⎦
0 Jp θ̈ Cθu Cθθ θ̇ Kθu Kθθ θ
(28)

Equation (28) is described explicitly for a five pad tilting pad fs
= ,
bearing in Appendix A. The stiffness and damping terms in 0
(28) are also described in terms of the local pad contributions (30)

Np
where Ms is the diagonal matrix of shaft masses, Jp is the % sc yθi + k yθi scθi x + kθi x
diagonal matrix of pad inertias, Cuu , Cuθ , Cθu , Cθθ represent A yx = , (39)
i=1
s2 J pi + scθi θi + kθi θi
the damping submatrices, Kuu , Kuθ , Kθu , Kθθ represent
the stiffness submatrices, us represents the shaft translation
Np
% sc yθi + k yθi scθi y + kθi y
degrees of freedom, θ represents the pad rotation degrees of Ay y = . (40)
freedom, and fs represents externally applied forces to the i=1
shaft. By expanding (30), the resulting tilting pad bearing
model equations take the form:
For a forced response analysis, where s = jΩ, the reduced
Ms üs + Cuu u̇s + Cuθ θ̇ + Kuu us + Kuθ θ = fs , direct horizontal stiffness as a function of excitation fre-
(31) quency is then
Jp θ̈ + Cθu u̇s + Cθθ θ̇ + Kθu us + Kθθ θ = 0.
Dynamic reduction is performed in the frequency ⎡N

⎤
domain by assuming a solution of the form us = Us est , θ = %p kxθi + jΩcxθi kθi x + jΩcθi x
kxx (Ω) = kxx − Re⎣ ⎦,
Θest , fs = Fs est . The damped excitation frequency s = p + jq i=1
kθi θi − Ω2 J pi + jΩcθi θi
is in general nonsynchronous except in the case of unbalance (41)
response. Use of the Laplace transform allows for a general
damped solution to the reduction problem [47, 52, 54, 63].
By substituting the assumed solution into (31), the following where Re is the real part, and the reduced direct horizontal
equations in the frequency domain are found damping as a function of excitation frequency is
(sCuu + Kuu )Us + (sCuθ + Kuθ )Θ = Fs − s2 Ms Us , (32)
⎡ Np

⎤
1 % kxθi + jΩcxθi kθi x + jΩcθi x
cxx (Ω) = Im⎣ jΩcxx − ⎦,
(sCθu + Kθu )Us + s2 Jp + sCθθ + Kθθ Θ = 0. (33) Ω kθi θi − Ω2 J pi + jΩcθi θi
i=1
(42)
Next equation (33) is solved in terms of the pad rotations Θ
as
−1 where Im denotes the imaginary part. The reduced stiffness
Θ = − s2 Jp + sCθθ + Kθθ (sCθu + Kθu )Us . (34)
and damping terms k xy , k yx , k y y , cxy , c yx , c y y are found
similarly to (41), (42). Returning to the time domain via
Then by back substitution of (34) into (32), the bearing
an inverse Fourier transform, the reduced model of bearing
coefficients are expressed in terms of the shaft degrees of
dynamics then becomes:
freedom Us as
[sCuu + Kuu − (sCuθ + Kuθ ) ⎡ ⎤ ⎡ ⎤

Ms 0 cxx (Ω) cxy (Ω) ẋ
−1 ⎣ ⎦ ẍ + ⎣ ⎦
× s2 Jp + sCθθ + Kθθ (sCθu + Kθu ) Us = Fs − s2 Ms Us . 0 Ms ÿ c yx (Ω) c y y (Ω) ẏ
(35) ⎡ ⎤ (43)
kxx (Ω) k xy (Ω)
+⎣ ⎦ x = fx .
Using the individual stiffness and damping coefficients from k yx (Ω) k y y (Ω) y fy
the full tilting pad representation, (35) can be written out in
detail as
⎡ ⎤
scxx + kxx − Axx scxy + kxy − Axy For a free response analysis where fx = f y = 0, the effective
⎣ ⎦ X stiffness and damping can be found numerically in the form:
sc yx + k yx − A yx sc y y + k y y − A y y Y

(36)
Fx − s2 Ms X ⎛ Np

⎞
= . % scxθi + kxθi scθi x + kθi x ⎠
F y − s2 Ms Y k xx (s) = Re⎝kxx + scxx − ,
i=1
The four A coefficients are defined for rigid pivots as ⎡ ⎛
Np

1
% scxθi + kxθi scθi x + kθi x cxx (s) = Im⎣ ⎝kxx + scxx
Axx = , (37) s
i=1
Np

⎞⎤
Np
% scxθi + kxθi scθi x + kθi x ⎠⎦
% scxθi + kxθi scθi y + kθi y − 2
.
Axy = , (38) i=1
s J pi + scθi θi + kθi θi
i=1
(44)
If the perturbation frequency s = jΩ is taken to be the The real and imaginary parts of the complex impedance
rotational speed of the machine, then (41), (42) represent functions zi j are then plotted as a function of frequency. The
the direct horizontal synchronously reduced stiffness and method is capable of discerning frequency dependence in
damping coefficients. The synchronously reduced k xy , k yx , both real and imaginary parts of the complex impedances,
k y y , cxy , c yx , c y y are found similarly. If the eigenvalue s = within the limits imposed by measurement uncertainty and
p+ jq is treated as a general perturbation frequency not equal repeatability. The subscripts i, j represent the appropriate
to the machine rotational speed, then the coefficients are rotor degree of freedom. Power spectral density functions are
the nonsynchronously reduced coefficients at the excitation used to reduce the effects of noise on the measurements. For
frequency of interest. The resulting coefficient matrices are the TPJB bearings reported in [92–94], the trends in the data
non-self-adjoint. The destabilizing tangential forces due to resulted in an adequate (r 2 ≥ 0.95) model described by
fluid structure interactions are represented as cross-coupled

stiffnesses kxy and k yx . However, these cross-coupled stiffness Re Zi j = k*i j − Ω2 m
* ij; Im Zi j = Ωc*i j . (46)
terms are generally 3 orders of magnitude less than the
direct stiffness terms in tilting pad bearings and typically The complex impedance of the system then takes the form:
neglected. In the frequency domain, the effective tangential
forces due to damping, proportional to Ωcxx , Ωc y y , are also Zi j = k*i j − Ω2 m
* i j + jΩc*i j . (47)
much greater than the cross-coupled stiffness terms, which
indicates that the destabilizing forces are small and do not Then, by substituting (47) into (45), and performing an
adversely affect bearing dynamic performance. inverse Fourier transform to return (45) to the time domain,
The reduced-order model with pad dynamics considered the resulting model for TPJB behavior is given by
implicitly, (43), is applicable to rigid pivot bearings. Repre- ⎡ ⎤ ⎡ ⎤
sentation of pivot flexibility in tilting pad bearings requires * xx
M+m * xy
m c*xx c*xy ẋ
consideration of additional degrees of freedom. Treatments ⎣ ⎦ ẍ + ⎣ ⎦
* yx
m * yy
M+m ÿ c*yx c*y y ẏ
of the pivot flexibility case have been addressed by several
authors, including [43, 50]. ⎡ ⎤ (48)
k*xx k*xy x fx
+ ⎣ ⎦ = ,
7.4. Reduced Order Nonsynchronous Bearing Models. An k* k* y f y
yx yy
alternative experimental approach to characterizing TPJB
behavior is based on an experimentally identified model in where the m * i j represent the identified lubricant mass coeffi-
the frequency domain. The experimentally derived model is cients, the c*i j represent the identified damping coefficients,
based on measurement of force inputs and bearing housing
and the k*i j represent the identified stiffness coefficients.
outputs. The shaft is held rigidly in rolling element bearings,
Equation (48) will be referred to as the KCM model.
and the bearing is allowed to move radially. The bearing
The second-order representation has been proposed as
housing is perturbed and displacements of the bearing
a nonsynchronous representation of TPJB behavior with
relative to the shaft are measured. The method, originally
twelve frequency-independent dynamic coefficients and two
applied to fixed pad hydrostatic bearings [88, 89], has been
degrees of freedom [92–94], which is in contrast with the
applied recently to flexible pivot bearings [90, 91] and four-
frequency-dependent KC model presented in Section 7.3.
pad and five-pad tilting pad journal bearings [92–95]. A
The effect of pad dynamics on the overall tilting pad bearing
brief description of the experimental identification method
dynamics is not considered explicitly in this formulation.
follows. A detailed description is available in the above
These results were reviewed by Childs [96]. In this paper,
references.
the results from several bearing tests were reviewed, and the
The system identification method employed in [92–
author stated that there was no apparent frequency depen-
95] assumes that the bearing dynamic properties can be
dency other than that captured by the KCM model, with
modeled as a two degree-of-freedom system based on tilting
the exception of the bearing originally reported in Childs
pad bearing housing (x, y) translations. The identification
and Harris [94], where an apparent frequency dependence in
procedure described in [88–95] is accomplished in the
the damping coefficient was observed. However, this paper
frequency domain by applying a sinusoidal excitation of
did not address the negative identified lubricant inertia
the form x = Xe jΩt , y = Y e jΩt , fx = Fx e jΩt , f y =
coefficients reported for several tests in [92–94].
F y e jΩt , where the excitation frequency Ω is in general non-
When the frequency response data for flexible pivot
synchronous. Excitation is accomplished by simultaneously
bearings reported in [90, 91] and tilting pad bearings [92–
applying several sinusoidal forces to the bearing housing with
94] is interpreted using (48) by the respective authors, several
mass M, resulting in a pseudorandom perturbation [88, 89].
common themes emerge. The reported data are compared to
The resulting complex impedance is then determined as a
two models: a tilting pad bearing model based solely on the
function of excitation frequency. In the frequency domain,
classic Reynolds equation, (1), and a thermohydrodynamic
the net bearing response is expressed in terms of complex
bulk-flow analysis developed by San Andrés and presented
impedances Zi j in the form:
in [74–76]. The model developed by San Andrés is based on
⎡ ⎤ theories developed by Hirs [58] and Constantinescu [29, 70]
Zxx Zxy
⎦ X = Fx + Ω2 MX .
2
⎣ (45) that include temporal and convective inertia terms from
Z yx Z y y Y F y + Ω MY the Navier-Stokes equations averaged across the lubricating
film. The quadratic behavior observed in the real part of the a Taylor series expansion about the reduced Reynolds num-
impedance occurs with both the Reynolds equation and the ber, resulting in an x-direction force:
bulk flow models. The data generally have better agreement
with the bulk-flow model than with the Reynolds equation fx = fx0(0) + Re∗ fx0(1) + kxx
(0)
+ Re∗ kxx
(1)
Δx
model, for example, [90], so the improved agreement is at-
tributed to both temporal and convective fluid inertia effects. + kxy
(0)
+ Re∗ kxy
(1)
Δ y + cxx
(0)
+ Re∗ cxx
(1)
Δx˙
(49)
8. Temporal Inertia Effects (0)
+ cxy + Re∗ cxy
(1)
Δ y˙ + Re∗ m
xx Δx¨ + m ¨
xy Δ y
The inclusion of temporal inertia effects in the bulk flow +
2 ,
+ O Re∗ .
model is justified in [90–94] using the work of Reinhardt
and Lund [97]. Reinhardt and Lund retained the temporal The hat symbol in (49) indicates nondimensional quantities.
and convective inertia terms in their nondimensional for- The y-direction force is similar.
mulation of the lubrication problem and obtained a solution Both analyses also essentially agree on the nondimen-
that indicated that temporal inertia effects were important in sionalization of force, damping coefficients, and stiffness
lubricating flows with Re ≥ 100. This is in contrast to results coefficients, with minor differences in expression of rota-
presented by Szeri et al. [98] for squeeze film dampers and tional speed:
Szeri [77] for fluid film bearings. A detailed discussion of the
fi
difference in the two models and the underlying physics is fi =
,
μωLR2 /πcr2
presented in [83] and is repeated in Section 8.1.
The theory proposed by San Andrés is in contrast to the ki j
TEHD analysis based on the generalized Reynolds equation ki j = +
,,
(50)
μLω/π (R/cr )3
proposed by He [63] and summarized in Section 5. The
generalized Reynolds equation presented by He accounts for ci j
ci j = +
,.
convective inertia effects, but not temporal inertia effects, μL/π (R/cr )3
through the use of an eddy viscosity model. The eddy
viscosity model used in the TEHD analysis by He [63] has There is a key difference in the two approaches to the inertia
an effect on both the fluid effective stiffness and damping, terms at this point. Reinhardt and Lund [97] considered the
but the TEHD theory does not predict inertia coefficients. fluid density when nondimensionalizing the inertia terms,
A discussion of the effect on stability analysis is also resulting in
presented in [90–94]. All of the papers discuss the use mi j
of synchronous versus nonsynchronous coefficients in sta- i j = -
m
.. (51)
bility analyses and state that a frequency-dependent KC ρπR2 L/π 2 (R/cr )
model requires an iterative solution to calculate the system When using (51) to dimensionalize results, Reinhardt and
eigenvalues. This is correct for older rotordynamic analyses, Lund derived a dimensional expansion with no influence of
especially transfer matrix analyses. However, modern rotor- reduced Reynolds number on the inertia terms, resulting in
dynamic codes such as the one documented by [53] are finite [97].
element based and can easily accept the additional degrees
of freedom required to represent pad motion for implemen- fx = fx0(0) + Re∗ Fx0
(1) (0)
+ kxx + Re∗ kxx
(1)
Δx
tation of the full KC TPJB model. Since the pad degrees
of freedom are explicit within this framework, an iterative (0)
+ kxy + Re∗ kxy
(1)
Δy + cxx
(0)
+ Re∗ cxx
(1)
Δẋ (52)
eigenvalue solution is not required. It has been recently
shown that the KCM model is not guaranteed to produce a (0)
+ cxy + Re∗ cxy
(1)
Δ ẏ + mxx Δẍ + mxy Δ ÿ.
conservative estimate of flexible rotor stability [99].
However, Szeri et al. [98] and Szeri [77] considered the
fluid viscosity when nondimensionalizing the inertia terms,
8.1. Comparison of Reinhardt and Lund to Szeri. There have
resulting in:
been two distinct approaches to temporal inertia effects in
hydrodynamic lubrication documented in the literature. The mi j
two approaches were compared originally in [83], and the i j = +
m
,
. (53)
μL/πω (R/cr )3
discussion is repeated here for completeness.
Both Reinhardt and Lund [97] and Szeri et al. [98] This choice of nondimensionalization results in a dimen-
considered the effects of fluid inertia on journal bearing sional expansion that indicates that reduced Reynolds num-
lubricating flows by investigation of the convective and ber is a coefficient on the dimensional inertia terms, or
temporal inertia terms in the Navier-Stokes equations. Szeri
et al. [98] considered the analysis for squeeze film dampers, fx = fx0(0) + Re∗ fx0(1) + kxx
(0)
+ Re∗ kxx
(1)
Δx
but expanded the analysis to journal bearings in [77].
Both analyses agree on the nondimensional form of the (0)
+ kxy + Re∗ kxy
(1)
Δy + cxx
(0)
+ Re∗ cxx
(1)
Δẋ (54)
perturbed N-S equations with inertia terms. The general
approach to calculate rotordynamic coefficients is to perform (0)
+ cxy + Re∗ cxy
(1)
Δ ẏ + Re∗ mxx Δẍ + mxy Δ ÿ .
When considering a laminar lubricating type flow, the Generally, consideration of more complex lubrication
flow results are dominated by fluid shear effects. As a result, models followed the experience of industrial users. As the
the fluid viscosity is more fundamental than fluid density classical Reynolds solution diverged from user experience,
in nondimensionalizing and scaling the results. This would the addition of energy and deformation effects into the
imply that for low reduced Reynolds number, the fluid inertia analysis became necessary to allow for reliable designs. The
(added mass) effects are not significant. Using data derived factors requiring these modeling improvements, including
in [97] for a plain journal bearing with a typical diameter increasing speeds and bearing specific loads, are demands
of 127 mm and Re∗ = 0.077, (54) results in an added mass by industrial users that continue to influence the need to
coefficient of 1 kg, versus an added mass coefficient of 12 kg improve tilting pad bearing models.
using (52) [83]. There has been a similar evolution in the understanding
of the bearing dynamics, especially with tilting pad journal
The Taylor series expansion resulting in (50) is also bearings. Initially treated as simple supports, inclusion of
based on Re∗ . For the Taylor series expansion to converge, stiffness effects and later damping effects improved the
Re∗ < 1 [77]. Modern bearings are reaching surface speeds understanding of the bearing contribution to the overall
where Re∗ ≥ 1, so the Reinhardt and Lund and Szeri rotordynamic system. These improvements came as user
analyses of the relative effect of temporal inertia on fluid film experience did not match with simpler bearing dynamic
bearings are no longer valid as the series does not converge. models.
These analyses are also not valid for low-viscosity process Tilting pad bearings were adopted to address self-
fluid lubricants such as water, where Re∗ 1 for typical excited vibrations from the fluid structure interactions
industrial applications. An alternative method will have to be within fixed pad bearings. The initial understanding was
found for linearized nondimensional estimates of turbulent, that the synchronous response was a sufficient representation
inertial flow. of the bearing dynamics regardless of excitation frequency
Temporal inertia terms will be important for bearings based on a misinterpretation of the work by Lund. More
with low viscosity lubricants or high Re∗ oil bearings so recent investigations, especially into rotordynamic stability,
temporal inertia terms will be present. However, those inertia indicate that the dynamic response is excitation-frequency-
terms are not very important in small amplitude linear dependent.
modeling of rotor dynamic calculations unless there is some The nonsynchronous modeling presented in Sections
sort of significant radial acceleration in the shaft. There will 7.1–7.3 is not comprehensive since pivot flexibility and foun-
dation flexibility effects are not considered. The discussion
have to be major research done as to whether an extension
does cover the basic ideas in the current understanding of
of Reynolds equation following Elrod and Ng, averaged film
bearing dynamic theory.
relations following Hirs and Constantinescu, or new forms
The KCM experimentally identified model is a funda-
of lubrication modeling will produce the most accurate mentally different model compared to the full bearing coeffi-
predictions of bearing performance. cient model. The full bearing coefficients are obtained from
first principles. The KCM model is based solely on system
9. Discussion and Conclusions identification experiments and arises from observation of the
system. The observations are consistent with a 12-coefficient
Since the original development of the lubrication equation second-order nonsynchronous dynamic representation with
by Reynolds [7], there has been an increasing level of frequency-independent stiffness, damping, and mass coef-
sophistication in the calculation of bearing properties, due ficients. This is a “black box” identification technique this
to the inclusion of thermal heating effects, mechanical is suitable for obtaining a tentative system model and is a
and thermal deformations, and turbulence corrections. The technique that is also popular in the controls community for
initial solutions only considered the fluid flow inside the developing an approximate model of a plant to be controlled.
bearing. The thermal effects were added through solutions The issue of the relative importance of temporal inertia
effects is still an open area of discussion. There are conflicting
to the energy equation and mechanical deformations were
treatments in the literature of the relative importance of the
included with deflection analyses. One of the key differences
temporal inertia term in the Navier-Stokes equations, and the
between the TEHD model presented in Section 5 and the
assumptions made in developing these treatments are being
bulk flow models presented in Section 6 is the treatment invalidated by current and projected operating speeds and
of the lubricating film. The TEHD analysis proceeds from loads in industrial bearings. Development of a new approach
a differential approach to the flow field. The mathemat- to the generalized Reynolds equation or another simplified
ics involved are more easily justified than the averaging form of the Navier-Stokes equations is an opportunity for
approaches employed in mixing length theory model of future research.
Constantinescu, which rely on an approximation of the The proper dynamic model for tilting pad journal
products of averages of the flow field. The Hirs model bearings is another area of research and discussion in the
gives good agreement with friction data because it is fit literature. This paper summarizes the two approaches, and
to that data—hence it relies entirely on extensive empirical new methods have been developed to directly compare
data to be implemented. As a result, the approach requires the two approaches [100]. A new area of research is to
experimental data for each new application. experimentally identify the pad transfer functions instead of
relying solely on implicit treatments based on measurements Appendix

of rotor or bearing housing motion. Early work in this
area has been recently reported [101]. It is anticipated A. Stiffness and Damping Terms
that the inclusion of these measurements, in conjunction
with multiple-input-multiple-output system identification A.1. Stiffness Matrices. Once the coordinate transformation
techniques, will be useful in determining the correct TPJB is applied, the single-pad stiffness matrix for a tilting pad
dynamic model. bearing with rigid pivots is
QT Kp Q
⎡ ⎤
kηη sin2 ψ + kξη + kηξ sin ψ cos ψ + kξξ cos2 ψ kξξ − kηη sin ψ cos ψ + kηξ sin2 ψ − kξη cos2 ψ −kηθ sin ψ − kξθ cos ψ
⎢ ⎥
⎢ ⎥
⎢ ⎥
= ⎢ kξξ − kηη sin ψ cos ψ + kξη sin2 ψ − kηξ cos2 ψ kηη cos2 ψ − kξη − kηξ sin ψ cos ψ + kξξ sin2 ψ kηθ cos ψ − kξθ sin ψ ⎥.
⎢ ⎥
⎣ ⎦
−kθη sin ψ − kθξ cos ψ kθη cos ψ − kθξ sin ψ kθθ
(A.1)
N p + ,
For a five pad tilting pad bearing, the full stiffness matrix %
becomes k yx = kξξ − kηη sin ψ cos ψ + kξη sin2 ψ − kηξ cos2 ψ i ;
i=1
⎡ ⎤
Kuu Kuθ
⎣ ⎦
Np + ,
Kθu Kθθ %
ky y = kηη cos2 ψ − kξη + kηξ sin ψ cos ψ + kξξ sin2 ψ i ;
i=1
⎡ ⎤
kxx kxy kxθ1 kxθ2 kxθ3 kxθ4 kxθ5
⎢ ⎥
⎢ k yx k y y k yθ1 k yθ2 k yθ3 k yθ4 k yθ5 ⎥
⎢ ⎥ kxθi = −kηθ sin ψ − kξθ cos ψ i ;
⎢ ⎥.
⎢k kθ1 y kθ1 θ1 ⎥
⎢ θ1 x 0 0 0 0 ⎥
⎢ ⎥
⎢ ⎥
= ⎢kθ2 x kθ2 y 0 kθ2 θ2 0 0 0 ⎥
⎢ ⎥ kθi x = −kθη sin ψ − kθξ cos ψ i ;
⎢ ⎥
⎢kθ3 x kθ3 y 0 0 kθ3 θ3 0 0 ⎥
⎢ ⎥
⎢ ⎥
⎢kθ4 x kθ4 y 0 0 0 kθ4 θ4 0 ⎥
⎣ ⎦ k yθi = kηθ cos ψ − kξθ sin ψ i ;
kθ5 x kθ5 y 0 0 0 0 kθ5 θ5
(A.2)
kθi y = kθη cos ψ − kθξ sin ψ i .
The terms in (A.2) are defined as
(A.3)
Np + ,
%
kxx = kηη sin2 ψ + kξη + kηξ sin ψ cos ψ + kξξ cos2 ψ i ;
i=1
N p + , A.2. Damping Matrices. Once the coordinate transformation

%
kxy = kξξ − kηη sin ψ cos ψ + kηξ sin2 ψ − kξη cos2 ψ i ; is applied, the single-pad damping matrix for a tilting pad
i=1 bearing with rigid pivots is
QT Cp Q
⎡ ⎤
cηη sin2 ψ + cξη + cηξ sin ψ cos ψ + cξξ cos2 ψ cξξ − cηη sin ψ cos ψ + cηξ sin2 ψ − cξη cos2 ψ −cηθ sin ψ − cξθ cos ψ
⎢ ⎥
⎢ ⎥
= ⎢ cξξ − cηη sin ψ cos ψ + cξη sin2 ψ − cηξ cos2 ψ cηη cos2 ψ − cξη − cηξ sin ψ cos ψ + cξξ sin2 ψ cηθ cos ψ − cξθ sin ψ ⎥
⎣ ⎦
−cθη sin ψ − cθξ cos ψ cθη cos ψ − cθξ sin ψ cθθ
(A.4)
For a five pad tilting pad bearing, the full stiffness matrix [6] E. J. Gunter, “Dynamic stability of rotor-bearing systems,”
becomes Tech. Rep. NASA SP-113, National Aeronautics and Space
⎡ ⎤ Administration, 1966.
Cuu Cuθ [7] O. Reynolds, “On the theory of lubrication and its appli-
⎣ ⎦
Cθu Cθθ cation to Mr. Beauchamp Tower’s experiments, including
an experimental determination of the viscosity of olive oil,”
⎡ ⎤ Philosophical Transactions of the Royal Society, vol. 177, pp.
cxx kxy cxθ1 cxθ2 cxθ3 cxθ4 cxθ5
⎢ ⎥ 157–234, 1886.
⎢ c yx k y y c yθ1 c yθ2 c yθ3 c yθ4 c yθ5 ⎥
⎢ ⎥ [8] A. Sommerfeld, “Zur Hydrodynamische Theorie der
⎢ ⎥
⎢c 0 ⎥
Schmiermittelreibung,” Zeitschrift für Mathematik und
⎢ θ1 x cθ 1 y cθ 1 θ 1 0 0 0 ⎥. (A.5)
⎢ ⎥ Physik, vol. 50, pp. 97–155, 1904.
⎢
= ⎢cθ2 x cθ 2 y 0 cθ 2 θ 2 0 0 0 ⎥
⎥ [9] A. Stodola, “Kritische Wellenstörung infolge der
⎢ ⎥
⎢ ⎥ Nachgiebigkeit des Oelpolsters im Lager,” Schweizerische
⎢cθ3 x cθ 3 y 0 0 cθ 3 θ 3 0 0 ⎥
⎢ ⎥ Bauzeitung, vol. 85/86, pp. 265–266, 1925.
⎢ ⎥
⎢cθ4 x cθ 4 y 0 0 0 cθ 4 θ 4 0 ⎥ [10] C. Hummel, Kritische drehzahlen als folge der nachgiebigkeit
⎣ ⎦
des schmiermittels im lager, Ph.D. thesis, Eidgenössischen
cθ 5 x c θ 5 y 0 0 0 0 cθ 5 θ 5 Technischen Hochschule in Zürich, 1926.
[11] F. C. Linn and M. A. Prohl, “The effect of flexibility of
The terms in (A.5) are defined as support upon the critical speeds of high speed rotors,”
Np + Journals & Transactions—Society of Naval Architects & Marine
% ,
Engineers, vol. 59, pp. 536–553, 1951.
cxx = cηη sin2 ψ + cξη + cηξ sin ψ cos ψ + cξξ cos2 ψ i ;
i=1
[12] J. E. L. Simmons and S. D. Advani, “Michell and the devel-
opment of tilting pad bearing,” in Fluid Film Lubrication—
N p + , Osborne Reynolds Centenary: Proceedings of the 13th Leeds-
%
cxy = cξξ − cηη sin ψ cos ψ + cηξ sin2 ψ − cξη cos2 ψ i ; Lyon Symposium on Tribology, D. Dowson, C. M. Taylor, M.
i=1 Godet, and D. Berthe, Eds., pp. 49–56, Elsevier Science, 1987.
[13] C. M. M. Ettles, “The analysis and performance of pivoted
N p + ,
% pad journal bearings considering thermal and elastic effects,”
c yx = cξξ − cηη sin ψ cos ψ + cξη sin2 ψ − cηξ cos2 ψ i ; Journal of Lubrication Technology, vol. 102, no. 2, pp. 182–
i=1 192, 1980.
Np + [14] J. Boyd and A. A. Raimondi, “An analysis of the pivoted-pad
% , journal bearing,” Mechanical Engineering, vol. 75, no. 5, pp.
cy y = cηη cos2 ψ − cξη + cηξ sin ψ cos ψ + cξξ sin2 ψ i ; 380–386, 1953.
i=1
[15] A. C. Hagg, “The influence of oil film journal bearings on the

stability of rotating machines,” Journal of Applied Mechanics,
cxθi = −cηθ sin ψ − cξθ cos ψ i ;
vol. 13, pp. A–211–A–220, 1946.
[16] B. Sternlicht, “Elastic and damping properties of partial
cθi x = −cθη sin ψ − cθξ cos ψ i ; porous journal bearings,” Journal of Basic Engineering, vol.
81, pp. 101–108, 1959.
c yθi = cηθ cos ψ − cξθ sin ψ i ; [17] D. M. Smith, Journal Bearings in Turbomachinery, Chapman
and Hall, London, UK, 1969.
cθi y = cθη cos ψ − cθξ sin ψ i . [18] O. Pinkus and B. Sternlicht, Theory of Hydrodynamic Lubri-
cation, McGraw-Hill, New York, NY, USA, 1961.
(A.6) [19] A. Tondl, Some Problems of Rotor Dynamics, Chapman and
Hall, London, UK, 1965.
References [20] J. W. Lund, “Spring and damping coefficients for the tilting-
pad journal bearing,” ASLE Transactions, vol. 42, no. 4, pp.
[1] T. Suganami and A. Z. Szeri, “A thermohydrodynamic anal- 342–352, 1964.
ysis of journal bearings,” Journal of Lubrication Technology, [21] F. K. Orcutt, “The steady state and dynamic characteristics of
vol. 101, no. 1, pp. 21–27, 1979. the tilting pad journal bearing in laminar and turbulent flow
[2] C. H. Li, “The effect of thermal diffusion on flow stability regimes,” ASME, Journal of Lubrication Technology, vol. 89,
between two rotating cylinders,” Journal of Lubrication no. 3, pp. 392–404, 1967.
Technology, vol. 99, no. 3, pp. 318–322, 1977. [22] C. W. Ng and C. H. T. Pan, “A linearized turbulent ubrication
[3] T. S. Brockett, Thermoelastohydrodynamic lubrication in theory,” Journal of Basic Engineering, vol. 87, pp. 675–682,
thrust bearings, Ph.D. thesis, University of Virginia, Char- 1965.
lottesville, Va, USA, 1994. [23] J. C. Nicholas, E. J. Gunter, and L. E. Barrett, “The influence
[4] S. Taniguchi, T. Makino, K. Takeshita, and T. Ichimura, of tilting pad bearing characteristics on the stability of high
“Thermohydrodynamic analysis of large tilting-pad journal speed rotor-bearing systems,” in Proceedings of the Design
bearing in laminar and turbulent flow regimes with mixing,” Engineering Conference, Topics in Fluid Film Bearing and
Journal of Tribology, vol. 122, no. 3, pp. 542–550, 1990. Rotor Bearing System Design and Optimization, pp. 55–78,
[5] H. Xu and J. Zhu, “Research of fluid flow and flow transition April 1978.
criteria from laminar to turbulent in a journal bearing,” [24] J. C. Nicholas, “Lund’s tilting pad journal bearing pad assem-
Journal of Xi’an Jiaotong University, vol. 27, no. 3, pp. 7–14, bly method,” Journal of Vibration and Acoustics, Transactions
1993. of the ASME, vol. 125, no. 4, pp. 448–454, 2003.
[25] J. C. Nicholas, E. J. Gunter, and P. E. Allaire, “Stiffness and 1, pp. 179–193, Turbomachinery Laboratory, Texas A&M
damping coefficients for the five-pad tilting-pad bearing,” University Press, College Station, Tex, USA, 1995.
ASLE Trans, vol. 22, no. 2, pp. 113–124, 1979. [41] J. M. Conway-Jones, “Plain bearing damage,” in Proceedings
[26] J. C. Nicholas and R. G. Kirk, “Selection and design of tilting of the 4th Turbomachinery Symposium, W. Tabakoff, Ed., vol.
pad and fixed lobe journal bearings for optimum turborotor 1, pp. 55–63, Texas A&M University Press, College Station,
dynamics,” in Proceedings of the 8th Turbomachinery Sym- Tex, USA, 1975.
posium, P. E. Jenkins, Ed., vol. 1, pp. 43–57, Texas A&M [42] P. Monmousseau and M. Fillon, “Transient thermoelastohy-
University Press, College Station, Tex, USA, 1979. drodynamic analysis for safe operating conditions of a tilting-
[27] J. C. Nicholas and R. G. Kirk, “Four pad tilting pad bearing pad journal bearing during start-up,” Tribology International,
design and application for multistage axial compressors,” vol. 33, no. 3-4, pp. 225–231, 2000.
ASME, Journal of Lubrication Technology, vol. 104, no. 4, pp. [43] R. G. Kirk and S. W. Reedy, “Evaluation of pivot stiffness
523–532, 1982. for typical tilting-pad journal bearing designs,” Journal of
[28] G. J. Jones and F. A. Martin, “Geometry effects in tilting-pad Vibration, Acoustics, Stress, and Reliability in Design, vol. 110,
journal bearings,” ASLE Trans, vol. 22, no. 3, pp. 227–244, no. 2, pp. 165–171, 1988.
1979. [44] P. Monmousseau, M. Fillon, and J. Frêne, “Transient
[29] V. N. Constantinescu, “On the pressure equation for tur- thermoelastohydrodynamic study of tilting-pad journal
bulent lubrication,” in Proceedings of the Conference on bearings—comparison between experimental data and the-
Lubrication and Wear, vol. 182–183, pp. 132–134, IMechE, oretical results,” Journal of Tribology, vol. 119, no. 3, pp. 401–
London, UK, 1967. 407, 1997.
[30] L. Malcher, Die Federungs und Dämpfungseigenschaften von [45] W. Shapiro and R. Colsher, “Dynamic characteristics of fluid-
Gleitlagern für Turbomaschinen, Ph.D. thesis, Karlsruhe Tech- film bearings,” in Proceedings of the 6th Turbomachinery Sym-
nische Hochschüle, 1975. posium, M. P. Boyce, Ed., vol. 1, pp. 39–53, Turbomachinery
[31] H. Hashimoto, S. Wada, and T. Marukawa, “Performance Laboratory, Texas A&M University Press, College Station,
characteritics of large scale tilting-pad journal bearings,” Tex, USA, 1977.
Bulletin of the JSME, vol. 28, no. 242, pp. 1761–1767, 1985. [46] P. E. Allaire, J. K. Parsell, and L. E. Barrett, “A pad
[32] J. D. Knight and L. E. Barrett, “Analysis of tilting pad journal perturbation method for the dynamic coefficients of tilting-
bearings with heat transfer effects,” Journal of Tribology, vol. pad journal bearings,” Wear, vol. 72, no. 1, pp. 29–44, 1981.
110, no. 1, pp. 128–133, 1988. [47] J. K. Parsell, P. E. Allaire, and L. E. Barrett, “Frequency effects
[33] D. Brugier and M. T. Pascal, “Influence of elastic defor- in tilting-pad journal bearing dynamic coefficients,” ASLE
mations of turbo-generator tilting pad bearings on the Transactions, vol. 26, no. 2, pp. 222–227, 1983.
static behavior and on the dynamic coefficients in different [48] American Petroleum Institute, API 684: API Standard
designs,” Journal of Tribology, vol. 111, no. 2, pp. 364–371, Paragraphs Rotordynamic Tutorial: Lateral Critical Speeds,
1989. Unbalance Response, Stability, Train Torsionals, and Rotor
[34] C. M. Ettles, “Analysis of pivoted pad journal bearing Balancing, American Petroleum Institute, Washington, DC,
assemblies considering thermoelastic deformation and heat USA, 2nd edition, 2005.
transfer effects,” Tribology Transactions, vol. 35, no. 1, pp.
[49] J. C. Nicholas, “Tilting pad bearing design,” in Proceedings
156–162, 1992.
of the 23rd Turbomachinery Symposium, J. C. Bailey, Ed., vol.
[35] K. Brockwell and W. Dmochowski, “Experimental deter- 1, pp. 179–193, Turbomachinery Laboratory, Texas A&M
mination of the journal bearingoil film coefficients by the University Press, College Station, Tex, USA, 1994.
method of selective vibration orbits,” in Proceedings of the
[50] K. E. Rouch, “Dynamics of pivoted-pad journal bearings,
12th Biennial Conference on Mechanical Vibration and Noise,
including pad translation and rotation effects,” ASLE Trans-
T. S. Sankar, V. Kamala, and P. Kim, Eds., pp. 251–259, ASME,
actions, vol. 26, no. 1, pp. 102–109, 1983.
New York, NY, USA, 1989.
[36] M. Fillon, D. Souchet, and J. Frêne, “Influence of bearing ele- [51] J. W. Lund and L. B. Pedersen, “The influence of pad
ment displacements onthermohydrodynamic characteristics flexibility on the dynamic coefficients of a tilting pad journal
of tilting-pad journal bearings,” in Proceedings of the Japan bearing,” Journal of Tribology, vol. 109, no. 1, pp. 65–70, 1987.
International Tribology Conference, pp. 635–640, Nagoya, [52] L. A. Branagan, Thermal analysis of fixed and tilting pad
Japan, 1990. journal bearings including cross-film viscosity variations and
[37] K. Brockwell, D. Kelinbub, and W. Dmochowski, “Measure- deformations, Ph.D. thesis, University of Virginia, Char-
ment and calculation of the dynamic operating characteris- lottesville, Va, USA, 1988.
tics of the five shoe, tilting pad journal bearing,” Tribology [53] J. A. Chaudhry, Rotor dynamics analysis in MatLab frame-
Transactions, vol. 33, no. 4, pp. 481–492, 1990. work, M.S. thesis, Universityof Virginia, Charlottesville, Va,
[38] G. Hopf and D. Schüeler, “Investigations on large turbine USA, 2008.
bearings working under transitional conditions between [54] L. E. Barrett, P. E. Allaire, and B. W. Wilson, “The eigenvalue
laminar and turbulent flow,” Journal of Tribology, vol. 111, dependence of reduced tilting pad bearing stiffness and
no. 4, pp. 628–634, 1989. damping coefficients,” Tribology Transactions, vol. 31, no. 4,
[39] C. H. Hyun, J. K. Ho, and W. K. Kyung, “Inlet pressure effects pp. 411–419, 1966.
on the thermohydrodynamic performance of a large tilting [55] L. L. Earles, A. B. Palazzolo, and R. W. Armentrout, “Finite
pad journal bearing,” Journal of Tribology, vol. 117, no. 1, pp. element approach to pad flexibility effects in tilt pad journal
160–165, 1995. bearings. Part I. Single pad analysis,” Journal of Tribology, vol.
[40] J. C. Nicholas and K. D. Wygant, “Tilting pad journal bearing 112, no. 2, pp. 169–177, 1990.
pivot design for high load applications,” in Proceedings of [56] L. L. Earles, A. B. Palazzolo, and R. W. Armentrout, “A finite
the 24th Turbomachinery Symposium, J. C. Bailey, Ed., vol. element approach to pad flexibility effects in tilt pad journal
bearings: part II—assembled bearing and system,” Journal of in turbulent regime part I: the model and perturbation
Tribology, vol. 112, no. 2, pp. 178–182, 1990. analysis,” Journal of Applied Mechanics, vol. 62, no. 3, pp. 674–
[57] M. F. White and S. H. Chan, “Subsynchronous dynamic 679, 1995.
behavior of tilting-pad journal bearings,” Journal of Tribology, [75] Z. L. Yang, L. San Andrés, and D. Childs, “Thermohydro-
vol. 114, no. 1, pp. 167–173, 1992. dynamic analysis of process liquid hydrostatic bearings in
[58] G. G. Hirs, “A bulk-flow theory for turbulence in lubricant turbulent regime part II: numerical solution and results,”
films,” Journal of Lubrication Technology, vol. 95, no. 2, pp. Journal of Applied Mechanics, vol. 62, no. 2, pp. 680–684,
137–146, 1973. 1995.
[59] T. S. Brockett and L. E. Barrett, “Exact dynamic reduction of [76] L. San Andrés, “Thermohydrodynamic analysis of fluid film
tilting-pad bearing models for stability analyses,” Tribology bearings for cryogenic applications,” Journal of Propulsion
Transactions, vol. 36, no. 4, pp. 581–588, 1993. and Power, vol. 11, no. 5, pp. 964–972, 1995.
[60] J. Kim, A. Palazzolo, and R. Gadangi, “Dynamic characteris- [77] A. Z. Szeri, Fluid Film Lubrication Theory & Design, Cam-
tics of TEHD tilt pad journal bearing simulation including bridge University Press, Cambridge, UK, 1998.
multiple mode pad flexibility model,” Journal of Vibration [78] G. G. Hirs, “A systematic study of turbulent film flow,”
and Acoustics, vol. 117, no. 1, pp. 123–135, 1995. Journal of Lubrication Technology, vol. 96, no. 1, pp. 118–126,
[61] M. Fillon, J. C. Bligoud, and J. Frene, “Experimental study 1974.
of tilting-pad journal bearings—comparison with theoretical [79] C. M. Taylor and D. Dowson, “Turbulent lubrication
thermoelastohydrodynamic results,” Journal of Tribology, vol. theory—application to design,” Journal of Lubrication Tech-
114, no. 3, pp. 579–588, 1992. nology, vol. 96, no. 1, pp. 36–47, 1974.
[62] B. W. Wilson and L. E. Barrett, The effect of eigenvalue- [80] L. Bouard, M. Fillon, and J. Frêne, “Comparison between
dependent tilt pad bearing characteristics on the stability of three turbulent models—application to thermohydrody-
rotor-bearing systems, M.S. thesis, University of Virginia, namic performances of tilting-pad journal bearings,” Tribol-
Charlottesville, Va, USA, 1985. ogy International, vol. 29, no. 1, pp. 11–18, 1996.
[63] M. He, Thermoelastohydrodynamic analysis of fluid film
[81] T. W. Dimond, A. A. Younan, P. E. Allaire, and J. C. Nicholas,
journal bearings, Ph.D. thesis, University of Virginia, Char-
“Modal frequency response of a four-pad tilting pad bearing
lottesville, Va, USA, 2003.
with spherical pivots, finite pivots, finite pivot stiffness and
[64] M. He and P. E. Allaire, “Thermoelastohydrodynamic analy-
different pad preloads,” in Proceedings of the ASME Turbo
sis of journal bearingswith 2D generalized energy equation,”
Expo, vol. 1, ASME, 2010.
in Proceedings of the 6th International Conference on Rotor
[82] T. W. Dimond, P. N. Sheth, P. E. Allaire, and M. He,
Dynamics, E. J. Hahn and R. B. Randall, Eds., vol. 1, IFToMM,
“Identification methods and test results for tilting pad
2002.
and fixed geometry journal bearing dynamic coefficients—a
[65] M. He, P. E. Allaire, and L. E. Barrett, “TEHD modeling of
review,” Shock and Vibration, vol. 16, no. 1, pp. 13–43, 2009.
leading edge groove tilting pad bearings,” in Proceedings of
[83] T. W. Dimond, A. A. Younan, and P. E. Allaire, “Comparison
the 6th InternationalConference on Rotor Dynamics, E. J. Hahn
of tilting-pad journal bearing dynamic full coefficient and
and R. B. Randall, Eds., vol. 1, IFToMM, Sydney, Australia,
reduced order models using modal analysis,” in Proceedings
2002.
of the ASME Turbo Expo, pp. 1043–1053, ASME, June 2009.
[66] M. He, P. Allaire, L. Barrett, and J. Nicholas, “Thermohy-
drodynamic modeling of leading-edge groove bearings under [84] N. O. Myklestad, “A new method of calculating natural
starvation condition,” Tribology Transactions, vol. 48, no. 3, modes of uncoupled bending vibration of airplane wings and
pp. 362–369, 2005. other types of beams,” Journal of the Aeronautical Sciences,
[67] P. Michaud, D. Souchet, and D. Bonneau, “Thermohy- vol. 11, pp. 153–162, 1944.
drodynamic lubrication analysis for a dynamically loaded [85] M. A. Prohl, “A general method for calculating critical speeds
journal bearing,” Proceedings of the Institution of Mechanical of flexible rotors,” Journal of Applied Mechanics, vol. 12, pp.
Engineers, Part J, vol. 221, no. 1, pp. 49–61, 2007. 142–148, 1945.
[68] B. R. Munson, D. R. Young, and T. H. Okiishi, Fundamentals [86] American Petroleum Institute, API 617: Axial and Centrifugal
of Fluid Mechanics, John Wiley & Sons, New York, NY, USA, Compressors and Expander-Compressors for Petroleum, Chem-
5th edition, 2006. ical and Gas Industry Services, American Petroleum Institute,
[69] H. G. Elrod and C. W. Ng, “A theory for turbulent fluid Washington, DC, USA, 2002.
films and its applicationto bearings,” Journal of Lubrication [87] J. A Kocur, “The state of rotordynamics—today and the
Technology, vol. 89, no. 3, pp. 346–362, 1967. future,” in Proceedings of the Rotating Machinery and Controls
[70] V. N. Constantinescu, “On turbulent lubrication,” Proceed- Laboratory (ROMAC ’09), June 2009.
ings of the IMechE, vol. 173, no. 38, pp. 881–900, 1959. [88] C. Rouvas and D. W. Childs, “A parameter identification
[71] V. N. Constantinescu, “On the influence of inertia forces in method for the rotordynamic coefficients of a high reynolds
turbulent and laminar self-acting films,” Journal of Lubrica- number hydrostatic bearing,” Journal of Vibration and Acous-
tion Technology, vol. 92, no. 3, pp. 473–481, 1970. tics, vol. 115, no. 3, pp. 264–270, 1993.
[72] V. N. Constantinescu and S. Galetuse, “On the possibilities [89] D. Childs and K. Hale, “Test apparatus and facility to identify
of improving the accuracy of the evaluation of inertia the rotordynamic coefficients of high-speed hydrostatic
forces in laminar and turbulent films,” Journal of Lubrication bearings,” Journal of Tribology, vol. 116, no. 2, pp. 337–344,
Technology, vol. 96, no. 1, pp. 69–79, 1974. 1994.
[73] V. N. Constantinescu and S. Galetuse, “Operating character- [90] A. M. Al-Ghasem and D. W. Childs, “Rotordynamic coef-
istics of journal bearings in turbulent inertial flow,” Journal ficients measurements versus predictions for a high-speed
of Lubrication Technology, vol. 104, pp. 173–179, 1982. flexure-pivot tilting-pad bearing (load-between-pad config-
[74] Z. L. Yang, L. San Andrés, and D. Childs, “Thermohy- uration),” Journal of Engineering for Gas Turbines and Power,
drodynamic analysis of process liquid hydrostatic bearings vol. 128, no. 4, pp. 896–906, 2006.
[91] L. E. Rodriguez and D. W. Childs, “Frequency dependency

of measured and predicted rotordynamic coefficients for a
load-on-pad flexible-pivot tilting-pad bearing,” Journal of
Tribology, vol. 128, no. 2, pp. 388–395, 2006.
[92] C. R. Carter and D. Childs, “Measurements versus predic-
tions for the rotordynamic characteristics of a 5-pad, rocker-
pivot, tilting-pad bearing in load between pad configura-
tion,” in Proceedings of the ASME Turbo Expo, vol. 1, ASME,
Berlin, Germany, June 2008.
[93] C. R. Carter and D. W. Childs, “Measurements versus
predictions for the rotordynamic characteristics of a five-
pad rocker-pivot tilting-pad bearing in load-between-pad
configuration,” Journal of Engineering for Gas Turbines and
Power, vol. 131, no. 1, Article ID 012507, 2009.
[94] D. Childs and J. Harris, “Static Performance characteristics
and rotordynamic coefficients for a four-pad ball-in-socket
tilting pad journal bearing,” Journal of Engineering for Gas
Turbines and Power, vol. 131, no. 6, Article ID 062502, 2009.
[95] A. Delgado, G. Vannini, B. Ertas, M. Drexel, and L. Naldi,
“Identification and predictionof force coefficients in a five-
pad and four-pad tilting pad bearing for loadon-pad and
load-between-pad configurations,” in Proceedings of the
ASME Turbo Expo 2010, vol. 1, ASME, 2010.
[96] D. W. Childs, “Tilting-pad bearings: measured frequency
characteristics of their rotordynamic coefficients,” in Proceed-
ings of the 8th IFToMM International Conference on Rotor
Dynamics, vol. 1, pp. 1–8, IFToMM, Seoul, Korea, 2010.
[97] E. Reinhardt and J. W. Lund, “The influence of fluid inertia
on the dynamic properties of journal bearings,” Journal of
Lubrication Technology, vol. 97, no. 2, pp. 33–40, 1975.
[98] A. Z. Szeri, A. A. Raimondi, and A. Giron-Duarte, “Linear
force coefficients for squeeze-film dampers,” Journal of
Lubrication Technology, vol. 105, no. 3, pp. 326–334, 1983.
[99] T. W. Dimond, A. A. Younan, and P. Allaire, “The effect
of tilting pad journal bearing dynamic models on the
linear stability analysis of an 8-stage compressor,” Journal of
Engineering for Gas Turbines and Power. In press.
[100] T. W. Dimond, A. A. Younan, and P. Allaire, “Compari-
son of tilting-pad journal bearing dynamic full coefficient
and reduced order models using modal analysis (GT2009-
60269),” Journal of Vibration and Acoustics, Transactions of the
ASME, vol. 132, no. 5, pp. 0510091–05100910, 2010.
[101] J. C. Wilkes and D. W. Childs, “Measured and predicted
transfer functions between rotor motion and pad motion
for a rocker-back tilting-pad bearing in LOP configuration,”
in Proceedings of the ASME Turbo Expo, ASME, Vancouver,
Canada, 2011.
International Journal of
Rotating
Machinery
The Scientific
Engineering Distributed
Journal of
Journal of

World Journal
Hindawi Publishing Corporation Hindawi Publishing Corporation
Sensors
Sensor Networks
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014
Journal of
Control Science
and Engineering
Advances in
Civil Engineering
Hindawi Publishing Corporation Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014
Submit your manuscripts at

http://www.hindawi.com
Journal of
Journal of Electrical and Computer
Robotics
Engineering
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014
VLSI Design
Advances in
OptoElectronics
Modelling &
Simulation
Aerospace
Hindawi Publishing Corporation Volume 2014
Navigation and
Observation
http://www.hindawi.com Volume 2014
in Engineering
Engineering
http://www.hindawi.com
International Journal of Antennas and Active and Passive Advances in
Chemical Engineering Propagation Electronic Components Shock and Vibration Acoustics and Vibration
Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

A Review of Tilting Pad Bearing Theory PDF

Uploaded by

Copyright:

Available Formats

A Review of Tilting Pad Bearing Theory PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Review of Tilting Pad Bearing Theory PDF

Uploaded by

Copyright:

Available Formats

Hindawi Publishing Corporation

International Journal of Rotating Machinery

Timothy Dimond, Amir Younan, and Paul Allaire

Correspondence should be addressed to Timothy Dimond, [email protected]

Received 24 January 2011; Accepted 5 May 2011

Academic Editor: R. Kirk

1. Introduction The hydrodynamic action in fluid film bearings is fun-

Many of the investigators of bearing dynamics were

V1 clearance cd as the characteristic length, and ωD as the

The first attempts to quantify the dynamic response of

results. Γ η, z, Re∗ = χ2 η, ξ, z, Re∗

5. Thermoelastohydrodynamic Tilting Pad

Bearing Lubrication Theory −

fixed geometry isoviscous analytical solutions, to advanced G η, z, Re∗ =

u η, ξ, z = χ2 η, ξ, z, Re∗ eﬀective viscosity is found by.

χ1 η, h, z, Re∗ (7) where ν is the kinematic viscosity. The eddy viscosity m is

× χ1 η, ξ, z, Re , In (11), based on turbulent boundary layer theory, κ = 0.4

6. Averaged Flow Approaches  =i

⎧ ⎫ approach was also heavily influenced by the performance

Dynamic reduction is performed in the frequency ⎡N

[sCuu + Kuu − (sCuθ + Kuθ ) ⎡ ⎤ ⎡ ⎤

relying solely on implicit treatments based on measurements Appendix

N p +  , A.2. Damping Matrices. Once the coordinate transformation

[91] L. E. Rodriguez and D. W. Childs, “Frequency dependency

Hindawi Publishing Corporation

Submit your manuscripts at

You might also like

6. Averaged Flow Approaches =i

[sCuu + Kuu − (sCuθ + Kuθ ) ⎡ ⎤ ⎡ ⎤

N p + , A.2. Damping Matrices. Once the coordinate transformation