non-contact acoustic emission measurements for condition...

September 3, 2012

Symbio International Workshop 2012 on Advanced Condition Monitors for Nuclear Power and Other Process Systems

Non-contact Acoustic Emission Measurements for Condition Monitoring of Bearings

in Rotating Machines using Laser Interferometry

Operation and Maintenance Technology Development Group, FBR Plant Engineering Research Center,

Japan Atomic Energy Agency (JAEA)

Yasufumi OHTA

1/18

Backgrounds

Nuclear power plants

2/18

Condition Based Maintenance (CBM)

Target equipments

An example of pump Rolling bearing

Main failure mode Rolling bearing damage

Misalignment Unbalance

Rolling bearings are one of the most important targets of condition monitoring in nuclear power plants.

Pumps

Valves

Piping

Normal

Early Coping

Early Detection

No Detection

“When and How” for an earlier detection of abnormal

symptoms

change point

Deterioration stage

Time

Coping Non-coping

In order to maximize bearing life, it is important to detect abnormal symptoms as early as possible.

Failure

Detection

3/18

AE is the most useful method for understanding abnormal states and deterioration symptoms of bearings.

Deterioration curb

Vibration analysis Oil analysis Acoustic Emission (AE)

Condition monitoring technologies

Abnormal symptom High frequency vibration resulting from bearing damage

Wear Fracture Deformation

AE (acoustic emission) method and sensor types

AE method

AE sensor types

Previous studies using interferometers have only been applied to the measurement of static objects.

Elastic waves

Frequency several 10 kHz to several 1 MHz

AE signals

Non-contact type

[1] Palmer（1977）, [2] Bruttomesso et al.（1993）, [3] Watanabe et al.（2003）

Laser interferometer [1][2][3]

Piezoelectric sensor Optical transducer Contact type

4/18

Rotating shaft

Rolling bearing

Measurement method point

Limited

Free

Noise in rotating machines

Large

Small

Piezoelectric sensor

AE source

Bearing housing

Free

Limited

The objective of this study

Select the most adequate AE parameters corresponding with changes of bearing defect size

Long term goal To be applied to condition monitoring technology in actual plants

Setup a rotating shaft laboratory test using bearings with artificial-defects

Setup a laser interferometer which can detect AE propagated on the rotating shaft

Work plan

5/18


Spatial filter

Collimating lens

Coupling

Configuration of the experimental setup

ND filter

Motor

Bearing test apparatus

Laser interferometer

Polished shaft

Beam splitter

Reference mirror

Laser

Focusing lens

6/18

Shaft Bearing housing

Bearing

AE signal processing unit

Side view Top view

Photodetector

abnormal symptom

Overview photograph of the experimental setup

7/18

Motor Piezoelectric sensor

Bearing housing

Shaft

Laser

Spatial filter

Collimating lens

Beam splitter

ND filter

Reference mirror

Photodetector

Focusing lens

Shaft

Photodetector

Laser

Reference mirror

ND filter

Focusing lens

Motor

Beam splitter

Collimating lens

Spatial filter

Measurement point


Bearing housing Bearing

Shaft


Bearing test apparatus Bearing

Conversion of AE into voltage (simplification)

8/18

Displacement

Intensity Voltage

Shaft

AE

2 Δd

Laser beam

Optical path difference

∝ cos 2π λ

(2 Δd)

Δd

Inte

nsity

Interferometer Photodetector

(Refractive index; 1)

Fringe

Δd

AE signal processing unit

Displacement

AE propagating on the shaft. schematic diagram of fringe intensity.

High pass filter

Preamplifier

Band pass filter

λ; 632.8 nm

ΔIntensity

Voltage

Time

Second hit

AE parameters

First hit

Schematic AE signals (left side), and a power spectrum (right side) obtained from one AE wave packet.

Frequency（kHz）

Peak frequency

9/18

Peak amplitude

Bearing specimen

inner diameter 30 mm outer diameter 62 mm

width 16 mm balls diameter 9.53 mm An example of a φ0.50 mm artificial defect,

which has been added on the inner race by electric discharge machining.

1 mm

Outer ring

Inner ring

Cage

Ball

10/18

6206 deep-groove ball bearing

Inner race

Defect

Test conditions

The reference value of dBAE is 1 µV

Defect size of bearing specimen (depth; 0.25 mm) defect-free, φ0.25 mm, φ0.50 mm, φ0.75 mm, φ1.00 mm

Rotational speed 837 rpm

Total number of rotations 100,000 (about 120 minutes)

Measurements using two different methods: rotating shaft using presented laser interferometer bearing housing using piezoelectric sensor

Threshold on each measurement point 62 dBAE at rotating shaft 70 dBAE at bearing housing

11/18

Number of AE hits

Measurement Time Defect size (mm)

method point (min) -free φ0.25 φ0.50 φ0.75 φ1.00


Rotating shaft

12 43 609 680 884 1380

120 1313 5437 6695 8859 13174


Bearing housing

12 0 3 3 111 188

120 0 11 26 815 1771

12/18

The number of AE hits increases with defect size. AE waves are generated stably through the test duration. When measuring on the bearing housing, similar trends are observed. However, the number of AE hits is much lower than for shaft measurements.

0

20

40

60

80

100

68 72 76 80Peak amplitude (dBAE)

Perc

enta

ge (%

)

0

20

40

60

80

100


Perc

enta

ge (%

)

0

20

40

60

80

100


Perc

enta

ge (%

)

0

20

40

60

80

100


Perc

enta

ge (%

)

0

20

40

60

80

100


Perc

enta

ge (%

)

0

20

40

60

80

100

60 64 68 72 76 80Peak amplitude (dBAE)

Perc

enta

ge (%

)

0

20

40

60

80

100


Perc

enta

ge (%

)

0

20

40

60

80

100


Perc

enta

ge (%

)

0

20

40

60

80

100


Perc

enta

ge (%

)

0

20

40

60

80

100


Perc

enta

ge (%

)

Peak amplitude distribution

The maximum peak amplitude values change irregularly and are independent of defect size.

13/18

φ0.25 mm φ0.50 mm φ0.75 mm φ1.00 mm

Defect-free Piezoelectric sensor, bearing housing

φ0.25 mm φ0.50 mm φ0.75 mm φ1.00 mm

Defect-free Laser interferometer, rotating shaft By increasing defect size, the distribution shifts towards higher amplitudes. However, the relation is rather weak.

No detection

RMS (Root Mean Square) and Peak frequency

14/18

φ0.25 mm

φ0.50 mm φ0.75 mm φ1.00 mm

Defect-free

The RMS voltage and the peak frequencies have almost the same distribution and can not be correlated with defect size.

Piezoelectric sensor Bearing housing

No detection

RMS and Peak frequency

For defect sizes between 0.50 mm and 1.00mm, the maximum RMS voltage and peak frequency increases with defect size.

15/18

Laser interferometer Rotating shaft

φ0.25 mm

φ0.50 mm φ0.75 mm φ1.00 mm

Defect-free

50 kHz 50 kHz 50 kHz

The RMS voltage and peak frequency can be used to detect defects larger than 0.50 mm.

No detection

Distribution of absolute AE energy and Frequency centroid

16/18

The correlation between the distributions and defect sizes is weak because there are few AE hits for defects smaller than 0.50 mm, and piezoelectric AE sensors have a non-flat frequency response.

10-11

10-12

10-13

10-14

10-15

Piezoelectric sensor Bearing housing

φ0.25 mm φ0.50 mm φ0.75 mm φ1.00 mm

Defect-free

Distribution of absolute AE energy and Frequency centroid

With an increase in defect size, the frequency centroid tends to broaden, particularly at lower frequency, and the AE energy reaches higher values.

17/18

10-11

10-12

10-13

10-14

10-15

Laser interferometer Rotating shaft

φ0.25 mm φ0.50 mm φ0.75 mm φ1.00 mm

Defect-free

Can clearly recognize a 0.25-mm defect.

Summary

We demonstrated that the non-contact AE measurement method using a laser interferometer can detect AE waves on a rotating shaft in a laboratory test. After analyzing various AE parameters, we observed that the frequency centroid and absolute AE energy carry the higher correlation with defects size on the rolling bearing. The distribution of frequency centroid and absolute AE energy obtained by shaft measurement can clearly detect smaller defects than bearing housing measurements using piezoelectric sensors. This method is therefore promising for condition monitoring on rotating machines in actual plants.

18/18

non-contact acoustic emission measurements for condition...

Documents