12-11-2012, 03:13 PM
FAULT DIAGNOSIS AND SIMULATION OF ROLLING ELEMENT BEARING USING CONDITION MONITORING
FAULT DIAGNOSIS.docx (Size: 253.9 KB / Downloads: 162)
ABSTRACT
Rolling element bearings find widespread domestic and industrial application as it is an important factor in failure of rotating machines and therefore bearings are the one which are exposed the most towards getting damaged and failure. In industrial applications, these bearings are considered as a critical mechanical components and a defect in such a bearing, unless detected in time, causes malfunction and may even lead to catastrophic failure of machinery which results in significant time and economic loss. These types of failures might take place during the manufacturing process and therefore it is important to review the problem and monitor the condition of roller bearings so that the details of failure would occur before any harsh consequences take place. Therefore an early detection and indication is necessary for the safety and reliability of machine. This paper describes various vibration monitoring techniques suitable to analyse the defect in bearing. By performing this test, these techniques would reveal information about the progressing faults. From the different maintenance techniques, conditioning monitoring which is one of the techniques is highlighted. It uses the vibration having high frequencies which are generated from faulty bearing, is therefore investigated and compared.
From one of the conditioning monitoring technique vibration analysis method has been employed and therefore utilised as a medium to fulfil the aim. Computer oriented programming software MATLAB has been used for finding faulty frequency at inner race. Also the vibration signatures caused due to damages at inner race of bearing are examined. The result indicates that faulty bearing has a strong effect on vibration spectrum. This paper therefore reveals comparison between frequencies and time domain signal from vibration analysis and also it is been verified within the Matlab programming. Overall this paperhas demonstrated that vibration analysis and MATLAB programming techniques are useful in detecting the problems in roller bearings.
Introduction:
In 21st century, managing the industries has become one of the challenging tasks due to change in management structure, increased global competition, intense change in technology, reliability, health and safety, consumer demands towards quality and environmental considerations (Rao,1996 p.1). Taking into consideration all the above factors there is great chance to improve the opportunities as well as strategic plans and therefore make the most benefits of the modern manufacturing techniques and methods. Advanced manufacturing methods and techniques, quality of human resources has greatly persuaded the productivity in manufacturing industry. Rao (1996) stated that manufacturing companies in the UK pay out three times as much every year in replacing the machinery and maintaining the existing plant. As there have been several sectors in manufacturing affected by the different issues related with the rotating machineries, one of the major issues in the manufacturing has been early detection of faults which has result in the unplanned breakdowns, maintenance cost and so it has been disaster failure in machinery or process (Shikari, 2004). To overcome and therefore prevent these failures, different techniques of maintenance management like corrective maintenance, breakdown maintenance, preventive maintenance, time based maintenance and condition based maintenance are widely utilised (Venkatesh, 2007, p.4). All these techniques are viable in some and the other way for the different failure machinery for manufacturing industry. Above all, condition based maintenance is one of the best method in early detection of faults in rolling element bearing. Usually these faults in the rolling bearing element occurs due to corrosion, overheating, excessive load, lubrication failure, misalignment, tight fits, normal fatigue failure and contamination
Vibration Analysis
Vibration analysis is most powerful tool for fault diagnosis of bearing. Manufacturing industries are using rolling contact bearings almost in every rotating machine. Accuracy of operation of bearing is related with elements like housing, shafts and nuts. Some of the bearings fail earlier in service because of the poor lubrication, tight fitting, loose fitting, contamination and misalignment. Above aspects increase the bearing vibrations so CM is necessary to detect the failure of the bearing. Vibration analysis helps to calculate the amplitude of the bearing.
Characteristic frequencies of the bearing:
The vibration analysis technique gives the precise and early information about the failure of bearing. According to Tondon and Choudhary (1992) faults in bearing (inner race, outer race and cage fault) produces the particular defective frequencies which is calculated by using the following equations.
Literature review
Condition Monitoring:
Since last two decades industries have spent huge amount on conditioning monitoring to produce efficient instrumentation for minimise the various problem. According to Hutton (1996) conditioning monitoring mostly focuses on the vibration data including sample of lubricant, temperature readings and measurement of shocks from rolling element bearing defects. Beebe (2004) defined conditioning monitoring as “conditioning monitoring on or off-line is a type of maintenance inspection where an operational asset is monitored and the data obtained analysed to detect signs of degradation, diagnose cause of faults, and predict for how long it can be safely or economically run”. There are several benefits of conditioning monitoring which potentially effects on the improved productivity, maintenance cost and increased plant availability (Mathew and Alfredson, 1984). In order to analyse the conditioning monitoring two different factors need to be considered, those factors are technical issues like measurement and analysis and organizational and environmental issues (Rao, 1996). This paper focuses on the technical issues of conditioning monitoring which includes the vibration analysis and its verification by using software programming. According to (Babson and Renwick, 1985) conditioning monitoring is the method for establishing the functionality and life of the machine or element which helps to determine the future failure. Conditioning monitoring includes the constant or periodical set of data, analysis of data and its diagnosis. Conditioning monitoring is nothing but the predictive maintenance in which prediction of failure is evaluated.
Vibration analysis:
Rao (1996) stated that vibration analysis is the most tangible and established technique in conditioning monitoring. McFadden and Smith, (1984) stated that vibration analysis has been used comprehensively in diagnosis of bearing in rotating machine. As all the machines vibrate, the link between the vibration and those machine conditions are measured and therefore its outcome is easily interpreted. It is also the most popular and a hopeful procedure in detecting developing faults in bearing. To identify the faults of bearings, the received vibrating signals are processed by different methods. More a less, analysing the vibration signals, needs a good exposure and knowledge in interpreting the vibration signals for the exact identification of faults and their respective location. This analysis is based on a fact that all the machines or systems produce vibrations. When the machine is working properly and running smoothly, it still produces small and constant vibration but when there is a fault developed and if there is a slight change in the dynamic process of the machine, changes in the vibration scale can be observed. The vibrating signals which are produced from a bearing contain precise information about the condition of bearing. In case of single defect on bearing components, many researchers have used vibration analysis as the premium technique for the diagnosis of faults on rolling element bearing. Rolling element bearing is a bearing which carries load by placing the round elements between the two pieces.
Time domain analysis
Karimi (2008) stated that almost all the vibration analysis was done in the time domain before the spectral analyser was available. The time domain analysis is nothing but display or analysis of the vibration data as a function of time. To detect the fault, the time domain method analyzes phase information and amplitude of the vibration time signals (Wang, 1998). Changes in vibration signals due to faults were detected by studying the time domain waveform using equipments like vibrographs, oscilloscope or oscillographs.
Frequency domain analysis
Frequency domain analysis is the classical bearing diagnostic technique also known as spectral analysis. Karimi (2006, pp.15) stated that frequency domain analysis method is more reliable and most sensitive than time domain analysis method. The vibration data is analysed as a function of frequency by frequency domain analysis. The spectrum of faulty bearing and the bearing in good condition is compared and its difference is used in detecting faults on bearing (Taylor, 1980). Obtaining narrowband spectra easily and more efficiently is mostly done by using Fast Fourier Transform (FFT). Fast Fourier transform algorithm is used to process time domain vibration signal into frequency domain. In other words, frequency spectrum is achieved as frequency domain method mainly uses numerical fast Fourier transform to the vibration signal. The proceeding analysis is carried out conventionally using vibration amplitude and power spectra. According to Li et.al. (1996) the damage of elements can be identified using the difference of power spectral density of signal which occurs due to fault of bearing. On the contrary to spectral comparison, various types of spectral trending can be utilised to give signs of rate of fault progression. Su and Lin (1992) researched providing details of expected spectral differences related with various bearing faults.
Conclusion
In this paper, condition monitoring and its different techniques of determining defects and faults were highlighted. An overall review of study on vibration measurement techniques for detection of faults in rolling bearing element was focused and therefore different vibration measurement methods were highlighted. Different researches on detection of faults are being done for ages. Although literature is available for detection of both localized and distributed faults, this paper revealed the proper hopeful or alternative technique to detect the fault more precisely and experimentally. For the early fault detection methods like vibration analysis, acoustic analysis, oil debris analysis, Thermography and corrosion analysis were taken into consideration. Vibrations in time domain and frequency domain were studied as it can be measured through parameters like skewness, Kurtosis, crest factor, RMS (root mean square value), probability density and pulse software. To find more precise results in finding those faults, vibration analysis and MATLAB software were merely focussed theoretically. Signal data was obtained for faulty bearings with induced inner race fault at 0.5mm, 1500rpm, In the data analysis, plots obtained for each test conditions are compared i.e. different frequencies obtained from Fast Fourier Transform (FFT) analyzer.