diagnosis and solution of high noise and vibration issues
TRANSCRIPT
Diagnosis and Solution of High Noise andDiagnosis and Solution of High Noise andVibration Issues After a PropyleneVibration Issues After a PropyleneRefrigeration Compressor Re-rateRefrigeration Compressor Re-rate
ByZheji Liu
Mark J. KuzdzalDresser-Rand
Kenneth E. AtkinsEngineering Dynamics Inc.
Randy RialsWestlake Chemical
IntroductionIntroduction
Noise is a working environment concern
excessive noise exposure can cause hearingdamage, raising stress level, and interfere withverbal communication.
Excessive noise and vibration may inflict fatiguedamage on mechanical systems
This is the focus of this study.
HistoryHistoryA propylene refrigeration turbo-compressor installed in a ChemicalPlant was revamped in 2006 to increase the volumetric flow ratewith new stationary and rotating components.
In October 2006, the centrifugal compressor was restarted andsignificant noise and piping vibration was observed from thedischarge pipe attached to the 4th nozzle.
The level of noise and vibration appeared to vary slightly with theflow rate, but even at relatively low flow rates the vibration andnoise was excessive.
The discharge piping downstream of nozzle 4 had some vibrationsince startup in 1998, but was nowhere near the levels that werenow being observed.
Description of the Process CompressorDescription of the Process CompressorThe subject compressor draws gasfrom two process pipes anddischarges gas through threedischarge pipes.
Main inlet 1: 42-inch-diameter .
2nd inlet: a 20-inch nozzle (2) to bringin additional gas.
1st outlet: a 10-inch nozzle (3) to takegas out of compressor.
2nd outlet: an 8-inch nozzle (4) (Thisimpeller/nozzle/discharge pipe is thefocus of this study). The 8-inch nozzleis then transitioned into a 12-inch-diameter discharge pipe (4D).
Final discharge: 20-inch-diameterfinal discharge nozzle (5). 1st Inlet
1st Outlet
2nd Outlet(4D)
Final Discharge
1
2nd Inlet
Noise and Vibration IssueNoise and Vibration Issue
High noise andvibration wasobserved only on the4D pipe.
It caused externalpiping connections onthe 4D pipe todevelop cracks.
Crack location
First AttemptFirst AttemptPlant engineering resources initiallycontacted an outside firm to assist indetermining the cause of the high levels ofvibration and noise.
Pressure pulsations, piping vibration anddynamic strain were measured on the 4Dpipe.
The conclusion from the first consultantwas that the “Mokveld” check valve locateddownstream of the compressor wasinducing an “acoustic” vibration resonanceon the piping. Vortex shedding wasclaimed as the root cause of the high noiseand vibration.
Noise and Vibration Remain HighNoise and Vibration Remain High
In December 2006, the production unit was shutdown and anew “Wafer” check valve and replacement piping wasinstalled.
Upon restart of the propylene compressor, the pipingvibration and noise levels returned to or near the samelevels as before.
The investigation into the high noise and vibration in thedischarge piping downstream of nozzle 4 continued withanother troubleshooting test on January 12, 2007. This testprogram was undertaken by the authors of this study.
Root Cause DiagnosisRoot Cause DiagnosisPressure pulsations, piping vibration, dynamic strain, and sound pressurelevel were measured on the 4D pipe.
Sound Intensity level was measured along 4D pipe. (Additional)
Test ConditionsTest ConditionsTest lasted about 9hours, dynamicsignals (pulsation,vibration, strain andnoise) werecontinuouslyrecorded.
Speed, flow rate,pressure, andtemperature werevaried during thetest.
__ 3D Press ___ 4D Temp ___ 4D Flow ___ Speed ___ Total Flow
Source IdentificationSource Identification- by Frequency Based Technique- by Frequency Based Technique
Both sound pressure level and soundintensity level were acquired in narrowband (FFT).
Noise source is discrete in nature withstrong tonal components - fundamental at480 Hz, and integer multiples at 960 Hzand 1440 Hz.
This type of noise signature is not likelyfrom a Vortex shedding phenomenon.
It can not be a sole result of eitheracoustical or structural resonance of the4D pipe.
The excitation source is not likely withinthe 4D pipe.
Calc. Intensity Spectrum(Mic B,Mic A) - Mark 3Intensity : S1_10:22 : Input : FFT Analyzer
0 500 1k 1.5k 2k 2.5k 3k 3.5k 4k 4.5k 5k 5.5k 6k
50
60
70
80
90
100
110
[Hz]
[dB(A)/1.000p W/m²] Calc. Intensity Spectrum(Mic B,Mic A) - Mark 3Intensity : S1_10:22 : Input : FFT Analyzer
0 500 1k 1.5k 2k 2.5k 3k 3.5k 4k 4.5k 5k 5.5k 6k
50
60
70
80
90
100
110
[Hz]
[dB(A)/1.000p W/m²]
60
70
80
90
100
110
120
0 500 1000 1500 2000 2500
Fre que ncy (Hz)
Soun
d Pr
essu
re L
evel
(dB
A)
Source Identification - by FrequencySource Identification - by Frequency(continued)(continued)
Pulsations,vibration, anddynamic strain datashow strongcoherence amongthem.
All have peaks at480 Hz(fundamental); 960Hz (2nd harmonic),and 1440 (3rdharmonics).
Source identification - by Spatially BasedSource identification - by Spatially BasedTechnique (Sound Intensity)Technique (Sound Intensity)
Measuring noise spatialvariation, sound intensityscanning is an excellenttechnique for noise sourceidentification.
Uniform intensity levelsalong the 4D pipe indicatesthe excitation source is fromupstream of the 4D pipe -the compressor.
70
80
9 0
10 0
110
Ove
rall
So
und
Int
ensi
ty L
eve
l (d
BA)
S1 S2 S3 S4
Locations
Source Identification - by ConditionSource Identification - by ConditionVariationsVariations
Peak pulsationfrequency doesnot track tospeed - not arotationalsource.
It also does notcorrelate to flowrate - not avortex sheddingsource. __ P1 Fmax ___ 4D Press ___ 4D Temp ___ Speed ___ 4D Flow
Root Cause and SolutionRoot Cause and SolutionThe source of the high noise and vibration on the 4D pipe was successfullydiagnosed with various noise source identification techniques
Frequencies - already known; strong discrete tones at 480, and 1440 Hz.
Source Location - not the 4D pipe but the compressor.
Further analysis of the geometry in the side stream volute upstream of the 4Ddischarge piping lead to the belief that standing waves in the modified collectorwere the root cause of the high noise and vibration.
With the excitation source correctly identified, a solution of using duct resonatorarray was prescribed to the Propylene compressor.
This provides quick turnaround time than modifying machine internal.
The excitation is not harmful to compressor due to poor coupling.
Duct resonator array is a proven noise control technology that has beensuccessfully implemented to 170 compressors.
Typically reduce noise level by 10+ dBA and vibration level by 90%.
Duct Resonator Array - InternalDuct Resonator Array - InternalUsing Duct Resonator array to reduce the compressornoise source internally by lowering
Pressure Pulsations
Dynamic Excitations
Structural Vibration
Noise Radiation.Benefits include
Reduce unsteady aerodynamic load onimpeller blades and diffuser vanes.
Reduce piping vibration and improvestructure integrity and instrumentation life.
Reduce compressor noise level. Require machineinternalmodification
Pipe DR Array - ExternalPipe DR Array - External
Applying the duct resonator array technology to the compressorinlet and discharge pipe.
Lining the inner wall of the inlet/discharge pipe (spool piece) withduct resonator array.
No need to disassemble compressor.
Acoustic liner
of D-R Array
Two DoF Pipe ArrayTwo DoF Pipe ArrayA two DoF pipe array was designed to reduce noise and vibration at 480 Hz and1440 Hz.
It requires no compressor internal modification and provides quick turnaround.
As it is mounted on the 4D nozzle, the pipe array minimizes the excitation energyentering the 4D pipe.
100% metal
No pressure drop
Impedance Tube TestImpedance Tube TestAn impedance tube with anacoustic driver was used to testthe pipe array via a transitioncone.
Microphones positionedupstream and downstream ofthe array to measure the soundpressure level difference acrossthe array.
A PC driven B&K Pulse FrontEnd was used for
Generating noise signal
Measuring noise
MicrophoneUpstream
MicrophoneDownstream
Sound Pressure Levels ComparisonSound Pressure Levels ComparisonDiscrete noise sourcewith tones at 700 Hzand 1900 Hz wasapplied.
This is to emulate the480 Hz and 1440 Hzdominant noisecomponents at site.
This frequency shift isdue to change in sonicvelocity of the gas(Propylene vs. Air).
Autospectrum(Signal 3) - Mark 1Working : Measurement2 : Input : FFT Analyzer
0 400 800 1.2k 1.6k 2k 2.4k 2.8k 3.2k 3.6k 4k
20
40
60
80
100
120
[Hz]
[dB/20.0u Pa] Autospectrum(Signal 3) - Mark 1Working : Measurement2 : Input : FFT Analyzer
0 400 800 1.2k 1.6k 2k 2.4k 2.8k 3.2k 3.6k 4k
20
40
60
80
100
120
[Hz]
[dB/20.0u Pa]
----- Upstream Array____ Downstream Array
Sound Pressure Levels ComparisonSound Pressure Levels Comparison(continued)(continued)
Broadbandnoise sourcewas also fedinto the pipearray.
Thiscomparisondemonstratesthe noiseattenuationbandwidth(s) ofthe array.
Autospectrum(Signal 3) - InputWorking : Input : Input : FFT Analyzer
0 400 800 1.2k 1.6k 2k 2.4k 2.8k 3.2k 3.6k 4k
30
40
50
60
70
80
90
100
110
[Hz]
[dB(A)/20.0u Pa] Autospectrum(Signal 3) - InputWorking : Input : Input : FFT Analyzer
0 400 800 1.2k 1.6k 2k 2.4k 2.8k 3.2k 3.6k 4k
30
40
50
60
70
80
90
100
110
[Hz]
[dB(A)/20.0u Pa]
----- Upstream Array____Downstream Array
Final TestFinal TestAfter the pipe array was insertedbetween the compressor 2nd outlet(nozzle 4) and the 4D dischargepipe, follow-up testing wasconducted to verify the results.
The test procedure was to installpressure transducers upstream anddownstream of the pipe array andrecord pulsation data at processconditions similar to those thatresulted in excessive pulsation,vibration, and noise levels duringthe initial testing in January 2007.
Data Comparison (Array Effect)Data Comparison (Array Effect)The final test showsvirtually no pulsation near480 Hz (red curve),whereas this location hadabout 2.3 psi pk-pk duringthe initial test on January12, 2007.
The pressure pulsation at1440 Hz was alsocompletely eliminated bythe introduction of theresonator array.
Adding the pipe arraycorrected the problem.
ConclusionsConclusionsIn-pipe pressure pulsations, piping vibration, pipe wall strain levels, externalsound pressure, and pipe surface sound intensity were measured to identifythe excitation source.
Analysis of the measured data indicated that the acoustic source originatedwithin the compressor, upstream of the side stream discharge nozzle.
With the excitation source correctly identified, a duct resonator pipe array wasdesigned to reduce the acoustic energy entering the side stream piping.
After the pipe array was installed in the discharge piping, a second site testwas performed to confirm the effectiveness of the array. Pressure pulsationdata were measured upstream and downstream of the array, which indicatedthat the pressure peaks at 480 Hz and 1440 Hz were both eliminated.
The noise and vibration issues on the side stream piping system weresuccessfully resolved. Since the duct resonator array was installed in June,2007, the chemical plant has not had any issue with the 4D discharge pipe.