An overview of DREAM/DCFIL bearing measurement systems


What DREAM does

Basic features of DREAM and DCFIL bearing test systems

Basic methods used by DREAM for diagnostics and prediction

The method used is based  on envelope spectrum analysis of random vibrations  created by friction forces in the bearings. In a non-defective bearing, the friction force is constant in time and the exited high frequency vibration is a stationary, random process. During the rotation of a defective bearing, the friction force and the random vibration become amplitude modulated with a period, T, and depth, m, because of a periodic change in the friction coefficient or pressure on the race surface. Shock pulses are considered to be  a particular case of amplitude modulated friction forces whose appearance results from a periodic change in the friction coefficient. Thus, in the case of the defects, harmonic components appear in the envelope spectrum which are absent in envelope spectrum of a non-defective bearing.  The type of defect can be identified by the harmonic component frequencies, the defect depth by the  difference L between the harmonic Lh and the random Lr components, or in the time function by T and  m respectively (see below).

12 types of bearing defects (below)

No. Type of defect Main frequencies of modulation Comments
1 Revolution around outer (fixed) race fr  
2 Nonuniform radial tension 2fr  
3 Misalignment of outer race 2fout  
4 Wear of outer race fout  
5 Cavities (cracks) on outer race kfout; k=1,2,3...  
6 Wear of inner race kfr; k=1,2,3... the decay of amplitudes with increase of k
7 Cavities (cracks) on inner race kfin  
8 Wear of cage and rolling elements fc or (fr-fc)  
9 Cavities, spallings on rolling elements kfrol  
10 Defects of several bearing surfaces fout + fin = zfr or 
fout + fr 
no fout-fr or (fr-fc)(z-1) 
11 Slip of race kfr k=1,2,3... no other frequencies of modulation
12 Defects of lubrication    increase of high frequency vibration level 

 Defect Frequency Equations (below)

- rotor frequency
- cage frequency
- rolling element rotation frequency
- rolling element pass frequency, outer race
- rolling element pass frequency, inner race
z  - number of rolling elements
drol - diameter of rolling element
dc - diameter of cage
- contact angle between balls and outer race

Defect type identification (above) and defect level characterization (below).

With one single  measurement, this method can give an accurate prediction for non-failure operation of a bearing for up to 20% of the specified bearing’s service life, as was proven by experimental work in a 15  year period with bearings having dimensions from 20  mm to 2 m in machines with rotating speeds from 10 to 10000 RPM. The program was  designed to make the process of  diagnostics and prediction automatically, increasing bearing diagnostic productivity and effectiveness. The combined use of Data Physics or Vibration Specialties Corporation hardware and DREAM/DCFIL software gives even an unqualified operator the following abilities:

The input data for the program are:

The output data for the program are:

What DCFIL does.

 Much of the superior performance of DREAM may can be attributed to the methods it uses to control the content of the envelope spectrum data. DREAM controls the relative sensitivities of bearing's friction forces and impact forces and the effective bandwidth of the measurement system by selecting specific bandpass filters applied to the input data before demodulation using criteria based on bearing's dimensions and shaft rotating frequencies. DCFIL generates the filters specified by DREAM, processes the input data, automatically removes interfering signals from the data before the data is demodulated, and then does the remaining demodulation and other signal processing necessary for DREAM. In most cases, DCFIL also controls the measurement hardware, although it may also be used to process previously collected time domain data.

Typical filter specified by DREAM and implemented with DCFIL.
Typical filter specified by DREAM and implemented with DCFIL and automatically modified to remove interference from other machine sources.

For more detailed information, see our on-line technical library