Introduction
In all moving parts of a machine, bearing failure forms one of the major causes of industrial breakdown for machines rotating at high or low speeds. The following paper provides a comparative analysis son the application of acoustic emission (EA) technique for condition monitoring of bearing element. The use of advanced signal processing and pattern recognition approach was established to monitor bearings conditions through automatic detection and diagnosis of localized defects in bearings. A short-time signal processing technique extracts two normalized and dimensionless features that establish linear discriminant functions used to detect defects on rollers of bearings and the outer race (Pandya, Upadhyay, and Harsha 2013).
The main aim of the paper is to analyze faults of bearings using acoustic emission approach in order to determine AE characteristics of bearing. In order to come up with an effective analysis, vibration frequencies of bearing using AE signal, time-domain approach, and short-time rms will be used. The AE signal will then be analyzed using acceptable signal standard parameters to explore the rolling element bearing faults. AE testing in bearing is significantly advanced state-of-the-art technology that all industries should implement. In addition, the process produces better results compared to other methods that use vibrations to detect faults in bearings (He, Zhang, & Friswell 2009).
Techniques to measure and analyze the waveform of AE
Many researchers have tried various techniques to measure and analyze waveform of AE and come up with several signal processing techniques for bearing conditioning monitoring using AE. For this analysis only two techniques will be used, the time-domain approach and the frequency-domain approach.
Time-domain approach, rms, peak amplitude, energy
In time-domain approach a series of time signals are used to perform fault and failure diagnosis through analyzing acoustic data obtained from the test equipment. To investigate random characteristics of a physical system statistical methods are mostly used in the time-domain approach. The process effectively analysis the data obtained and draw meaningful conclusions. The use of overall root-mean-square (RMS) level forms the simplest method of analyzing data collected from the test equipment. Another method commonly used is the crest factor which is the ratio of peak value to rms. The above methods are mostly used in the diagnosis of localized defects. On the other hand, probability density is also used in bearing defect detection (Grosse & Ohtsu 2008).
The main focus of this approach is to determine the relationship between the rotational speed of a bearing and achieved parameters. RMS is suited for steady-state signals. In addition amplitude, A, is analyzed in the time domain approach. Amplitude in this experiment is the highest measured voltage measures in decibels (DB). Amplitude is a significant parameter in AE because it makes signals detectable. If a signal shows amplitude below the operator-defined, the experiment will not yield expected results. The maximum amplitude can be seen from the signal display (Fakhfakh 2012).
AE energy is the area envelope of the amplitude-time curve. AE energy depends on the viscosity and quantity value of the lubricant (Fakhfakh 2012).
Frequency-domain approach
Frequency domain is the treatment of signals expressed as a function of frequency using time domain signal. The most important analysis of a rotating element in a machine is conducted using the frequency domain and the widely used approach for bearing defect detection. Frequency domain approach is used because some other approaches such as signature spectral comparisons cannot effectively detect damages found in rolling element bearings. Dominant signals coming from other rotating elements, like shafts and pulleys, overcome the energy produced by bearing elements. Frequency domain approach involves a high-energy noise filtering operation that gets rid of dominant low frequency components from the expected signal. The signal is then rectified, demodulated, and the carrier of high frequency is eliminated by low-pass filter. The frequency spectrum then displays the processed signal that shows isolated bearing defects (Abdullah, Al-Ghamd & Mba 2006).
The frequency-domain approach data acquisition process is carried out at low speed in the range 250-1500 revolutions per minute (RPM) at intervals of 250 rpm. The data is made more realistic by acquiring 5 sets of data in a single test. The influence of varying loads and rotating speeds on AE equipment produces different amplitude. Generally, the process of acquiring AE signals in testing faults in a bearing element can be summarized into four steps:
- Capturing AE signals
- Performing band pass filtering process to acquire the desired signal rates and getting rid of unwanted vibrations
- Collecting sets of measurements using a definite frequency band, and
- Extracting time-frequency domain features through intrinsic mod function.
Conclusion
The advancement in technologies has made work easier and faster to perform. The fault diagnosis of a ball bearing can be conducted within 10 minutes using acoustic emission analysis as discussed above. From the analysis, two main approached of determining faults in bearings have been discussed. The choice of EA measurement technique depends on the speed of rotation and the type of lubricant used in the bearing. Using the test equipment correctly and following the laid down procedures leads to better results that can be relied on. On the other hand, AE testing in bearing is the most advanced method used by top engineers in detecting faults in bearings.
References list
ABDULLAH, M., AL-GHAMD & MBA, D. (2006). “A comparative experimental study on the
use of acoustic emission and vibration analysis for bearing defect identification and estimation of defect size”. Mechanical Systems and Signal Processing, 20, 1537–1571.
FAKHFAKH, T. (2012). Condition monitoring of machinery in non-stationary operations
proceedings of the Second International Conference "Condition Monitoring of Machinery in Non-Stationnary [ie. Non-Stationary] Operations" CMMNO'2012. Berlin: Springer.
GROSSE, C. U., & OHTSU, M. (2008). Acoustic emission testing [basics for research,
applications in civil engineering]. Berlin, Springer.
HE, YONGYONG., ZHANG, XINMING. & FRISWELL I. MICHAEL. (2013). “Defect
diagnosis for rolling element bearings using acoustic emission,” Journal of Vibrations and Acoustics, 131(2), 1-5.
PANDYA, D.H., UPADHYAY, S.H. & HARSHA S.P. (2013). “Fault diagnosis of rolling
element bearing with intrinsic mode function of acoustic emission data using APF-KNN,” Expert Systems with Application, 40 (1), 4137-4145.
