Exhaust fan bearing defects and vibration frequency solutions analysis

Source: www.sourceindex.com

Grease hardening and insufficient lubrication can be caused by mixing the grease with different bases and thus should be completely avoided. To change from the lithium-based grease used in factory manufacture of exhaust fan bearings, continually add the new grease to the bearing until only the new grease built with another base such as sodium or synthetic base completely replaces the factory lithium based grease.

The galvanized steel base bracket of the exhaust fan is welded to provide the strength and support needed for it’s cleaning, preventing spillage and loss into the building. A factory-covered aluminum grease plug is welded to the fan housing to ensure that it does not spill oil and grease.

Low oil levels do not provide adequate lubrication. The level meters are used to control the oil level must be adjusted to read correctly. Since it has been indicated when the fan is on, it’s better to check the oil levels when the fan is off.

Maintenance of the bearing while ensuring proper lubrication ensures the long life of the bearing. To prevent damage to the bearing a sufficient supply of lubrication should be ensured to avoid metal to metal contact between the bearing and the fan.

Static or circulating oil should only be used with separation High speed or high temperature cushion bearings applications where grease is inadequate. Lubricated with oil bearings require more rigorous maintenance.

During testing air inside the chamber should be avoided at all costs and at all times. High oil levels can prevent the bearing from venting the air inside rolling while the bearing warms up in revolutions. The accumulated pressure can blow all the oil on landing in a few minutes.

CONDITION MONITORING

Identification of fan bearing defaults involved the collection of fault frequencies, which include the harmonics and revolutions per minute, fundamental train frequency (FTF), the ball spin frequency (BSF), ball passing frequency-outer race (BPFO), and the ball passing frequency inner race (BPFI).

Data analysis of the ball passing frequency-outer race (BPFO), and the ball passing frequency inner race (BPFI) in conjunction with the revolutions per minute was applied to identify the fan-bearing defect.

EXHAUST FAN BEARING AND ITS USES

The fan bearing chosen for this research study is the

UCP204-12-AH-SP4 – Bearing 3/4″ pillow block bearing with cast housing and grease fitting. (8B0750) 1/4-28 x 1/8 Socket Set Screw (2@62°). Serrated Tip Set Screw With Loctite Patch. Set Screw Torque = 85 inch lbs. Grease fitting on opposite side of standard.

The UCP204-12-AH-SP4 bearing has been used in many applications including those mentioned below:

DETECTION

Exhaust fans are built to work efficiently. Evident failure in new and old exhaust fan bearings is testimony enough that manufacture and reliance on long life usage is nowhere near perfection. Detecting the causes of such failures or the problems with the bearing is important in predicting, gauging and guarding against eminent loss in usage of the exhaust fan bearing.

A series of high-energy pulses at a rate equal to the frequency of passing the ball in relation to the inner raceway result from a small fault on the inner raceway. Because of the inner ring rotating, the defect will enter and leave the load area causing a variation in contact raceway-bearing raceway force, so deflections.

Contamination is a very common source of bearing damage and premature failure and is due to the entry of strange element either because of improper treatment or during operation. Naturally the size of the vibration caused by the contamination varies and in the beginning steps can be difficult to recognize, but depends very much on the type and nature of impurities. Contamination can cause wear and tear damage to the bearing contact surfaces and generate vibrations over a wide range of frequencies.

Other causes of failure exhaust fan bearings include rusting of the bearing result from existence of moisture, broken bearings, low quality wrapping, acidic wear and tear and high temperatures. Buildup of metal in front of the rollers resulting in flaking, contamination due to defective bearing seals, arcing and burning at the point between the races and the rolling elements in electric currency damage of the bearing, mis-alignment and improper mounting and existence of foreign matter in the exhaust fan bearing are reasons, which also cause bearing damage and defects.

Accelerometers placed at keys positions on the motor and fan bearings are used to measure vibrations caused by the exhaust fan bearing during live factory testing and diagnosis. Because the bearings are the support part, mechanical train accelerometers should be placed in the entrance and exit of rolling bearings for vibration measurement levels.

Permanent accelerometers are set Fan and fan bearings Vibration sensors must be placed radial (vertical and horizontal) and axial positions on the motor and exhaust fan bearings. This provides the best recognition to every vibration components, including bearing vibration, imbalance failure, electrical disturbance, wings flow (aerodynamic disturbances) and frequency bands.

The piezoelectric accelerometer is widely regarded as a standard vibration sensor for vibration measurement of the machine. System configuration the piezoelectric crystal and the seismic mass depend on the frequency range of the sensor desired.

The prufteck vibroexpert does not only offers frequency information, but also general amplitudes for data analysis. It is useful for tendency or comparison measurements on similar machines but for diagnosis problems on machines. The other two large vibration instruments are tunable filters and FFT (Fast Fourier Transform) analysts. Analyzer is a default because analysis is a human function. One electronic analyzer “analyzed”. It only measures and display electrical signals.

The electrical signals of Accelerometers and speed transducers are very small AC voltages, typically millivolts. Therefore, the tunable filters and FFT instruments are nothing but fancy AC voltmeters with a frequency display axis. Any of these two large vibration-measuring instruments, of any one manufacturer, can be used effectively to diagnose machine vibrations.

The balancing and alignment method can be used to correct the blower vibration problems. This entails disassembly, manual sight inspection and the reassembling the bearing. It might result in replacement of the fan bearing or the defective parts, balancing and realignment and lubrication.

Clear and repeatable vibration measurements and data capture keys are not difficult to understand and set up. They are based on sound reason, care, organization and consistency. Precise vibration measurement requires the selection of the best measuring stations fan assembly. It also requires the best vibration converter specific frequencies or frequencies of general interest.

Last years, piezoelectric accelerometers have become the most widely used sensor type because of their general frequency characteristics, size, reliability and overall stability. Unless low frequency vibration measurement is required, the unusually low amplitude accelerometer is usually the best choice for accurate data.

The latest vibration measuring devices have the option of selecting a transition, speed or acceleration vibration measurement parameters, regardless of the type of sensor used. It is desirable, for the sake of accuracy, to use the parameter that provides the best (lowest) interest rate.

It turns out that speed is usually the best choice parameter for measuring vibration of the machine. The most common exception to this rule is using higher frequency accelerations to detect problems with certain components such as roller bearings.

How to use vibration converters to the measuring point is critical aspect vibration measurement accuracy. The stronger the sensor is connected the machine is your best answer. Machine vibration has different categories of causes that we’ve discovered after so many repairs, though it is now useful to check them to get more confidence about diagnosis, B.A Kardile, 2012.

CLASSIFICATION

There are four phases of classifying exhaust fan bearing defects. These phases include the phase one of wear and tear of the fan bearing where the possibility of complete failure of the bearing is quite low. At this phase, further monitoring for defect extents can be carried out on the fan bearing.

In the second phase most damage to the fan bearing can be seen with the naked eye and at the same time it can be sensed in the increased quantity and amplitude of fault frequency harmonics. In this second phase, outer race fault frequency is detected in the first bearing band. Inner race fault frequency is detected in second bearing band. Repair should continue at scheduled time intervals

In the third phase three, the ball spin frequencies appear in the outer bearing band. Harmonics of bearing fault frequencies should be easy to detect in a velocity spectrum. The possibility of catastrophic failure is imminent and repair should be more periodical.

In the fourth phase, the discrete frequency indicator will shows the extent of fundamental train frequency FTF (cage frequency) indicates that the bearings life is almost over or already over.

RESULTS

Analytical tests conducted using the chosen exhaust fan bearing data yielded various results, which enabled the process of decision-making. The determination of defect and lifetime remaining for a fan bearing enabled the compilation of specific benchmark predictors. In this section of the report follows the discussion of the results from data analysis.

TABLE 1 : HARMONIC AND FREQUENCY TABLE AT 179.99^O WITH 50,000 REVOLUTIONS PER MINUTE FOR UCP204-12-AH-SP4 – Bearing 3/4″ pillow block bearing

TABLE 2 HARMONIC AND FREQUNCY TABLE AT 179.99^O WITH 50,000 REVOLUTIONS PER MINUTE FOR UCP204-12-AH-SP4 – Bearing 3/4″ pillow block bearing

SOURCE: www.sourceindex.com

VIBRATION STATISTICAL PARAMETER VELOCITY VIBRATION LEVELS

FIGURE 6 INTERVAL PLOT FOR THE CONFIDENCE LEVEL BPFO RPM (VELOCITY) TO FREQUENCY VIBRATION OUTER RACE

SOURCE: Author

TABLE 3 BEARING FAULT FREQUENCIES AT VARIOUS REVOLUTIONS PER MINUTE

FTFI FTFO BSF BPFO BPFI RPM

0.385 0.615 2. 0614 3.08 4.92 1

48.138 76.862 257.6750 385.13 614.88 125000

96.275 153.725 515.3600 770.25 1229.75 250000

192.550 307.450 1030.7000 1540.50 2495.50 500000

Results of multiplying RPM with all frequencies to get velocities

FTFI (RPM) FTFO (RPM) BPFO RPM BPFI RPM

1 1 3 5

6017188 9607813 48140625 76859375

24068750 38431250 192562500 307437500

96275000 153725000 770250000 1247750000

SOURCE: Author

FIGURE 7 SCATTERPLOT OF BPFO RPM vs BPFI

The scatter plot for BPFO RPM (or velocity) v/s frequency of vibration outer race defect vibration BPFI shows anomaly in the fan bearing because the data is not linearly co-related, but significant variances are recorded in the data points plotted.

SOURCE: Author

FIGURE 8 INDIVIDUAL VALUE PLOTS VELOCITY AND INNER RACE (BPFI) AND OUTER RACE (BPFO) VIBRATION FERQUENCIES

Source: Author

The plot shows deviations from the densely located data thus indicating anomalies in both defect analyzing fault frequencies and therefore possible defects in the exhaust fan bearing. Figure 9 below clearly illustrates the deviations.

FIGURE 9

Source: Author

FIGURE 10 INDIVIDUAL VALUE PLOTS VELOCITY AND OUTER RACE (BPFO) VIBRATION FERQUENCIES LARGE DEFECT AT 360^O BEARING CONTACT

Source: Author

Figure 10 shows that more points vary at wider gaps from the densely distributed data indicating more defect possibilities at a contact angle of fan bearing of 360^O as opposed to the distribution in figure 9, which shows less possibility of defects.

The possible causes would be looseness of the fan bearing due to the extent of defect causing variations in the frequency vibrations or a technical error in mounting the fan bearing during lab testing.

THE K-NEAREST NEIGHBOR (KNN) HEALTH ALGORITHM

The KNN classifier requires a metric d and a positive integer K. In KNN Euclidean distance computes distances in multidimensional input space. The Euclidean distance between point p and q is the length of the line between them. In Cartesian coordinates, if pi and qi are two points in Euclidean n-space, then the distance from p to q is given by dE = (pi – qi)². Position of training samples and input sample can be visualized on 2D and 3D Cartesian coordinates.

BALL PASS FREQUENCY OUTER RACE BPFO AND INNER RACE BPFI DATA 179 ^O

VELOCITY BSF_1 BPFO_1

50000 1030704 2.459,499.2636

100000 2061407 4,918,998.5271

150000 3092111 7,378,497.7907

200000 4122814 9.837,997.0542

250000 5153518 12,297,496.3178

300000 6184222 14,756,995.5813

350000 7214925 17,216,494.8449

400000 8245629 19.675,994.1084

450000 9276333 22,135,493.3720

Source: Author

BALL PASS FREQUENCY OUTER RACE BPFO AND INNER RACE BPFI DATA 360^O

Source: Author

BPFO RPM REVOLUTIONS

1.540,500.7364 50000 50000

3,081,001.4728 50000 100000

4,621,502.2091 50000 150000

6,162,002.9455 50000 200000

7,702,503.6819 50000 250000

9,243,004.4183 50000 300000

10,783,505.1546 50000 350000

12,324,005.8910 50000 400000

13.864,506.6274 50000 450000

15,405,007.3638 50000 500000

16.945,508.1001 50000 550000

18,486,008.8365 50000 600000

20,026,509.5729 50000 650000

21,567,010.3093 50000 700000

23,107,511.0457 50000 750000

24.648.011.7820 50000 800000

26,188,512.5184 50000 850000

27,729,013.2548 50000 900000

29,269,513.9912 50000 950000

30.810,014.7275 50000 1000000

32,350,515.4639 50000 1050000

33.891,016.2003 50000 1100000

35,431,516.9367 50000 1150000

36.972,017.6730 50000 1200000

The distance between incoming velocities, frequency is compared to the distance between harmonics and revolutions per minute for the two scenarios of Ball Pass frequency Outer BPFO outer race defects.

The distance is greater for the 360 degrees bearing data as compared to the 179.99 degrees frequencies, thus revealing more defects in the 360-degree dimension.

CONCLUSION

Vibration frequency analysis is an effective method of ascertaining exhaust fan bearing defects. It is a meticulous technique, which when coupled with the K – nearest neighbor analysis combines to form a formidable force in exhaust fan bearing defect detection through analysis of vibration frequency data and finding of requisite solutions for enhanced bearing health and maintenance.

REFERENCES