Case History Rotating

machinery
Self-Excited Vibration Caused by Labyrinth Seal of Steam Turbine Rotor
Self-excited (turbine &
Vibration generator)

SE s
BA er
Object Machine Steam turbine for LNG cryogenic power generation (Fig.1)

v_ ine
During the initial phase of commissioning on site, a vibration component of about one half
Observed (38 Hz) of the rotational speed (4,545 rpm) was observed, but the overall shaft vibration
Phenomena amplitude was as small as 25 μm PP. However, after performing a periodic inspection (shaft

SE ng
alignment adjustment and bearing disassembling & reassembling), the shaft vibration of the
38 Hz component suddenly increased up to 90 μm PP (maximum). In particular, the vibration
depended to a large extent on the output, and started to develop at about 70% of the rated

BA l E
output, and the amplitude took its maximum value at the rated output. Fig.2 shows the result
of frequency analysis of this shaft vibration (displacement).

v_ ica
This phenomenon was considered to be self-excited vibration because 38 Hz was close to the
Cause 3UHVXPHG rotor natural frequency (first order). After examining several possible causes of the
self-excited vibration in the machine, the cause estimation was made by elimination method.

Processing
SE han
Analysis and Data
Since this machine used five-pad tilting-pad bearings, and there were almost no preceding
examples, oil whirl was considered unlikely to occur. On the other hand, diaphragm
couplings were used and turbine disks were blisks made by cutting work, so that the
influence of internal friction was also considered unlikely. Judging from the remarkable
characteristics of output dependency (Fig.3) of the shaft vibration, it was finally estimated
BA ec
that the most probable cause was a fluid induced self-excited vibration. In calculations
without considering such fluid force, the shaft damping ratio at the rated rotational speed was
v_ f M

as small as 0.01~0.015, which was sensitive to such destabilizing force. As for the flow
induced self-excited vibration, the fluid force of labyrinth seals was specifically estimated to
be the cause.
SE y o

Countermeasures were taken for the following two points.

Countermeasures (1) Improvement of bearing damping characteristics (increase by about 0.1% of radial
and Results clearance, preload factor 0.8→0.0, pad layout change; bearing load direction was
shifted to pad center (LOP: load on pad configuration) by rotating 20°; this aimed to
BA iet

gain damping ratio by increasing the anisotropic nature of spring coefficients Kxx ≠
Kyy.)
(2) Installation of a swirl breaker (swirl flow prevention plate at the labyrinth seal entry)
v_ Soc

The increase in radial gap and decrease in preload factor under (1) above had an effect of
reducing the amplitude by nearly half, while the effect of the pad location change was
significant, which decreased the amplitude to about 15 μm.
Furthermore by taking countermeasure (2), the self-excited vibration component has
completely disappeared.
SE an

Although it is possible to qualitatively grasp unstable vibration caused by seals at present, it

is hard to quantitatively assess such phenomenon. In this type of excitation force, the swirl
BA ap

flow at the seal entry has large effect on its destabilizing force, so that a swirl breaker (swirl
brake) is installed as a stabilizing means.
v_ e J

References Nothing in particular.

In many instances, it is hard to find out the causes for self-excited vibrations, thus it takes
Lesson/HDUQHG time to determine appropriate countermeasures. Consequently, as a cardinal rule, we are to
Th

take indirect countermeasures such as improvement of damping characteristics at the same

time.

Keyword flow induced vibration, labyrinth seal, swirl breaker, swirl brake, tilting pad bearings

Vibration Database (v_BASE) Committee, The Japan Society of Mechanical Engineers 0004-1
 Flow direction
Flow direction
Labyrinth seal Labyrinth seal

SE s
Thrust bearing Shaft coupling side

BA er
v_ ine
Tilting pad bearing Tilting pad bearing
(5pad) High pressure turbine Low pressure turbine (5pad)

SE ng
VB004)
Fig.1
Fig.Schematic of steam
8·13 Steam turbine
turbine for rotor for cryogenic
cryogenic power power generation
generation

BA l E
Shaft vibration amplitude [µmp-p]

62.5

v_ ica
50
38Hz O.A.
40.3 µmp-p 42.4 µmp-p

SE han 25

4 545 CPM (rotational component)

7.8 µmp-p
BA ec
0 40 80 120 160 200
Frequency [Hz] Load Between pad (LBP)
It is reported this pad
v_ f M

Fig.2 Example
Fig. of frequency
8·14 Shaft analysis
vibration at theresult of the shaftofvibration
neighborhood bearing layout is less damping
to fluid excitation.
↑
Around here,
SE y o

Changing the valve opening order load between pads

improved the vibration condition to
some extent (in some cases,
Shaft vibration amplitude [µm p-p]

significantly effective).
BA iet
v_ Soc

Vibration vector as per

the valve opening order

Pad loading direction Swirl Breaker

SE an

Labyrinth
BA ap

F
Rotor
v_ e J

OK NG
LOP LBP
Th

Turbine output non-dimensionalized by maximum load

Fig.3 Shaft vibration change with changing turbine output

Vibration Database (v_BASE) Committee, The Japan Society of Mechanical Engineers 0004-2