Sie sind auf Seite 1von 18

What Are Shaft Orbits Anyway?


by MARK A. JORDAN Sr. Rotating Equipment Vibration Engineer
Industrial Machinery Diagsnotics, LLC

About the Author Mark Jordan is a Senior Field Vibration Engineer in the field of rotating equipment As a professional industry consultant, he is responsible for the diagnosis of mechanical & process malfunctions in all types of industrial machinery for both petrochemical and power generation industries. Additioanlly, Mark is a qualified shaft alignment engineer, and uses optical surveing and lasers to rectify alignment problems on all variants of machinery. Until 2001 he was a Sr. Field Engineer for Bently Nevada Corporation's MDS (Machinery Daignostic Services) group for 22 years before the company was purchased by General Electric Power Systems (GEPS). He received an Associate in Applied Science degree in Electronics Technology in 1984, and a Baccalaureate in Mechanical Engineering from the University of Nevada-Reno in 1989.

The importance, and usefulness of shaft orbits and the average shaft position as they relate to malfunction diagnostics, provides vital vibration information which would otherwise go unchecked. A distinct advantage is gained by making use of these valuable tools in the problem solving process regarding todays complex machinery malfunctions.

A myriad of vibration transducers are available in todays marketplace, but choosing the correct transducer for a specific application is not only crucial for accurate vibration monitoring, but for advanced diagnostic capabilities also. The technical elements presented here first intend to establish a viable working relationship between proximity probe generated orbits, and shaft average position information. Then, further develop an understanding of these data formats as they relate to the field of advanced machinery diagnostics. A very common question often asked by our customers is... Just what are shaft orbits anyway, and furthermore how can they help me solve my machinery problems? First, there are many ways to observe signals generated by the non-contacting proximity probe, the most common of which is the bode and polar plot formats. These plots establish a rotors frequency filtered amplitude and phase components, usually through transient and steady state operations.

This paper discusses basic theory concepts covering considerations associated with observing static and dynamic motion from non-contacting proximity vibration probes. Dynamic shaft displacement information is available from the proximity probes, but very often is not used to identify potentially harmful mechanical malfunctions. Using this proven diagnostic tool, it is possible to isolate and identify potentially harmful machinery malfunctions by observing shaft orbits in conjunction with the average rotor position within a given bearing clearance.

Page 1

What Are Shaft Orbits Anyway?

However, an understanding of orbit-timebase and average shaft centerline position can prepare the reader to comprehend how the dynamics of machinery malfunctions take place, and more importantly, how these problems can be accurately identified before a failure occurs. Therefore, by monitoring shaft orbits and average shaft centerline position within the bearings, important and relevant information to rapidly changing machinery conditions is made available. These concepts, illustrated in the following theoretical discussion, are followed by a case history in which shaft orbits and average shaft position information become the key components in identifying and solving a serious and very real vibration problem.

in coplanar fashion resulting in what is known as shaft absolute motion. The term absolute motion was used historically because antique shaft riders originally yielded this reading. Unfortunately, shaft riders do have problems of reliability, and furthermore have no capability to provide slow roll data. This severely limits their use in machinery diagnostics, and even more so in balancing. When non-contacting eddy current probes and Proximitors are used to monitor lateral shaft motion, this transducer system provides the following individual signal components: 1. A DC (Direct Current) signal which monitors the shaft average position relative to the probe mounting. 2. An AC signal (in this case, negatively fluctuating) which monitors shaft dynamic motion relative to the probe mounting. In most plant applications, transducer signals are usually processed by radial vibration monitors; these values are typically the machinery information that is displayed as the amount of DIRECT (or overall) machine vibration in mils peak to peak (pp). Use of proximity probes are primarily applicable to those machines using fluid film lubricated bearings such as seen in turbines, motors, pumps and compressors. Although a variety of different bearing types do exist, the use of proximity probes is universally ideal, and the diagnostic capability is afforded equally to all. As with so many other diagnostic applications, the comparison of what machinery condition is normal or ideal, to what actually exists is appropriate by observing proximity probes that measure the machines dynamic rotor motion. An arrangement for adequate machinery monitoring (and protection) is to install orthogonal (X&Y) proximity probes mounted at each bearing. This provides the required AC/DC signals for on line monitoring and diagnostics. When used in conjunction Page 2

Diagnostic conclusions concerning rotating equipment using fliud film bearings are generally based upon the quality of data acquired from a machines transducer system. Failure to observe proximity probe data in its simplest form can lead to gross shortcomings in the interpretation of the data obtained. Bearing cap vibration information alone cannot truly indicate the rotors dynamic response while in a state of malfunction. Casing absolute measurements acquired by seismic transducers (either velocity or accelerometer) can sometimes be grossly inaccurate in the lower frequency ranges. Therefore, by utilizing a case mounted transducer system by itself can only be viewed as an indirect method of quantifying a machines operational condition. Conversely, proximity probes can measure the direct relative response of the rotor to the stationary bearing housing. And for those machines that possess high bearing cap activity, both the proximity probe and casing transducer may be used

What Are Shaft Orbits Anyway?

with a once-per-turn reference probe (also known as the Keyphasor), the diagnostic capability is even more pronounced. It is from this transducer arrangement that machinery information such as shaft orbits and average shaft position becomes available. DC Component-Average Shaft Position For identification of a rotors average position within a bearing clearance, voltage fluctuations are seen by the proximity probes as relative distance changes caused by dynamic rotor motion under operating conditions. A method of obtaining accurate shaft centerline data requires that one be cognizant a 'zero machine speed' gap voltage reference is required (at rest or <100 rpm). For horizontal machines, this reference is generally obtained with the rotor at rest, on turning gear or a very very slow rotative speed.
POINT: BRG 1 Vertical POINT: BRG 1 Horizontal MACHINE: Turbine From 06JUN95 16:44:30 To 06JUN95 16:53:05 Startup (not orbit or polar plot) UP 45 Left 45 Right REF: -9.14 Volts REF: -9.01 Volts 0.476 0.159 (-1.10, 2.26) 2990

As machine speed increases (or decreases) during startup/shutdown, measurements are made of changing gap voltages from the two orthogonally mounted probes that indicate the amount traveled inside that bearing clearance by the rotor. At running speed, the rotors average position within the bearing is then easily identified when the referenced zero speed gap values are used. Values are calculated by subtracting the running speed gap from the reference gap, then dividing through by the transducers scale factor. By analyzing average shaft position within the known diametral bearing clearance, valuable information regarding coupling and bearing alignment, overall bearing condition, oil film thickness and shaft radial loading etc., becomes available. Shaft Eccentricity Ratio , is a dimensionless quantity representing the average shaft position within a bearing (or seal). The average eccentricity ratio, obtained by dividing the distance between the average position of the shaft centerline and the bearing (or seal) centerline by the radial clearance, can vary between zero and one. Zero represents shaft concentricity with the bearing, and one represents the shaft in contact with the bearing. A trend of decreasing eccentricity ratio (0) indicates a potential stability problem. Conversely, an increasing eccentricity ratio (1) suggests the rotor is approaching the constraints of the bearing wall. Another property available from the shaft average centerline plot is the Rotor Position Angle . It is defined as the angle between a vertically oriented line drawn through the center of a bearing (horizontal machines), and the line connecting the bearing and shaft centers, measured in the direction of rotation (positive angles). This parameter helps identify rotor preloads. AC Component - Orbits The AC component of a transducers signal produces a periodic waveform from each probe in the orthogonal probe pair. A typical output waveform is shown in Figure 2 below. Page 3

2930 2340 1830 1240 590



-4 0.5 mil/div


Figure 1: Shows the shaft average centerline position within a 0.007 diametral bearing clearance using two orthogonally mounted displacement probes. Note the REF values, these are the at rest gap measurements. In this condition, it is assumed the rotor is at rest in the bottom of its bearings; therefore, all subsequent gap voltage changes are REFerenced to this starting position.

What Are Shaft Orbits Anyway?

P O I N T: B R G 3 V e r tic a l 6 0 L e f t V E C TO R0 . 6 3 2 m il p p : M A C HINE : C om p resso r 1 0 M A R 9 5 2 2 :0 4 :5 7 ta r tu p 1 X C O M P S 0 .4 0 .2 0 .0 0 .2 0 .4 0 0 .0 5 m il/d iv 20 40 C W R O TA TI O N


S R : 0 .1 0 8 2 1 5

60 5 m s e c /d iv


100 2792 rpm

P O I N T: B R G 3 V e r tic a l 6 0 L e f t D I R A M P T: 0 . 6 9 9 m il p p M A C HINE : C om p resso r 1 0 M A R 9 5 2 2 :0 4 :3 9 ta r tu p D I R E C T S 0 .4 0 .2 0 .0 0 .2 0 .4 0 0 .0 5 m il/d iv 20 40 C W R O TA TI O N 60 5 m s e c /d iv 80 100 2790 rpm

Figure 2:

The upper waveform represents the rotors synchronous (filtered to 1X) lateral vibration response, while the

lower is a representation of DIRECT vibration.

AMPLITUDE VS. TIME 2 1 0 T1 3 4 5 6 8 7 T2 VERTICAL INPUT Y AXIS 0 3 VERTICAL TIMEBASE WAVEFORM 0 45 90 135 180 225 270 315 360 270 X AXIS 1 0.5 T1 -0.5 -1 HORIZONTAL TIMEBASE WAVEFORM 0 45 90 135 180 225 270 315 360 1 2 3 4 5 6



-1 4 5 -0.5 6

2 1 9 8 7 90




-1 180 ROTATION=XtoY (CCW)



Figure 3: The graphical result of plotting Equations 1 and 2 from time T1 to time T2. To the right of the waveforms, the associated shaft orbit plotted as amplitude vs. amplitude. Numerical points (1, 2, 3 etc,.) along the timebase waveforms correspond to specific points on the orbit precession. The same inputs are used for Figure 4.

Page 4

What Are Shaft Orbits Anyway?

Note that in Figure 2, two separate waveforms are shown; the upper waveform, filtered to running speed (1X), shows a smooth sinusoid while the lower represents the unfiltered DIRECT vibration. Note, when observing DIRECT data, unless it has been specifically compensated, this data includes every component between 0 (DC) and 10 KHz (600,000 cpm) including the operating parameters such as the 1X, 2X and nX components plus scratches, shaft ovality and electrical/mechanical runout. To establish how shaft orbits are formed, we must first consider that the waveform produced by each transducer is an individually processed vibration signal. This signal, generated at a specific angular location on the rotor, relates the amount of shaft lateral motion in that plane. When two probes are mounted orthogonally (XY configuration, separated by 90), the two individual signals (waveforms) are representative of shaft peak to peak displacement in their respective angular planes, and are plotted as amplitude vs. time (Again, refer to Figures 2 and 3). The origin of a shaft orbit begins when two XY waveforms are paired together so that the time element is removed, leaving the X (horizontal plane) amplitude component vs. Y (vertical plane) amplitude component, plotted in what is commonly known as the Cartesian Coordinate system (or polar coordinate system). To better illustrate this concept, take a pair of XY timebase waveforms which are separated by a phase difference of 90, and whose waveform amplitudes are 2.00 mil pp (from the vertical probe), and 1.00 mil pp (from the horizontal probe). Subsequently, these two signals are best described by the following equations:

X()horizontal = 1.00 Cos () Y()vertical = 2.00 Sin ()

(Eq. 1) (Eq. 2)

Where =t (=rotational rotor frequency, t=time) representing the angle over one shaft revolution (T1 to T2) in radians and the numerical values (2.00 mil & 1.00 mil) are the amplitudes of lateral shaft vibration.

In Figure 3, equations 1 and 2 are plotted in the amplitude vs. time domain (waveform). The same result is achieved using a machines 1X filtered XY waveform pair, (or unfiltered waveform pair) per each bearing, resulting in a 1X filtered orbit, (or a DIRECT vibration orbit). Waveforms and DIRECT orbit precessions are typically available for display on a two channel oscilloscope. It is important to note that the scope should have a third channel, a Z channel for a Keyphasor signal input. Refer to Figure 4.

SCOPE SCALE: 0.500 mil/DIV

2 mil









1mil pp.
T1 T2


2mil pp.

Figure 4: Shows two individual time base signals being fed to an oscilloscope along with a Keyphasor signal. The orbit amplitude spans 2.00 mil pp vertically (Y direction), and 1.00 mil pp horizontally (X direction) with the XY channel amplitude scales set at 0.500 mil per division.

Page 5

What Are Shaft Orbits Anyway?

When two vibration signals are placed into a dual channel oscilloscope (or data acqusition device) and observed on its display, the amount of vibration can be displayed in timebase (sinusoid waveform, Figure 2) or in orbital form (as shown in Figure 4).In its orbit mode, the oscilloscope places the Vertical (Y) and Horizontal (X) signals along their respective axis to create a display of amplitude vs. amplitude. The form in which this takes place is governed by the following equations: X(r,)horizontal axis r Cos () Y(r,)vertical axis r Sin ()
(in radians), and r denotes lateral shaft amplitude.

Notice the orbit representation in Figure 5 is slightly elliptical. This data suggests the rotor is in good operational position with normal minor influential elements present, such as gravity, fluidic, and bearing load forces.

Y: BRG 1 Vertical 45 Left VECTOR: 0.360 mil pp 87 SR: 0.73 26 (man) X: BRG 1 Horizontal 45 Right VECTOR: 0.211 mil pp 174 SR: 0.76 117 (man) MACHINE: Turbine 13JUL95 10:52:22Startup 1X COMP UP 0.4 0.2 0.0 0.2 0.4 X: BRG 1 Horizontal Y: BRG 1 Vertical

(Eq. 3) (Eq. 4)

Where =t (=rotational rotor frequency, t=time) represents one shaft revolution

0.4 0.2 0.0 0.2 0.4

Simply put, an orbit pattern as seen on an oscilloscope is nothing more than a light beam dot moving very rapidly such that to the eye it appears on the screen as a continuous line. This rapidly moving dot represents the centerline motion of the shaft as seen by the proximity probes. Therefore, the orbit simply represents the PATH of the rotor centerline at the lateral position of the proximity probes. The Keyphasor pulse, when fed to a Z intensity input of an oscilloscope, intensifies the dot at the instant in time when the keyway (once-per-turn event) is passing under the Keyphasor probe. The Keyphasor dot on an orbit (or waveform) represents the centerline location of the shaft in its path of travel (or high spot) at the instant that the key way is in front of the Keyphasor probe. This allows identification of a fixed physical reference to the shaft. This arrangement produces not only peak to peak amplitude, but important phase information that is used for diagnostics. Figure 5 below shows actual field data processed by a vibration diagnostics software package. By taking the orbit/shaft centerline experience one step further, it is rotor loading and/or differences in dynamic stiffness at a bearing location that is indicated by the average shaft position within the bearing clearance, and the orbits elliptical shape.

0 0.05 mil/div ROTATION: Y TO X (CW)

20 2 msec/div


60 3770 rpm

Figure 5: Shows four revolutions of the rotor with the 1X

timebase waveforms to the right, and the associated 1X shaft orbit to the left. Shaft rotation is Y to X (CW).

Understanding the Data

Machinery diagnostics depends on knowledge of a machines specific bearing parameters. Elements such as diametral clearance and specific bearing type are helpful when applying the diagnostic techniques discussed here. But larger benefits are realized when the vibration data is slow roll compensated. Slow roll vectors are obtained during a period of low rotative speed when dynamic motion effects from forces such as unbalance are negligible. At this speed (usually 100-400 rpm), shaft bow and electrical-mechanical runout can be measured, and subtracted. Typically, slow roll speed should be below 10% of the first balance resonance.Let us now review a simple example of slow roll compensation.

Page 6

What Are Shaft Orbits Anyway?

P O I N T : B R G 3 V e r tic a l 4 5 L e f t V E C TO R0 . 1 7 2 m il p p : M A C HINE : C om p resso r 0 6 J U N 9 5 1 7 : 0 8 : 0 9S ta r tu p 1 X U N C O M P 0 .2 0 .1 0 .0 0 .1 0 .2 0 0 . 0 2 m il/ d iv 10 20 C W R O TA TI O N


30 2 m se c /div


50 2990 rpm

P O I N T : B R G 3 V e r tic a l 4 5 L e f t D I R A M P T: 0 . 6 2 6 m il p p M A C HINE : C om p resso r 0 6 J U N 9 5 1 6 : 5 3 : 0 5S ta r tu p D I R E C T 0 .2 0 .1 0 .0 0 .1 0 .2 0 0 . 0 2 m il/ d iv 10 20 C W R O TA TI O N 30 2 m se c /div 40 50 2990 rpm

Figure 6a: Uncompensated 1X and DIRECT data.

P O I N T : B R G 3 V e r tic a l 4 5 L e f t V E C T O R0 . 0 6 9 m il p p : M A C H IN E : C o m p re ss o r 0 6 J U N 9 5 1 7 : 0 8 : 0 9S ta r tu p 1 X C O M P 0.2 0.1 0.0 0.1 0.2 0 0 . 0 2 m il/ d iv 10 20 C W R O TA TIO N


S R: 0.105 303

30 2 m s e c / d iv


50 2990 rpm

P O I N T : B R G 3 V e r tic a l 4 5 L e f t D I R A M P T : 0 . 2 3 9 m il p p M A C H IN E : C o m p re ss o r 0 6 J U N 9 5 1 6 : 5 3 : 0 5S ta r tu p D I R E C T C O M P 0.2 0.1 0.0 0.1 0.2 0 0 . 0 2 m il/ d iv 10 20 C W R O TA TIO N 30 2 m s e c / d iv 40 50 2990 rpm

Figure 6b: This figure serves to illustrate the marked difference in the vibration waveforms when slow roll compensation
is employed. This data is the compensated 1X and DIRECT waveforms as shown in Figure 6a. SR indicates the slow roll vector.

By subtracting these slow roll vectors from a given data set, the true vibration response is achieved. Slow roll compensation can be performed on 1X, 2X, nX and DIRECT data sets. Figure 6 strongly illustrates this concept. Three simple cases exemplifying changing machinery conditions are now presented. Each case shows a relative shaft position within the bearing clearance, along with an associated orbit shape. These examples progress from normal operation, to a state of high preload malfunction. They are illustrations taken from actual data sets.

Page 7

What Are Shaft Orbits Anyway?

Case 1. Acceptable Operation

Orbit-timebase and shaft centerline data shown in Figure 7 (Case 1) represents a horizontally mounted machine in good operating condition, with a shaft rotation of Y to X (clockwise). This machine operates using a plain sleeve journal bearing design. The shaft centerline data shows the rotor to be in the lower left quadrant ( 0.5), which is acceptable and normal for the Y to X (clockwise) shaft rotation. The orbit shape can also be considered acceptable under the aforementioned parameters.

Y: IB-Y 0 DIR AMPT: 0.735 mil pp X: IB-X 90 Right DIR AMPT: 1.40 mil pp MACHINE: Rotor Kit 08MAY96 08:28:32Startup DIRECT

Y: IB-Y 0 VECTOR: 0.740 mil pp 123 SR: 0.319 0 (man) X: IB-X 90 Right VECTOR: 1.07 mil pp 244 SR: 0.411 187 (man) MACHINE: Rotor Kit 08MAY96 08:28:32Startup 1X COMP

POINT: IB-Y POINT: IB-X 0 90 Right REF: -10.21 Volts REF: -9.79 Volts 0.403 -0.183 (-0.916, 2.01) UP 8 MACHINE: Rotor Kit From 08MAY96 08:28:23 To 08MAY96 08:29:17 Startup (not orbit or polar plot)



0 -4
0.1 mil/div ROTATION: Y TO X (CW) 1840 rpm 0.2 mil/div ROTATION: Y TO X (CW) 1840 rpm


0.5 mil/div

Figure 7:

An acceptable unfiltered orbit at 3600 rpm along with its associated shaft centerline travel in the bearing during start up. Notice the machine rotation is Y to X (clockwise), as indicated by the shaft position in the LOWER LEFT QUADRANT of the shaft centerline plot. (for CCW rotation, this shaft position would be in the lower right quadrant, and the orbit shape would be a mirror image of the one shown). Eccentricity ratio shown is approximately 0.5.

Case 2. Increasing Preloading

The next data set (Case 2) demonstrates a malfunction, such as misalignment between two machines, that affect both the shaft position, and orbit display as a result of induced shaft preloading. A change of shape in the shaft orbit can, for example, provide an indication to changing preloads (i.e. misalignment) that may be acting on the rotor. A preload is typically defined as unidirectional, axial or radial (side) load due to external or internal mechanisms. It can also act to stabilize or destabilize the dynamic condition of

the machine. If the restraining forces (the dynamic stiffness) in the machine are equal in all radial directions, with the only force acting on the rotor is its residual imbalance, then the orbit, theoretically, should be completely circular. Dynamic stiffness refers to the spring stiffness of the mechanical system, complemented by the dynamic effects of mass and damping. Dynamic stiffness is a characteristic of the mechanical system, and it opposes an applied dynamic force to limit vibration response. A common (and mistakenly Page 8

What Are Shaft Orbits Anyway?

so) notion is to think that machines in good working order should possess perfectly round orbits. This is simply not true, especially for machines with plain sleeve type bearings. Other forces, or unequal restraints cause the rotor to respond in various noncircular shapes such as those illustrated in Figures 7,

8, and 9.

Y: IB-Y 0 DIR AMPT: 3.18 mil pp X: IB-X 90 Right DIR AMPT: 0.966 mil pp MACHINE: Rotor Kit 08MAY96 09:57:30 Steady State DIRECT COMP

0 153 VECTOR: 2.59 mil pp Y: IB-Y X: IB-X 90 Right VECTOR: 0.740 mil pp 224 MACHINE: Rotor Kit 08MAY96 09:57:30 Steady State 1X COMP SR: 0.452 SR: 0.380 129 (man) 221 (man)

0 90 Right REF: -10.38 Volts REF: -9.94 Volts 1.28 -0.500 (-2.50, 6.41) UP 3890 POINT: IB-X MACHINE: Rotor Kit From 08MAY96 09:54:29 To 08MAY96 09:59:11 Shutdown (not orbit or polar plot)




1485 1483 5 1486 1486 1484 780 610 500 0 420 410

0.2 mil/div


3940 rpm

0.2 mil/div


3940 rpm

-5 1 mil/div

Figure 8:

A condition of lateral loading malfunction evidenced by shaft centerline travel into the upper left quadrant, and an elliptical orbit shape caused by a heavy resultant force acting on the rotor. Conditions such as this are typically seen on the inboard bearings of turbines/compressors when misalignment is present. Additionally, this may also be seen on the number one bearing of steam turbines, mainly due to the summation of dynamic forces present i.e., steam flow, gravity, bearing preload, fluid properties, etc.. Eccentricity ratio is very high, = .70-.80.

Case 3. Heavy Preloading

Illustrated in Case 3, the condition of progressive preload may result in the orbit shape transgressing to a shape resembling a figure 8. The associated shaft average position is shown to be adjacent to the bearing wall in the shaft centerline plot. All are indicators of excessive rotor loading that may be the result of misalignment coupled with excessive pipe strain and/or severe bearing problems. A machinery train experiencing such problems should be considered a candidate for immediate analysis and physical inspection The DIRECT orbit shows a reverse vibration precession, physically caused by the figure 8 orbit precession. Also shown is the 1X vibration precession which shows a forward vibration precession. The shaft average centerline depicts that a high eccentricity ratio exists, and this value can be determined in approximately the 0.95 region (rotor close to the bearing wall).

Page 9

What Are Shaft Orbits Anyway?

Y: IB-Y 0 DIR AMPT: 5.47 mil pp X: IB-X 90 Right DIR AMPT: 2.10 mil pp MACHINE: Turbine 08MAY96 09:07:47 Steady State DIRECT COMP

0 VECTOR: 5.13 mil pp 138 Y: IB-Y X: IB-X 90 Right VECTOR: 1.75 mil pp 225 MACHINE: Turbine 08MAY96 09:07:47 Steady State 1X UNCOMP

0 90 Right REF: -10.44 Volts REF: -9.48 Volts 1.10 -0.977 (-4.88, 5.49) UP 3840 POINT: IB-X MACHINE: Turbine From 08MAY96 09:04:50 To 08MAY96 09:08:09 Shutdown (not orbit or polar plot)




3660 3320 3210 3080 2870 2660 2390

Figure 9:

0.2 mil/div


3990 rpm

0.2 mil/div


3990 rpm

0 -5 1 mil/div

360 270 0 5

Shows a condition of heavy loading evidenced by shaft centerline travel against the bearing wall. A Figure 8 DIRECT orbit shape suggests reverse vibration precession, indicating the preload condition is severe. Eccentricity ratio very high, = 0.95. The 1X orbit shows a planar type orbit (highly restricted).

Preloads - A General Summary The three cases illustrated here show how shaft centerline position and orbit shape might respond with time to an increasing steady state unidirectional preload, progressing from normal conditions to heavy preloading, resulting in the classical Figure 8 pattern. Therefore, by observing orbital patterns over time, the degree and plane of a preload condition can be determined and tracked. Typically, a heavy preload is not a perfectly shaped Figure 8, but is more prone to displaying loops of different sizes as shown in case 3 above. The case history following this discussion illustrates this point precisely. Preloads affecting a rotor system can fall into several different categories. Radial preloads can be caused by gravity, fluidic forces, abnormal bearing loads (especially internally adjustable types), seals, and the effects of pipe strain on the machine case itself. Preloads can be categorized into two types: Soft Preloads: Soft preloads include the effects of gravity, thermal misalignment and process changes that can act on the rotor system. Hard Preloads: Due to high eccentricity ratios resulting in increased direct stiffness in the bearing. Probable causes include gross shaft to shaft misalignment, and piping forces acting on the machinery casing. In either case, preloads can be damaging if the forces involved are strong enough to create fatigue. The results of this can be realized by excessive bearing wear, shaft fatigue, and shaft cracks etc.. By observing orbit shapes, increasing rotor preloads will force the rotors precession into a more elliptical shape. An initial response is a change in the 1X and DIRECT amplitudes followed by an increase in other frequencies such as 2X, indicating that severe misalignment, or other malfunctions are present. Either case can present harmful effects on the operational condition of the machine which can lead to permature part failure or even a cracked rotor.

Page 10

What Are Shaft Orbits Anyway?

Orbit Direction & Vibration Precession It is important to recognize when the vibration precession is reversed. Figure 10 shows two orbits of the same size and shape, with different orientations and precessions. By simple observation of the orbits, the differences are clear. Orbit A might be considered normal for a bearing with the journal rotating X to Y (counterclockwise), but abnormal for Y to X (clockwise) rotation. The converse is true for orbit B - fairly normal for CW rotation, and abnormal for CCW rotation. Note the Keyphasor dot and the preceding blank section in the orbit precession. From this mark, the direction of vibration precession can be determined. The term blank-bright infers the vibration direction given by the Keyphasor location on the orbit precession, noting that the machines actual rotational direction under normal conditions should coincide with vibration direction. This convention is established by always viewing the transducer locations from the driving element, regardless of shaft rotation.

this powerful diagnostic tool. -Acceptable Figure 11 illustrates a rotor with minimal preload (other than normal loading forces, gravity, fluidic etc.), whose predominant frequency is 1 times running speed.

0.200 mil/DIV


Figure 11:

Representative of a machine with minimal preload characteristics, and no abnormal problems. Predominant amplitude frequency is 1X. Shaft rotation is X to Y (counterclockwise).

-Lateral Load Malfunctions


Figure 10: Orbit precession (shown inside the bearing

clearance). The Keyphasor blank-bright spot shows the direction of vibration precession when viewed on an oscilloscope. (Note, this serves to illustrate the rotor direction of two different machine trains).

Common Malfunctions-An Overview Orbit data presentations hold great value to the machinery vibration specialist. Many different type of malfunctions can be identified through orbit analysis. A few of these malfunctions are illustrated here to give a flavor of how much can be recognized by using

When machine-to-machine misalignment across the coupling (the most common preload cause) is present, the shaft orbit loading might resemble something similar to that shown in Figure 12. Increased angular and/or parallel offsets between two rotors is the most common cause of machinery misalignment. Other factors affecting the overall alignment condition can be attributed to something more complex such as foundation degridation when subsidence is detected due to elements such as concrete contamination (oil, cracks) or the support capability of the ground itself. By experience, the number of failed foundations that have been found by this author are many. Page 11

What Are Shaft Orbits Anyway?



0.500 mil/DIV


For a common plain sleeve fluid film bearing, the self excited vibration frequency of a whirl instability is typically within the range of 30% to 48% of machine operating speed. When fluid whirl is viewed on an oscilloscope, a key identity to look for is that the two Keyphasor dots rotate slowly against machine rotation. This indicates that the subsynchronous frequency is less than 50% of machine speed (usually indicating a fluidic malfunction caused by a bearing deficiency). Conversely, if the dots remain stationary, then the frequency is exactly half (50%) of machine speed (indicative of a rub). Lastly, if the dots appear to rotate forward, (with machine rotation) then the subsynchronous frequency is above 50% of running speed (usually indicative of a rotor, or structural resonance excitation). But relating to this topic further, there is always the age old question bantered about regarding the capability of a fluidic induced instability occuring in tilting pad bearings. In short, the answer to this is a definitive "NO!" In the bearing itself? - absolutely not. But, experience has shown that fluid whirl (and even whip) can be generated on the bearing casing itself, usually on the small lip between the casing and the shaft (which if the clearnance is too tight acts like a minature bearing). The best way to stop this is to cut small marks into the casing land area to prevent circumferential fluid flow of the oil exiting the bearing between the bearing carrier and the shaft. Normally, fluid whirl is a self excited fluidic malfunction that typically occurs in plain sleeve bearings. The most common instance is misalignment (unloading of a bearing where the eccentricity ratio approached 0) or a much more common cause is an excessive bearing clearance. Progressing beyond whirl is a condition known as whip which can be extreamly destructive to machinery if not properly diagnosed and rectified. Figure 14 shows a destructive bearing malfunction in progress. Typically known as fluid whip, its vibration

Figure 12: Representative of a machine with poor alignment, or a cocked bearing assembly. Further investigation warranted. Note, large amplitude present in the major plane of vibration. -Fluid Induced Instabilities Figure 13 shows the presence of a subsynchronous component, illustrated by the two Keyphasor dots. The precession of the whirl vibration component is always forward (same direction as machine rotation), and the orbit shape is usually circular.

2.00 mil/DIV


Figure 13: Subsynchronous components at approximately

half running speed. Suggests self excited fluid whirl. Note TWO Keyphasor dots. Usually accompanied by larger vibration amplitudes.

Page 12

What Are Shaft Orbits Anyway?

precession is forward. The period of this self excited malfunction may or may not be harmonically related to the rotational period of the shaft. When it is not, then the Keyphasor dots move in a seemingly random pattern; when it is, the multiple Keyphasor dots will appear to be stationary.

bearing, abnormal case anomalies due to thermal warping, etc.. A rub occurs as a secondary effect to a machinery malfunction, resulting in increased vibration levels, a change of orbit shape and average shaft centerline position. Once the rub continues, this malfunction response becomes dominant. Unlike fluid induced whirl and whip, a rub can take on many different orbit shapes. From looping figure 8s, with increasing amplitudes over time, to a complete circular orbit shape (full annular rub - which would be fully encompassing the seal or bearing clearance in which the rotor is in contact with). Indications that a machine is in this state of malfunction may also be identified by increased heat in the system. Lubrication oil and process temperatures may rise due to friction as heat is added to the system. This runs concurrent with increased vibration levels throughout the train, most likely due to shaft bow. The two main types of rub malfunction may be classified as follows: Partial Rub: Occurs when the rotor occasionally contacts a stationary part of the machine. The fundamental frequency most often at 50% of running speed. Therefore, the orbit may resemble a figure 8 , but with two stationary Keyphasor dots present. Due to its non regularity through partial rotor contact with the stationary part(s), other frequencies may also appear in the orbits progression. Sub multiples of running speed such as 0.25X, 0.32X etc., may be recognized by additional Keyphasor dots. For example, a 33% (1/3X) frequency may appear as a twisted looping orbit with three Keyphasor dots in its vibration precession as shown in Figure 15.

2.00 mil/DIV


Figure 14: Representative of a machine with subsynchronous excitation at a first balance rotor resonance when the machine speed is well above two times that resonance. Typically referred to as fluid whip. NOTE Multiple Keyphasor dots - THIS MALFUNCTION CAN BE VERY DESTRUCTIVE. This condition is accompanied by large vibration amplitudes that usually traverse the bearing clearance.

-Rotor Contact Malfunctions - Rub Shaft position and orbit shapes can indicate another form of machinery malfunction commonly seen in todays applications. Rubs are encountered when the rotating shaft contacts a stationary part of the machine. Malfunctions include shaft contact with seals (usually with minimal radial clearances), newly installed steam packing, contact of turbine/compressor blading due to a failed thrust

Page 13

What Are Shaft Orbits Anyway?

rub excitation force, or the machine self destructs. In any case, with this malfunction present, the machine is in imminent danger of total destruction.

3600 RPM

When the rotor and seal are in contact, the two major regimes of the rotor vibration under steady state conditions are: 1.
A forced synchronous precession due to unbalance excitations. The rotor bounces inside the seal, the lowest frequency being 1X.


0.500 mil/DIV



Forced self-excited circular reverse precession with a frequency corresponding to one of the natural frequencies of the rotor system.

Figure 15: Representative of a machine with 1/3X subsynchronous excitation caused by a partial rub condition. Note the positioning of the Keyphasor dots. For ease of rub identification, the following conditions apply to partial rub analysis:
Common Frequencies Generated by Partial Rub

Therefore, the orbit shape is predominantly circular to the extent of the seal clearance, with a reverse precession orbit. The only recommended course of action under these circumstances is to shut the machinery train down, further inspect the damage, and correct the initial problem.

A Case History
A case history is presented in which the usefulness of shaft centerline and orbit data are clearly demonstrated. The subject machine consists of an Elliott P-90 overhung centrifugal compressor driven by a six stage Elliott 2QNV-6 steam turbine. The range of operation is between 1800 and 3100 rpm. Throughout the last several months, this plant experienced a multitude of vibration problems during the operation of this machine. Bently Nevadas Machinery Diagnostic group was contracted to perform an in depth vibration analysis of this unit. As a result, through data acquisition, the analysis determined that the turbine rotors vibration precession was severely restricted at the inboard bearing (#2). The XY proximity transducers mounted permanently at each radial bearing provided the required vibration data to accurately diagnose this malfunction. Page 14

When < 2 resonance 2 resonance 3 resonance 4 resonance

Frequencies Generated 1X 1X or X 1X, X or_X 1X, X,_X or X

With other ratios possible

Where: is Rotor Speed, and is Rotor First Balance Resonance.

Full Annular Rub: Most often encountered in seals, when the seal clearance interferes with the rotating element. The precession of vibration in the full annular rub case is reversed to that of rotor direction. This condition is very destructive, with its orbit shape traversing the full extent of the bearing or seal clearance. Once initiated, the vibration may worsen until the system restraint forces overcome the

What Are Shaft Orbits Anyway?

Figure 15

Subject machinery arrangement. This machine operates with plain sleeve journal type bearings.

vibration precession at the same location is shown in Figure 17.

Y: BRG 2 Vertical 60 Left VECTOR: 0.971 mil pp 89 SR: 0.4 147 (man) X: BRG 2 Horizontal 30 Right VECTOR: 1.28 mil pp 277 SR: 0.329 238 (man) MACHINE: HP Turbine 06MAR95 09:53:09 Steady State 1X COMP UP 0.5 0.0 0.5 X: BRG 2 Horizontal 0.5
0.5 1.0 Y: BRG 2 Vertical 60 Left DIR AMPT: 1.26 mil pp X: BRG 2 Horizontal 30 Right DIR AMPT: 2.13 mil pp MACHINE: HP Turbine 06MAR95 09:53:09 Steady State DIRECT COMP UP 1.0 Y: BRG 2 Vertical 0.5 0.0

Y: BRG 2 Vertical

0.0 0.5 0 0.1 mil/div ROTATION: Y TO X (CW) 10 20 30 40 50

1.0 X: BRG 2 Horizontal 0.5 0.0

2 msec/div

2684 rpm

0.5 1.0 0 0.1 mil/div ROTATION: Y TO X (CW) 10 20 30 40 50 2 msec/div 2684 rpm

Figure 16: Shows a planar (highly preloaded) reverse orbit

vibration precession at the No. 2 bearing.

The data contained in plot 16 shows heavy rotor preloading at turbines No. 2 bearing. Note that the preload force is predominantly oriented in the vertical plane (as indicated by the upward arrow). This data was produced once coast down and slow roll vectors were acquired, which were then subsequently applied to the steady state data. Similarly, the DIRECT

Figure 17:

Shows a DIRECT highly preloaded reverse orbit vibration precession at bearing #2. A small figure 8 can be seen on the lower left tip of the orbit.

It is feasible to imagine that the vibration precession shown in Figure 17 looks as if it is part of a larger radius. The circle in question would be the radius of the bearing itself. Shown in Figure 18, the DC shaft average centerline plot for bearing No. 2. shows that the shaft position appears to be highly restricted in Page 15

What Are Shaft Orbits Anyway?

the lower right quadrant of the bearing. This would be fundamentally incorrect for a plain sleeve type bearing with a shaft rotation of Y to X (clockwise).

During that past outage, it was discovered that the turbine had been physically removed from its foundation for some off site casing work. Armed with this historical information, we turned to the vibration data to complete the analysis. Using both plot formats together, the extent of this malfunction was quickly realized. There are several elements provided by the machinery vibration data to show that the malfunction is indeed very real. To achieve an accurate diagnosis and provide a clear recommendation to resolve the problem, the vibration data was analyzed in detail. We shall call data set #1, the machinery as found , or original machinery condition data, which was acquired on 6 March 1995. Using this data, the following preliminary observations were made:

POINT: BRG 2 Vertica POINT: BRG 2 Horizontal MACHINE: HP Turbine

60 Left 30 Right

REF: -8.69 Volts REF: -9.05 Volts

-0.136 0.212 (1.10, 0.549) UP 2695

From 06MAR95 09:53:09 To 06MAR95 10:02:46 Steady State (not orbit or polar plot)


-4 0.5 mil/div


1. The rotors vibration precession clearly shows severely restricted orbits which are presented in plots 16 & 17. The malfunction state was more pronounced once the data set was slow roll compensated. 2. The average DC shaft position (plot 18) indicates the rotor is positioned in the wrong quadrant. Diagnosis The rotor progression within the bearing clearance is very minimal, so much so that it appears that the rotor is hard against the bearing wall. Additionally, the rotor travel should not be so restricted, rather the eccentricity ratio should be in the region of 0.3 to 0.4. The eccentricity ratio for the rotor position shown in plot 18 is approximately 0.99. Under normal operating conditions for a machine with Y to X (CW) shaft rotation and one that is equipped with plain sleeve type bearings, the rotor should be in the lower left quadrant. This evidence of hard preloading of the rotor to the bearing corresponds with high bearing metal temperatures observed by the plant on this bearing. The induced preload is severe enough to produce a Page 16

Figure 18:

Shows a highly restricted rotor precession within the No. 2 bearing clearance.

As a part of the original analysis, several pertinent factors were discovered by asking questions pertaining to this machines past performance. It was discovered that the No. 2 bearing temperature fluctuated over time, and was initially though to be a problematic thermocouple. The vibration levels changed unexpectedly at bearing No. 2 as process changes were initiated. However, large amplitudes were not seen, and the monitor alarms were not initiated System performance was under suspicion, basically because steam consumption was marginally higher than before the last outage. Finally, the major pieces of the puzzle were drawn together after discovering that the machine had experienced mechanical problems only since the last major overhaul.

What Are Shaft Orbits Anyway?

reversed orbit vibration precession. This suggests the presence of very high preload forces, subsequently, bearing damage at No.2 has been predicted. A recommendation to was made to the plant to remove the machine from service, inspect the bearings and verify the shaft alignment. The machinery train was subsequently shutdown, and allowed to cool. Using a coupling laser alignment system, the static (cold) shaft offsets were checked. It was discovered that the turbine elevation was too high by as much as 0.010 (10 mil). The lateral (horizontal) offsets appeared to be within the desired tolerance. The correct static shaft offsets were calculated, and the turbine elevation was adjusted accordingly.

The high preload condition previously illustrated in data set #1, figure 17, was alleviated after specific machinery alignment corrections. This course of action resulted in the orbit and timebase data presented below.
Y: BRG 2 Vertical 60 Left DIR AMPT: 0.254 mil pp X: BRG 2 Horizontal 30 Right DIR AMPT: 0.206 mil pp MACHINE: HP Turbine 10MAR95 22:00:06 Startup DIRECT COMP UP 1.0 Y: BRG 2 Vertical 0.5 0.0 0.5 1.0 1.0 X: BRG 2 Horizontal 0.5 0.0 0.5 1.0 0 0.1 mil/div ROTATION: Y TO X (CW) 10 20 30 40 50 2 msec/div 2660 rpm

As a matter of good investigative procedure, both turbine bearing caps were removed for physical inspection. Bearing No.2 (turbine inboard) was found damaged, and was subsequently replaced. The machine was prepared for startup, and excerpts from that data set (data set #2) are presented here to illustrate the dramatic improvement in the operating conditions.
Y: BRG 2 Vertical 60 Left VECTOR: 0.272 mil pp 61 SR: 0.272 189 X: BRG 2 Horizontal 30 Right VECTOR: 0.257 mil pp 133 SR: 0.293 278 MACHINE: HP Turbine 10MAR95 22:00:06 Startup 1X COMP UP 0.5 0.0 0.5 X: BRG 2 Horizontal 0.5 Y: BRG 2 Vertical

Figure 20: Shows the DIRECT response to the correction of abnormal coupling alignment. This plot is scaled accordingly with plot 17 for comparison purposes.

POINT: BRG 2 Vertical POINT: BRG 2 Horizontal MACHINE: HP Turbine

60 Left 30 Right

REF: -8.96 Volts REF: -8.89 Volts

0.378 0.403 (-0.610, 2.69) UP 2650

From 10MAR95 20:22:34 To 10MAR95 22:00:06 Startup (not orbit or polar plot)

2185 2 1480 1145 1235 1155 1020 760 535 1545



0.5 0 0.1 mil/div ROTATION: Y TO X (CW) 10 20 30 40 50

-4 -2 0 2 4 0.5 mil/div

2 msec/div

2660 rpm

Figure 19: Shows the 1X compensated orbit vibration precession at the No. 2 bearing after the alignment corrections. Notice the orbit orientation, and correct vibration precession. This plot is scaled for direct comparison to plot 16.

Figure 21: Represents a significant improvement in rotor travel. Again, the plot is scaled accordingly with plot 18 for comparison.

Page 17

What Are Shaft Orbits Anyway?

Finally, the shaft average centerline data shows a marked improvement in rotor travel within the bearing clearance, and shaft eccentricity is approximately 0.3. It was discovered that the turbine shims had been removed, and 'cleaned' in preparation for the turbine's return from the machine shop. Apparently, some of the original shims had been misplaced, and were replaced by "equivalent" shim plates. This was thought to be the main cause of malfunction in this particular case.

In todays continuously changing machinery needs, there is one thing that remains constant, namely, the need for a reliable method of accurately monitoring industrial machinery applications. This concept becomes critical when dealing with machinery that rely on fluid film hydrodynamic bearings. Continuing to observe signals from proximity probes in the simplest form is basic to quality machinery monitoring and machinery diagnostics. Many parameters needed in complex machinery malfunction analysis cannot be determined accurately by any other means. Full comprehension of rotating equipment problems and impending failures can best be viewed as insurance policy against costly mechanical failures that could have easily been identified and corrected with minimal financial impact. In our experience, many machinery failures were allowed to progress becuase they were not delt with, recognized or identified correctly leading to extreame financial losses as a result. _____________________________________ 2003 Mark A. Jordan
Revision 5

Page 18