Unexpected Vibration on

a Centrifugal Compressor
Caused by Vibration
Probe Support

S Y M P O S I A : 2 4 – 2 6 M AY 2 0 2 2 Yves Bidaut
SHORT C O U R S E S : 2 3 M AY 2 0 2 2
MAN Energy Solutions Schweiz AG
Author - Biography
Yves Bidaut
Manager Mechanical Development
R&D and Calculation
Engineering Turbomachinery
MAN Energy Solutions Schweiz AG
Zurich, Switzerland
Responsible for providing technical support in rotordynamics
and stress analysis.
Function: development and analysis of components of
centrifugal compressors for oil and gas application.
Before joining the site in Switzerland in 2003 was employed
for 6 years in MAN Energy Solutions, Berlin where he was
involved in the design, finite element analysis, rotordynamic
analysis, testing and development of centrifugal
compressors.
Received his diploma (Mechanical Engineering, 1995) from
the University of Valenciennes (France).
2 / 23
Synopsis
 During the commissioning of a Main Air Compressor in an Air Separation Unit,
the rotor showed increased radial vibrations at the Non-Drive End probe.
 The RCA revealed: The vibration resulted from the excitation of a natural
frequency of the vibration probe support, triggered by the pressure fluctuation at
the suction of the 11 blades-impeller.
 The probe support was redesigned to increase its stiffness in order to shift the
natural frequency. After replacement of the support, no particular vibration
appeared anymore.
 Generally: The vibration probe support requires careful attention, especially if the
machine is standardized for a very large operation speed range.

3 / 23
Outline

1. Background
2. Description
3. Findings
4. Root Cause Analysis
5. Actions
6. Measurements after modification
7. Lessons learned, Conclusion

4 / 23
Background – Train Arrangement, Compressor

Main Air Compressor Feature Main Air Compressor SI Unit

3 stages, in-line Suction Pressure bara 0.9
Internal coolers between St. 1 - 2 and St. 2 - 3 Discharge Pressure bara 5.6
Tilting Pad Journal Bearing Suction Temperature °C 28
Discharge Temperature °C 90
Mass Flow kg/s 160
Gas (MW) - (g/mol) Air (29)
5 / 23
Description
 Standard Compressor, designed for a
wide speed range. Same machines
(100% repeat) already in operation
w/o any issue.

 Machine delivered on site w/o test

 During commissioning:
increased radial vibrations appeared
at Non-Drive End probe
 Vibrations below trip limit
 Vibrations increased with opening IGV
 Because of unexpected vibrations
 Origine of phenomenon investigated. 6 / 23
Findings – Lateral Vibrations (1/3)
110 4’500

Anti Surge Valve

100
4’000 Discharge
IGV
90 Speed
3’500
Vibration pk-pk (mm), ASV & IGV - Position (%)

Inlet
80

3’000

Speed (rpm), Pressure (mabr)

70

60 2’500 NDE-Probe DE-Probe


50 2’000

40
Discharge Pressure

1’500  Increased NDE-Vibration up to
30
Vibration (X,Y) - NDE 1’000
60 mm with speed () then with
20 IGV- Position load ().
500
 Vibration below alarm (70 mm)
10
 Vibration (X,Y) - DE

0 0 and trip (93 mm)

11.30 h

12.00 h

12.30 h

13.00 h

13.30 h

14.00 h

14.30 h

15.00 h

15.30 h

16.00 h

16.30 h

17.00 h

17.30 h

18.00 h

18.30 h

19.00 h
Time  DE-Vibration still low ()
7 / 23
Findings – Lateral Vibrations (2/3)
110 5’000

Anti Surge Valve

100
Discharge Pressure
4’500  Sudden decrease of NDE-
90 Vibration at both probes
4’000
Vibration pk-pk (mm), ASV & IGV: Position (%)

Speed
() during increased
80
3’500 discharge pressure @
70 constant IGV & speed

Speed (rpm), Pressure (mbar)

 3’000
60

50
2’500
 All other vibration signals
IGV
2’000 (DE-MAC, Steam Turbine,
40
Axial Position) barely
1’500
30 affected by operational
20
 Vibration (X,Y) MAC - NDE 1’000 conditions.
10 500
Vibration (X,Y) MAC - DE
0 0
20.00 h

20.10 h

20.20 h

20.30 h

20.40 h

20.50 h

21.00 h

21.10 h

21.20 h

21.30 h

21.40 h

21.50 h
Time
8 / 23
Findings – Lateral Vibrations (3/3) - Waterfall diagrams
110 5’000

Anti Surge Valve Discharge Pressure

100 4’500
NDE

90
4’000

Vibration pk-pk (mm), ASV & IGV: Position (%)

Speed
80
3’500


70

Speed (rpm), Pressure (mbar)


3’000
60
2’500
50
IGV
2’000


40

1’500
30
1x 3x 5x 7x 9x 
11x
20 Vibration (X,Y) MAC - NDE 1’000

10 500
Vibration (X,Y) MAC - DE
0 0

20.00 h

20.10 h

20.20 h

20.30 h

20.40 h

20.50 h

21.00 h

21.10 h

21.20 h

21.30 h

21.40 h

21.50 h
DE Time

 Vibrations: 11x super-synchronous

 Overall Vibration increase due to 11x

1x 3x 5x 7x 9x 11x

9 / 23
Root Cause Analysis – Further Investigations of Vibrations
 Additional measurements performed at different operating conditions in order to characterize
phenomenon. Test IGV Speed Discharge Pressure ASU
3’960 rpm ≈ 2.9 bara 100%  90%
 Test Procedure :
1 35% 4’036 rpm 2.9  4.7 bara 100%  65%
4’155 rpm ≈ 3.1 bara 100%  80%
3’960 rpm ≈ 3.1 bara 100%  90%
2 45% 4’036 rpm 3.2  3.7 bara 100%  80%
4’155 rpm ≈ 3.3 bara 100%  90%
3’960 rpm ≈ 3.3 bara 100%  90%
60%
4’036 rpm 3.3  3.6 bara 100%  85%
3
50% 3’960 rpm ≈ 3.2 bara 100%
60% 4’155 rpm ≈ 3.7 bara 100%
4 45% 3’960  4’155 rpm 3.2 - 3.4 bara 100%
5 45% 3’960  4’155 rpm 3.2  3.4 bara 75%

IGV: 0% : closed – 100% : open

10 / 23
RCA – Further Investigations – Test 1
110 5’000
ASV
Speed (rpm) 100 4’500

Vibration pk-pk (mm), ASV & IGV: Position (%)

90 4'036 rpm
4'155 rpm 4’000
3'960 rpm Speed
80

Speed (rpm), Pressure (mbar)

3’500
70
Discharge Pressure 3’000
60
2’500
50
2’000
40 IGV
1’500
30

20 1’000
Vibrations (X,Y) - NDE
10 500

Vibration NDE (mm, pk-pk) 0 0

13.50 h

14.10 h

14.30 h

14.50 h

15.10 h

15.30 h

15.50 h

16.10 h

16.30 h

16.50 h

17.10 h

17.30 h

17.50 h
direct
Time

11x Vibrations NDE:

1x  Sudden increase @ ≈ 4’000 rpm
 Then slight decrease with increased pressure
 More sensitive to speed than to pressure 11 / 23
RCA – Further Investigations – Test 1

11x

 Origin of high vibrations: 11x super-synchronous

 11x synchronous always present, independly of operating condition 12 / 23
 RCA – Further Investigations – Test 4
110 4’500
Speed (rpm) 4'155 rpm
ASU
100
4’000

Vibration pk-pk (mm), ASV & IGV: Position (%)

3'960 rpm Speed
90
3’500
80

Speed (rpm), Pressure (mbar)

Discharge Pressure 3’000
70

60 2’500

50 IGV 2’000

40
1’500
30
1’000
20
Vibration (X,Y) - NDE
500
Vibration NDE (mm, pk-pk) 10

0 0
direct

9.20 h

9.30 h

9.40 h

9.50 h

10.00 h

10.10 h
Time

11x Vibrations NDE:

1x  Clear dependence on speed
 Max amplitude (80 mm) reached @ ≈ 4’070 rpm
13 / 23
RCA – Further Investigations – Test 4

direct

4’064 rpm 11x

1x

745.25 Hz (11x)

67.75 Hz (1x)

 Clear resonance @ 4’064 rpm and 11x harmonic

 Resonance frequency: 745 Hz 14 / 23
RCA – Origin of 11x - Excitation
 Listing of blade numbers of rotor and stator parts

NDE-
 Probe
DE-
 Probe




 

Stage 1 Stage 2 Stage 3

Name # blades Name # blades

 IGV ≠ 11  Impeller 2 11+11 (11 main + 11 splitter blades)
 Impeller 1 ≠ 11  Diffuser 2 >>11
 Excitation from non-uniform
 Diffuser 1 >>11  Impeller 3 ≠ 11
flow at inlet of impeller 2
 AxBearing Pads ≠ 11  Diffuser 3 >>11 15 / 23
RCA – Modal Analysis of Probe Support
 Model of assembled parts
Probe support
Probe

Bearing
support

Cross Section

3D-NX-Model
16 / 23
RCA – Modal Analysis of Probe Support – FEA-Results

464 Hz 798 Hz*

Bearing Support Probe Support

* 798 Hz: Probe considered as hollow tube.
703 Hz: Probe considered as “full” tube
 Expected natural frequency: between 703 Hz and 798 Hz
Meshed Half-Model Eigenmodes
 Excitation of the natural frequency of probe support @ approx. 750 Hz 17 / 23
RCA – Summary and Countermeasures
 Findings
 Bearing, Casing and Probes were correctly mounted
 Vibrations correlated with 11 times super-synchronous :
Number of impeller’s blades near NDE-bearing
 Higher vibration appeared at only one particular speed  resonance

 RCA
 Increased vibration results from the excitation of natural frequency of
probe support, triggered by pressure fluctuation at the suction of the
11 blades-impeller (forced vibration)

 Countermeasure
 Design of support revised  Increased stiffness to shift natural frequencies
18 / 23
Countermeasure : Modified Probe Support

3D-Model with revised Original

probe support

Modified

19 / 23
Countermeasure : Modified Probe Support
 FEA - Results

464 Hz 1’487 Hz

Bearing Support Probe Support

Meshed Half-Model Eigenmodes

 Natural frequency of probe support increases up to 1’487 Hz 20 / 23

Measurements after Modification
110 5’000
ASV Discharge Pressure
100 4’500
Vibration pk-pk (mm), ASV & IGV: Position (%)

90 4’000
80 Speed

Speed (rpm), Pressure (mbar)

3’500
70
3’000
60
2’500
50
2’000
40
1’500
30 IGV
Vibration (X,Y) - NDE 1’000
20

10 500
Vibration (X,Y) - DE
0 0  All vibrations
6.00 h
4.00 h

5.00 h

7.00 h

8.00 h

9.00 h

10.00 h
low and stable
Time
21 / 23
Lessons learnt / Summary

 The support of vibration probes requires careful attention (especially if

machines are standardized for very large operating speed range)

 If necessary, a simple FE Analysis should be performed to determine the

natural frequencies of the support and for optimization of the design.

22 / 23
Case Study – Support Probe

Thank you !
Questions ?

23 / 23