Fusion of the Dimensionless Parameters and
Filtering Methods in Rotating Machinery Fault
Diagnosis
Jianbin Xiong*, Qinhua Zhang, Guoxi Sun, Zhiping Peng, and Qiong Liang
Guangdong Provincial Key Lab. of Petrochemical Equipment Fault Diagnosis, Maoming, China
School of Computer and Electronic Information, Guangdong University of Petrochemical Technology, Maoming, China
*Corresponding author, Email: 276158903@qq.com, fenglianren@vip.tom.com; 158011382@qq.com;
417863130@qq.com;
Abstract—For the problem of large dimensionless index
fluctuations in rotating machinery complex fault and that
the corresponding scope is difficult to determine. In this
paper proposes a rotating machinery complex fault method
that combined dimensionless and the least squares method
filtering. This method implementation filtering and
determine the scope of the dimensionless index. By doing
experiments with 8 kinds of bearing failure data of
petrochemical rotary sets, comparing four filtering methods,
the scope of the dimensionless index was established, and
the text combined dimensionless index respectively with
Kalman (EKF), the weighted average, moving average, the
least squares method filtering.
Index Terms—Dimensionless Parameters; Kalman Filter;
Combination Faults; Fault Diagnosis; Moving Average
Filter; Weight Filter
I. INTRODUCTION
That large engineering system complexity continues to
increase requires higher safety and reliability of the
system. Through the analyzing system health status, fault
diagnosis technology determine the type of failure, for the
timely and effective maintenance and health systems
management provides a scientific basis, so that in the
field of aviation, aerospace and other need higher security
requirement, has a good application prospect [1, 2].
Particularly, rotating machinery and equipment (such as
rotation bearings, turbines, compressors, fans, etc.) is the
key equipment in petroleum, chemical, metallurgy,
machinery manufacturing, aerospace, and other important
engineering filed. Therefore, the study of such equipment
fault diagnosis method has been a hot topic in this field.
In rotating machinery fault diagnosis usually use the
time domain or frequency domain analysis of vibration
monitoring data for fault diagnosis [2-5]. Rotating
machinery in the event of a failure, however, vibration
monitoring signals tend to have a large number of
non-linear, random, non-ergodic information, and bring
great difficulty in fault signal analysis [6]. Considering
the time-domain signal of vibration is the most basic and
original signal, if failure characteristics can be extracted
directly from the time-domain signal, and analyze fault
diagnosis, so that maintain the basic characteristics of the
signal will be very beneficial [1-2]. In the time domain
analysis, the probability density function of vibration
signals can better reflect the fault information. Through
the probability density function of the vibration signal, it
has been derived dimensional index (such as the mean
and RMS values, etc.) and dimensionless index (such as
waveform, margin index, pulse, etc.) in the amplitude
domain [1-2, 6-7]. In practice, although a dimensional
index is sensitive to the fault characteristics, its value will
increase with the development of the fault, but also
because working conditions (such as load, speed, etc.)
changes, it is easily affected by interference, performance
is not stable enough [1]. By contrast, the dimensionless
index is not sensitive to the disturbance of vibration
monitoring signal, performance is stable. In particular,
these dimensionless index are not sensitive to the change
of amplitude and frequency of the signal, namely, it has
little relationship with working conditions of the machine
[1-3, 5-7]. Therefore, the dimensionless index has been
widely used in fault diagnosis of rotating machinery. In
dimensionless index, pulse index and kurtosis index is
more sensitive to impact type fault, especially in the early
failure, the large amplitude of the pulse is less, other
parameter values increase is not much, but kurtosis index
and pulse index rise faster, so that failure of the range is
larger. It is difficult to determine the scope of the
composite fault interval [1-3, 5-7].
In order to reduce the error, determine the scope of the
dimensionless index, and narrow the scope of the
machinery recombination fault interval, first excluding
outliers of the dimensionless index, and then filtering.
There are many ways to achieve signal filtering, Kalman
filtering is usually used in aviation and aerospace aspects,
for example, against the data collected by marine
dynamic positioning multiple sensors in real-time
fluctuation, causing controller move frequently, resulting
in boat actuators adjust frequently, and increasing its
mechanical wear, Jianbin Xiong proposes a method to
collect data online by DPS multi-sensor and filtering
based on OPC technology, achieving ship signal filtering
indoor, but the method is easy to diverge [9]. In order to
solve the problem of Kalman filtering divergence, Wang
Qinruo proposed a blend adaptive Kalman filter
JOURNAL OF NETWORKS, VOL. 9, NO. 5, MAY 2014
doi:10.4304/jnw.9.5.1201-1207