A Case Study of Bearing Condition Monitoring Using

SPM
Ruifeng Yang1, Jianshe Kang1, Jinsong Zhao2, Jie Li1, Haiping Li1
1. Mechanical Engineering College, Shijiazhuang, China
2. Military Transportation University, Tianjin, China
[email protected]

Abstract—Bearings consist one of the most widely used industrial Shock pulse method (SPM), which is patented by SPM
rotating machine elements, especially rolling bearings and instrument in the 1970s, has gained wide acceptance for
journal bearings. The capability to detect fast, accurately and condition monitoring of bearings [8, 9, 10]. SPM can provide a
easily the existence and severity of bearing fault during operation standardized shock value, which can be used as a measure of
is very important as an unexpected failure of machine can lead to bearings condition, and quantitatively divide the bearings into
significant economic losses. This paper mainly introduces an three states, namely normal, minor defective and severe
actual case study of bearing condition monitoring in industry. As defective and expressed in blue, yellow and red respectively
an effective tool in inspection the bearing condition, shock pulse [11, 12]. Therefore, maintenance engineers or managers can
method (SPM) is utilized to monitor the bearings of test rig which
make maintenance decision according to the given shock
includes an electric motor and a wind turbine. Through analysing
values. Through years of developing, SPM has developed as a
the frequency spectrum derived from the SPM instrument, the
faulty bearing is located accurately. The case study validates that
quantitative method for detecting the bearings fault [13, 14].
SPM is quite effective for monitoring the bearings condition in The relevant research of SPM up to now only mainly
industrial applications. focuses on the rolling bearings, but journal bearings are not
considered [5]. However, SPM can monitor the journal bearing
Keywords-condition-based maintenance; bearing condition utilizing the embedded vibration module in the same way. We
monitoring; SPM; fault diagnosis; frequency spectrum analysis
can also derive the actual frequency spectrum. The
combination of rolling bearings and journal bearings is not
I. INTRODUCTION considered before. Moreover, the laboratory data is mostly used
Maintenance policies include corrective maintenance, to validate the efficiency of SPM. But the actual operating
preventive maintenance and condition-based maintenance. environment of industry is more complicated than laboratory.
Compared with the other two policies, condition-based Therefore, it is needed to develop the correlative case study as
maintenance (CBM) is increasingly used in industrial a contribution to industrial applications.
applications, which has the obvious advantages in condition In this paper, an experimental system including an electric
monitoring and signal processing techniques [1]. Condition motor and a wind turbine gearbox is built. This experiment is
monitoring (CM) techniques are becoming grateful in industry actually a delivery test in an overhaul industry. The SPM
because of its efficient role in detecting potential failures [2]. instrument and shock pulse sensor are used to monitor the
The advances in condition monitoring (CM) techniques have condition of motor and gearbox. The real-time frequency
made CBM even more popular. spectrums are derived. Through the analysis and comparison of
Bearings are one of the most widely used industrial frequency spectrum and vibration data, the faulty bearing is
machine elements, especially rolling bearings and journal located accurately.
bearings. In industrial applications, these bearings are The remaining sections of this paper are organized as
considered as critical mechanical components and a defect may follows. In Section 2, the structure of the experimental system
even lead to catastrophic failure of the machinery [3]. Current is presented. Section 3 analyzes and discusses the experiment
research on bearings is driven by a growing demand for higher results through frequency spectrum and vibration data. Finally,
performance and greater availability. On the one hand, modern the conclusions are drawn in section 4.
industry needs to push the performance boundaries to faster
speed, lighter components and more efficiency. Increasing
availability, on the other hand, means enhancing the flexibility II. EXPERIMENTAL SYSTEM SPECIFICATIONS
of the machine to adapt to different operation conditions and SPM can calculate shock value quantitatively and detecting
environments, eliminate or minimize service intervals, and the bearings faults efficiently. The hand-held SPM instrument
extend their residual useful life. Such extreme needs not only and shock pulse sensor are used to implement the data
very accurate design, optimization and manufacturing, but also acquisition and processing. When the bearing element contacts
continuous monitoring and control of vibration levels [4]. with a damaged area in the rotating bearing, high frequency
Recent years have seen an increasing number of researchers to shock waves are generated and transferred to the bearing
explore the condition monitoring techniques of bearings [4, 5, housing through the inner material of bearings. SPM can derive
6, 7]. the shock wave signals using the shock pulse sensors which are

978-1-4799-7958-5/14/$31.00 ©2014 IEEE 695

placed on the bearing housing. Through the embedded SPM
frequency spectrum module, the diagram of frequency
spectrum can be derived and analyzed. Accordingly, the faulty
bearing can be detected and located.
In order to validate the efficiency of SPM on monitoring the
bearing fault, the experimental system is introduced in this
section. Fig. 1 shows the experimental system used for shock
signals collection, including the electric motor, the wind
turbine gearbox, SPM instrument and shock pulse sensor.

Fig. 1. The diagram of the experimental system, (a) illustration of the test
rig, (b) SPM instrument and shock pulse sensor.

The 1600 kW electric motor is employed to provide the

drive of the wind turbine gearbox. The bearings at load and
non-load side both are journal bearings. However, both the
bearings of the wind turbine gearbox at load and non-load side Fig. 2. The diagram of the frequency spectrum at the measuring point of
are 22236E SKF spherical rolling bearings. In the present the wind turbine non load side.
work, we only consider the steady state. The rotating speed of
motor and wind turbine gearbox is constant at 992 rpm.
In this experimental system, the fault of bearings most
likely occurs at load side and non-load side. Therefore, four
measuring points are chosen in order to measure the shock
vibration signals, as shown in Fig. 1. At each measuring point,
three directions are detected, namely horizontal, vertical and
axial direction. Since the SPM vibration module is versatile for
journal bearings and rolling bearings, the vibration module is
employed. Therefore, the vibration frequency spectrum and
vibration value are derived, which are shown in the next
section.

III. RESULTS ANALYSIS AND DISCUSSION

The diagrams of vibration frequency spectrum at each
measuring point are derived utilizing SPM instrument and
shock pulse sensor. The experimental results at different
measuring points are shown in Figs. 2-5.

696
Fig. 3. The diagram of the frequency spectrum at the measuring point of
the wind turbine load side.

Fig. 5. The diagram of the frequency spectrum at the measuring point of

the motor non load side.

As shown in Figs. 2 and 3, the operating condition of

rolling bearings at the load and non-load side of the wind
turbine gearbox can be analysed. From the frequency spectrum,
the spectrum lines are relatively seldom and gentle. The
frequency of the spectrum lines mainly concentrates between 0
and 50 Hz. Moreover, the occurrence of spectral peaks at the
first time is different, for example, the spectral peak firstly
occurs in the horizontal direction is 16.875 Hz. However, the
peaks in the vertical and axial directions are 16.25Hz. On the
other hand, the frequency doubling is not occurred using the
marking function of SPM (expressed in green lines). Therefore,
Fig. 4. The diagram of the frequency spectrum at the measuring point of the preliminary conclusion that the bearings in the wind turbine
the motor load side. gearbox are normal can be drawn.
However, from the Figs. 4 and 5, the frequency spectrum
lines are relatively uniform and occur repeatedly. The space
between frequency spectrum lines is equal and characteristic of
frequency doubling is quite obviously. All the fundamental
frequencies are the same and occur at 16.25 Hz. As we all

697
known, the purpose of frequency spectrum is to reveal the line effective tool in monitoring the bearing condition. In order to
patterns associated with bearing faults. Characteristic for many research the efficiency of SPM in inspecting bearings in actual
fault patterns is the presence of ‘multiples’ or ‘harmonics’, industrial environment, an experimental system including a
which means that the line (or group of lines) is repeated two, motor and a wind turbine gearbox is built. The actual vibration
three or more times further up in the frequency spectrum. In frequency spectrum and vibration data are derived utilizing the
other words, the spacing is 1Z, 2Z, 3Z, ..., nZ, where Z is the SPM instrument and shock pulse sensor. According to the
frequency of the first line. Moreover, with the SPM marking analysis of frequency spectrum and vibration data, the faulty
function, the position of each frequency doubling can be clearly bearing is located accurately. The case study validates that
indicated (expressed in red lines). Theoretically speaking, the
SPM technology is potentially effective for monitoring the
journal bearings at the load and non-load side of electric motor
bearings condition in industrial applications.
are faulty.
REFERENCES
TABLE I. THE VIBRATION DATA AT THE FOUR MEASURING POINTS [1] W Wang, “Overview of a semi-stochastic filtering approach for residual
life estimation with applications in condition based maintenance,”
Vibration testing(0-1000HZ) Proceedings of the Institution of Mechanical Engineers, Part O: Journal
The location of
displacement velocity acceleration of Risk and Reliability, pp. 225: 185, 2011.
measuring points
(μm) (mm/s) (m/s2)
horizontal 59.4 3.13 1.46 [2] W. Wang, “A two-stage prognosis model in condition based
Measuring maintenance,” European Journal of Operational Research, vol. 182, pp.
vertical 6.92 0.62 1.45
point 1 1177–1187, 2007.
axial 24.4 2.32 1.38
horizontal 44.17 3.76 1.03 [3] N. Tandon and A. Choudhury, “A review of vibration and acoustic
Measuring measurement methods for the detection of defects in rolling element
vertical 36.55 0.93 0.6
point 2 bearings,” Tribology International, vol. 32, pp. 469–480, 1999.
axial 14.43 1.33 0.75
horizontal 54.24 5.53 0.7 [4] S. Morosi and I. Santos, “Active lubrication applied to radial gas journal
Measuring bearings. Part 1: Modeling,” Tribology International vol. 44, pp. 1949–
vertical 8.22 0.76 0.36
point 3 1958, 2011.
axial 18.51 0.87 0.43
horizontal 74.69 7.1 0.82 [5] N. Tandon, G. S. Yadava and K. M. Ramakrishna, “A comparison of
Measuring some condition monitoring techniques for the detection of defect in
vertical 8.93 1.12 0.63
point 4 induction motor ball bearings,” Mechanical Systems and Signal
axial 9.31 2.95 1.8
Processing, vol. 21, pp. 244–256, 2007.
In order to further explain the efficiency of SPM in
[6] R. B. Randall and J. Antoni, “Rolling element bearing diagnostics—A
detecting the bearings fault, the vibration data at the four tutorial,” Mechanical Systems and Signal Processing, vol. 25, pp. 485–
measuring points are listed in Table 1. The three directions at 520, 2011.
each measuring point are all present. Moreover, all the values [7] S. W. Buuren, H. Hetzler, M. Hinterkausen and W. Seemann, “Novel
of displacement, velocity and acceleration are measured and approach to solve the dynamical porous journal bearing problem,”
listed as follows. Tribology International, vol. 46, pp. 30–40, 2012.
[8] The shock pulse method for determining condition of anti-friction
The vibration standard of the electric motor is different bearings, SPM Technical Information, Sweden: SPM Instruments AB.
depending on the power of motor. However, the normal [9] D. E. Butler, “The shock pulse method for the detection of damaged
velocity is between 1.5 and 5 on the whole. If the velocity rolling bearing,” NDT Int. vol. 6, pp. 92–95, 1973.
exceeds 5, the bearing can be considered as faulty bearing. [10] L. Morando, “Measuring Shock Pulse,” SPM Instrument Inc, 2005.
Therefore, in comparison with the vibration standard of the [11] L. E. Morando, “Measuring shock pulses is ideal for bearing condition
motor, we can found that the velocity values at the horizontal monitoring,” Pulp Paper, vol. 62, pp. 96–98, 1988.
function of the motor load and non-load side exceed the normal [12] N. Tandon and K. S. Kumar, “Detection of defects at different locations
value (marked with red). So the journal bearings of motor are in ball bearings by vibration and shock pulse monitoring,” Noise and
faulty, which is consistent with the analysis results of vibration Worldwide, vol. 34, no. 3, pp. 9-16, 2003.
frequency spectrum. [13] Z. Li, Z. J. He, Y. Y. Zi and X. F. Chen, “Bearing condition monitoring
based on shock pulse method and improved redundant lifting scheme,”
Mathematics and Computers in Simulation, vol. 79, pp. 318-338, 2008.
IV. CONCLUSION [14] F. Masaru and K. Minoru, “SPM shock pulse method for diagnostic
system of rotating roll bearing,” Japan Tappi Journal, vol. 57, no. 4, pp.
Bearings are one of the most widely used rotating machine 52-57, 2003.
elements in industrial applications. SPM has been proven as an

698