Fatigue Analysis in Caesar II

Dynaflow Lectures – Reciprocating compressors Acoustics and Mechanical Response Rotterdam, December 10th 2009 Compress

Views 138 Downloads 1 File size 3MB

Report DMCA / Copyright

DOWNLOAD FILE

Recommend stories

Citation preview

Dynaflow Lectures – Reciprocating compressors Acoustics and Mechanical Response Rotterdam, December 10th 2009

Compressor piping vibration analysis

EXAMPLE

Two parts: 1. Acoustical/pulsation study 2. Mechanical response analysis

•Labor intensive modeling •Large number of load cases.

Copyright 2009 © Dynaflow Research Group BV

2

Sequence of dependence f Acoustics is about propagation of pressure pulsations in piping systems f Source of Pressure pulsations:  Reciprocating compressors and pumps

f Pressure waves are propagated thru the piping system. f Pressure waves are reflected (partly) and transmitted (partly) at geometrical discontinuities f Pressure pulsations generate unbalanced forces that are the source of piping vibration f Sustained vibration may result in fatigue failures Copyright 2009 © Dynaflow Research Group BV

3

Agenda

f Elements of Acoustics f Aspects of Mechanical Response f Examples of Mechanical Response

Copyright 2009 © Dynaflow Research Group BV

4

Reciprocating compressors and pumps inherently produce pulsations in the suction and discharge piping

Double acting cylinder: Piston displacement opens and closes suction and discharge valves

Copyright 2009 © Dynaflow Research Group BV

5

Actual Piston movement (not purely sinusoidal) due to finite rod length

Copyright 2009 © Dynaflow Research Group BV

7

Valve openings result in a “Sawtooth” type of gas flow Due the sequence of piston movement and valve opening and closing

The shape of the sawtooth is determined by the rotational speed of the compressor, the geometry of the cylinder and the pressure ratio. Copyright 2009 © Dynaflow Research Group BV

8

Flow time history for a single acting cylinder With ideal instantaneous reacting valves

Copyright 2009 © Dynaflow Research Group BV

9

Resulting Flow Frequency Spectrum (discrete) for single acting cylinder

Copyright 2009 © Dynaflow Research Group BV

10

Double acting cylinder (slightly unsymmetrical) Head end ≠ cranck end because of the piston rod volume

Copyright 2009 © Dynaflow Research Group BV

12

Flow Frequency Spectrum (discrete) for double acting unsymmetrical cylinder

Uneven frequency components finite but small due to imperfect symmetry

Copyright 2009 © Dynaflow Research Group BV

13

Flow pulsations result in pressure pulsations

f Pressure pulsations propagate thru the piping system at the speed of sound f Speed of sound depends on:  Gas composition  Gas Temperature  Gas Density f Pressure/Flow pulsations reflect at geometrical discontinuities f Wave length of propagating wave depending on speed of sound and pulsation frequency c λ= f f Wave reflection and wave interaction results in system acoustical natural frequencies. e.g. for wave length/frequency that match a geometrical length scale standing waves may occur f Presence of Acoustical natural frequencies may result in Acoustical resonance f System will show an acoustical response to an acoustical excitation Copyright 2009 © Dynaflow Research Group BV

14

Example of acoustical natural frequency result

Copyright 2009 © Dynaflow Research Group BV

15

Limited accuracy of acoustical model

f Accuracy of prediction of acoustical natural frequencies relatively large f Error margin relatively small: +/- 5% f Errors controlled by limited number of parameters:  Geometry  Speed of sound  Compressor RPM

Copyright 2009 © Dynaflow Research Group BV

16

Guidelines for acoustical pulsation levels according API618

f Guidelines for acceptable pulsation levels. f Acceptable levels are related to (inversely proportional to) frequency, pipe diameter and (proportional to) average pressure level f Measures to control pulsation levels:  Geometry changes (controlling acoustical natural frequencies)  Changing pipe diameters to reduce pulsation level  Introduction of damping (orifice plates at location of max oscillating flow)  Introduction of additional volumes with or without internals (creating filters)  Increasing size of bottles (“windkessel” function).

Copyright 2009 © Dynaflow Research Group BV

17

Pulsation Bottles are a way to reduce pulsations Bottles serve two effects: (1) Surge volume and (2) Filter function 1. SURGE VOLUME

2. FILTER FUNCTION

Pulsation reduction is proportional to surge volume size

Maximum filter function for pulsations with a wave length that matches the bottle length

Minimum filter function (attenuation) for pulsation with a half wave length that matches the bottle length

Copyright 2009 © Dynaflow Research Group BV

18

Pulsation Bottles located near the compressor

EXAMPLE

COMPRESSORS Inlet scrubbers

Two bottles per compressor

Copyright 2009 © Dynaflow Research Group BV

Multiple pistons per compressor

19

Guidelines for Pulsation Bottle sizing 1. SINGLE CYLINDER BOTTLE

Copyright 2009 © Dynaflow Research Group BV

2. MULTICYLINDER BOTTLE

20

Acoustical filters

Volumes connected by choke tubes Filter frequency fh:

Filter frequency response

Copyright 2009 © Dynaflow Research Group BV

21

Agenda

f Elements of Acoustics f Mechanical Response f Example of Mechanical Response

Copyright 2009 © Dynaflow Research Group BV

22

Mechanical response calculation fifth edition of API 618

f Guidelines for pulsation levels. f If pulsation levels exceed guidelines system may be qualified by means of mechanical response analysis. f Vibration control by mechanical means is a possible option f Large uncertainty margin in mechanics during design (minimum 10-20%) f Acoustic is more accurate (typically +/- 5%) f Preference for reduction of pulsations and thereby shaking forces by means of acoustical measures e.g. filtering (e.g. Helmholtz resonator)

Copyright 2009 © Dynaflow Research Group BV

23

Accuracy of prediction of mechanical natural frequencies Error margin: 10-20% or many time even larger

f Modeling of Boundary conditions f Modeling of rack structures f Support clearance f Support lift off (thermal), support settling f Support stiffness i.e. stiffness of clamps and restraints f Influence of friction f Nonlinear supports (supports with gaps or single acting supports) f Uncertainties in masses f Differences between “as built” and “design” f Interaction between parallel pipes in pipe racks f Stiffness of concrete sleepers and pedestals Copyright 2009 © Dynaflow Research Group BV

24

Many vibration problems related to attached components

Examples: f Valve Actuators f Small bore branch connections f Instrument connections f Level indicators f Stairs & Ladders

Copyright 2009 © Dynaflow Research Group BV

25

Mechanical properties and pulsations

Rule of thumb: minimum mechanical natural frequency 20% above second compressor harmonic. Question: is this feasible??? Copyright 2009 © Dynaflow Research Group BV

26

Mechanical properties and pulsations (2)

f Mechanical resonance difficult to avoid due to uncertainty in mechanical nat. freq.. f Variable speed compressor makes separation virtually impossible. f At resonance condition amplitude limited by damping only (typical damping factors of 2%-3% of critical) f High stiffness results in lower amplitudes. Copyright 2009 © Dynaflow Research Group BV

27

Application of filters in combination with high mechanical natural frequencies looks ideal

Copyright 2009 © Dynaflow Research Group BV

28

Agenda

f Acoustics f Mechanical Response f Example of Mechanical Response analysis in design

Copyright 2009 © Dynaflow Research Group BV

29

Example: Mechanical Response of NAM Oude Pekela Compressor plant

EXAMPLE

Air cooler A-174

Copyright 2009 © Dynaflow Research Group BV

30

Acoustical results of suction piping

EXAMPLE

Focal area

Copyright 2009 © Dynaflow Research Group BV

31

Unbalanced shaking forces in [kN peak to peak] per pipe section and per compressor harmonic

EXAMPLE

Nodal correspondence: 3360-3430 = C2 node 1085 3350-3360 = C2 node 1070 3000-3350 = C2 node 1033

Copyright 2009 © Dynaflow Research Group BV

32

Acoustical results of interstage piping

EXAMPLE

Focal area

Copyright 2009 © Dynaflow Research Group BV

33

Unbalanced shaking forces in [kN peak to peak] per pipe section and per compressor harmonic

EXAMPLE

Nodal correspondence: 3330-3340 = C2 node 5060 3340-3350 = C2 node 5076 3380-3350 = C2 node 5097

Copyright 2009 © Dynaflow Research Group BV

34

EXAMPLE

Summary of shaking forces Conservative selection: maximum value of all harmonics

Acoustic pipe section

Caesar II node number

Force [N.] [peak-peak]

Force [N.] [0-peak]

3330-3340

5060

131

65.5

3340-3350

5076

355

177.5

3350-3380

5097

815

407.5

3360-3430

1085

535

267.5

3350-3360

1070

240

120

3000-3350

1033

81

40.5

Copyright 2009 © Dynaflow Research Group BV

35

Two-stage compression combined model

EXAMPLE

Suction (partly), Interstage (upto cooler), Discharge (complete)

Compressor discharge bottles Interstage Line

Additional discharge volumes to reduce pulsation levels in remaining piping

Discharge Line Copyright 2009 © Dynaflow Research Group BV

Aircooler E-174 nozzles

Suction Line

36

Additional discharge volumes

Copyright 2009 © Dynaflow Research Group BV

EXAMPLE

37

Harmonic frequency assessment in CAESAR II

EXAMPLE

Sweep from 4 -56 Hz with 1 Hz steps

Copyright 2009 © Dynaflow Research Group BV

38

Harmonic forces are inserted in the model

EXAMPLE

Conservative Shaking force set taken from acoustic pulsation report

Copyright 2009 © Dynaflow Research Group BV

39

Maximum dynamic stress amplitude calculation

EXAMPLE

Max amplitude 6 MPa

Copyright 2009 © Dynaflow Research Group BV

40

At a stress amplitude level of 6 MPa the number of cycles is > 1011

EXAMPLE

Carbon Steel Fatigue Curve in the high cycle range

6 MPa

Copyright 2009 © Dynaflow Research Group BV

41

Agenda

f Acoustics f Mechanical Response f Example of Mechanical Response analysis “as built”

Copyright 2009 © Dynaflow Research Group BV

42

Issue: Unacceptably high vibration level in compressor suction piping

EXAMPLE

In 5 steps to solution

1. Vibration Measurements: identification of main contributions in frequency domain 2. Acoustical Resonance: verification of acoustical natural frequencies 3. Mechanical Resonance: verification of mechanical natural frequencies 4. Identification of source of vibration problem 5. Modification proposal

Copyright 2009 © Dynaflow Research Group BV

43

Compressor plant

Copyright 2009 © Dynaflow Research Group BV

44

Structure and support details around the compressor (I)

Copyright 2009 © Dynaflow Research Group BV

45

Structure and support details around the compressor (II)

Copyright 2009 © Dynaflow Research Group BV

46

Details of the compressor location

Copyright 2009 © Dynaflow Research Group BV

47

EXAMPLE

Step 1. Vibration Measurements

120.00

33 Hz 66 Hz

49 Hz

100.00

16 Hz

99 Hz 83 Hz

Amplitude (dB)

80.00

60.00

40.00

20.00

0.00 0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

Frequency (Hz)

Copyright 2009 © Dynaflow Research Group BV

48

Intermediate conclusion from step 1

f Vibrations are at compressor harmonics f Vibrations must be result of f 1 Acoustical resonance or f 2 Mechanical resonance or f 3 High pulsation forces without resonance (compressor bottle sizing problem)

Copyright 2009 © Dynaflow Research Group BV

49

Step 2. Acoustical natural frequencies & Compressor Harmonics

EXAMPLE

250

16 Hz 200

Amplitude

150

100

50

0 10.00

20.00

Copyright 2009 © Dynaflow Research Group BV

30.00

40.00

50.00

60.00

Frequency (Hz)

70.00

80.00

90.00

100.00

50

Intermediate conclusion from step 2

EXAMPLE

f Maybe near-to-resonance condition at first compressor harmonic (16.5 Hz.) f No further acoustical resonance f Vibration peak at 16.5 Hz, most probably is due high shaking forces as a result of near resonant condition f The other vibration peaks must be the result of: 1

Mechanical resonance or

2

High pulsation forces without resonance (compressor bottle sizing problem)

Copyright 2009 © Dynaflow Research Group BV

51

EXAMPLE

Step 3. Vibration Measurements & Calculated Mechanical Natural Frequencies (Search for Mechanical Resonance)

100.00

66 Hz

90.00

83 Hz

33 Hz

80.00

Amplitude (dB)

70.00 60.00 50.00 40.00 30.00 20.00

Purple vertical lines represent pipe system natural frequencies

10.00 0.00 0.0

10.0

Copyright 2009 © Dynaflow Research Group BV

20.0

30.0

40.0

50.0 Frequency (Hz)

60.0

70.0

80.0

90.0

100.0

52

Conclusion from step 3 & Identification of cause of vibration problem f Apparently there is mechanical resonance at 33 Hz and 66 Hz and near mechanical resonance at 83 Hz f No mechanical resonance condition at the first compressor harmonic (16.5 Hz.) and at 49 Hz. and 99 Hz f The high vibration levels 33 Hz, 66 Hz and 83 Hz are of mechanical nature f The high vibration level at 16.5 Hz most probably is an acoustical resonance problem f The high vibration level at 49 Hz and 99 Hz. must be the result of High pulsation forces without resonance (compressor bottle sizing problem)

Copyright 2009 © Dynaflow Research Group BV

53

Step 4. Identification of cause of vibration problem

EXAMPLE

f The high vibration level at 16.5 Hz most probably is an acoustical resonance problem. f Apparently there is mechanical resonance at 33 Hz and 66 Hz and near mechanical resonance at 83 Hz. f The high vibration levels 33 Hz, 66 Hz and 83 Hz are of mechanical nature f No mechanical resonance condition at the first compressor harmonic (16.5 Hz.) and at 49 Hz. and 99 Hz. f The high vibration level at 49 Hz and 99 Hz. must be the result of:  High pulsation forces without resonance (compressor bottle sizing problem)

Copyright 2009 © Dynaflow Research Group BV

54

Examination of mechanical behavior

EXAMPLE

Example of 66 Hz. mode shape

Large amplitude movement in suction manifold

Copyright 2009 © Dynaflow Research Group BV

55

Step 5. Modifications

EXAMPLE

1. The high vibration levels 33 Hz, 66 Hz and 83 Hz are of mechanical nature and need a mechanical solution  Better supporting  Improved support stiffness 2. The high vibration level at 16.5 Hz is due to acoustical resonance and needs an acoustical solution, I.e. different bottles and/or orifice plates to introduce more damping 3. The high vibration level at 49 Hz and 99 Hz. are the result of high pulsation forces without resonance and must be resolved by compressor bottle (re)sizing.

Copyright 2009 © Dynaflow Research Group BV

56

Modified structure implemented and connected to attached piping AS BUILT SITUATION

Copyright 2009 © Dynaflow Research Group BV

EXAMPLE

IMPROVED AND IMPLEMENTED SITUATION

57

Conclusion from example

EXAMPLE

f Compressor vibration problems many cases are of a mixed nature f Part is mechanical f Part is acoustical f Each category requires a different approach and result in different solutions f Not all vibration problems can be solved by mechanical measures.

Copyright 2009 © Dynaflow Research Group BV

58

END

Copyright 2009 © Dynaflow Research Group BV

59