Fluid Film Bearing Protection Due to Loss of Lubrication

Alexander M. Tomsick - P.E. Director of Engineering

Case Study - Abstract

Given the wide usage of fluid film bearings in rotating and reciprocating equipment worldwide and the extremely short time to failure of fluid film bearings, the technology used to monitor fluid film bearing health must respond quickly to the start of a failure. This paper compares the change in response of bearing dynamic strain measurements and relative shaft displacement measurements to a loss of lubricating oil pressure supplied to plain bore bearings in a laboratory test rotor. This paper aims to provide results on the efficacy of both technologies tested in responding to the loss of oil pressure in the bearings and compare their usage as a protection measure to shut down equipment before a catastrophic failure.

Hydrodynamic Bearing Background

Fluid film bearings are widely utilized in industrial rotating equipment due to their many beneficial characteristics. These benefits include their ability to tolerate higher operating speeds and loads, as well as having a much longer theoretical life when compared to rolling element bearings. While there are many benefits to the use of fluid film bearings, fluid film bearings have an extremely short failure interval once a failure has begun. For this reason, protection systems are typically installed on highly critical rotating equipment with fluid film bearings which can shut down the equipment based on measured vibration and temperature values being compared to predetermined limits.

Figure 1: Plain Bore Fluid Film Bearing

A plain bore hydrodynamic bearing is the most basic design of a fluid film bearing. Figure 1 shows an example of a plain bore bearing. As the shaft rotates in a fluid film bearing, oil in the bearing is dragged under the rotating shaft into a reducing clearance. This reduction in clearance leads to an increase in oil pressure before the location with the minimum oil film thickness. This high-pressure section produces the lift required to support the shaft as it rotates and lifts the shaft off the bearing surface. The oil continues to be dragged by the shaft as it rotates. As the oil leaves the location of minimum oil film thickness and the clearance increases, the oil pressure decreases. This decrease in oil pressure increases the oil velocity accordingly.

Proximity Probes

Historically, proximity probes have been installed on critical rotating equipment to measure the position of the shaft inside of the fluid film bearing. Proximity probes can be used for both bearing health monitoring and protection. However, since proximity probes measure relative displacement between the probe tip and the shaft surface, shaft runout whether mechanical or electrical must be removed by properly preparing the probe target area. Any remaining mechanical or electrical runout must be removed by performing a runout compensation. If the shaft surface is not properly prepared and demagnetized and runout compensation is not performed the measured displacement from a proximity probe can be unrealistic due to significant runout obscuring the actual vibration signal. An additional negative to the use of proximity probes is the large costs required to install the probes in rotating equipment above and beyond the hardware costs.

Dynamic Strain Sensors

Given the significant costs associated with installing proximity probes, the use of proximity probes is typically reserved for the most critical equipment. However, there are many less critical pieces of rotating equipment seen throughout industrial facilities that utilize fluid film bearings but operate without any form of protection. Another limitation is some bearings such as the main bearings in large engines and reciprocating compressors do not have adequate room to mount proximity probes. The limitations of mounting space and cost were the reasons Voyager Dynamics first began researching the use of dynamic strain to monitor fluid film bearing health.

U.S. Patent No. 9,841,329 was awarded in 2017 for the use of strain-based measurements in the health monitoring of fluid film bearings. Additionally, U.S. Patent No. 10,684,193 was issued in 2020 for the use of strain-based measurements to diagnose common machinery faults including unbalance and misalignment.

Figure 2: Voyager FDS Sensor

The Voyager FDS Sensor is a dynamic strain sensor to be used in industrial environments to monitor the health of fluid film bearing supporting rotating equipment. Just as with proximity probes, the two Voyager FDS sensors can be used for bearing health monitoring and protection. Figure 2 shows an image of the Voyager FDS sensor.

The Voyager FDS sensor is adhesively mounted to the face of the bearing with the longitudinal axis of the sensor pointing towards the center of the bearing. The ideal dynamic strain measurement locations depend greatly on the style of bearing that is used on the machine. This is because the highest strain amplitudes are expected at the locations of maximum oil pressure. On a plain bore bearing like the one utilized in this test, the locations of maximum oil pressure exist at the bottom of the bearing as shown in Figure 3. For this reason, the location of the highest bearing dynamic strain is expected to be located at the bottom dead center (BDC) of the bearing.

Figure 3: Location of high pressure within a plain bearing.

ll testing was performed on a laboratory test rotor to measure the change in response to a loss of lubricating oil. The rotor that was used was arranged in a center-hung configuration and weighed 220 lbs. Plain bore hydrodynamic bearings with a length-to-diameter ratio of 0.4 were used throughout all testing. Oil with a viscosity meeting an ISO grade 30 was used for lubrication.

Figure 4: Final shaft alignment compared to coupling manufacturer guidelines.

Before testing, the rotor was balanced to a G1.7 balance quality on a soft bearing balancing stand. After the rotor was balanced, the rotor was installed in the bearing supports and was aligned to a 5 hp motor. The rotor was coupled to the motor with a Dura-Flex flexible coupling. Figure 4 compares the remaining angular and offset misalignment measured during the alignment procedure to the coupling manufacturer guidelines.

Test Cases

Several different test cases were used to compare the effects of oil viscosity, bearing clearance, and rotor load on the overall RMS displacement and dynamic strain amplitudes. To measure shaft displacement, two proximity probes were mounted orthogonally on the inboard bearing of the rotor. To measure the dynamic strain experienced by the bearing one Voyager FDS sensor was mounted to the face of the inboard bearing below the shaft. Table 1 details the parameters set in each tested case.

Table 1: Set test parameters in each tested case.

To understand the effect oil viscosity has on shaft displacement and dynamic strain amplitudes, the temperature of the oil was varied from 70°F up to 130°F with an oil heater. To understand the effect on bearing clearance both a worn bearing with a total radial clearance of 0.005” and a new bearing with a total radial clearance of 0.003” were used during testing. The load on the bearings was varied by collecting data at a balance quality of G1.7 and collecting data with a significant static unbalance added to the rotor increasing the balance quality to a G14.75.

In all test cases, the overall amplitudes from the three transducers were trended at 0.25-second intervals before and after the flow of lubricating oil to the bearing was stopped. All test data was collected with the rotor operating at 1,798 RPM which is above the rotor’s first critical speed of 1,300 CPM.

Case 1

Figure 5: Case 1 - % overall amplitude change

Figure 5 shows the change in overall dynamic strain and shaft displacement measured at the inboard rotor bearing when the rotor was balanced to a G1.7 and cold oil was supplied to the high clearance bearings. In this condition, the overall dynamic strain amplitude measured at the BDC location on the bearing increased by 102% within one second after the oil supply was removed.

Also seen, the X-direction proximity probe measured only a 20% increase in overall shaft displacement within the same period followed by a sharp decrease in measured displacement to 20% below the starting amplitude. The Y-direction proximity probe did measure an increase in overall shaft displacement of nearly 40% but did not reach its maximum until twenty seconds after the oil supply was removed.

Case 2

Figure 6: Case 2 - % overall amplitude change

Figure 6 shows the change in overall amplitude for all three transducers with the same balance quality and bearing clearance that were used in Case 1, but the oil temperature was increased to 130°F. With the higher oil temperature and thus lower viscosity, the measured change in overall dynamic strain was higher than what was measured at the same location in Case 1.

The lower oil viscosity in Case 2, while maintaining a consistent load and speed as the Case 1 measurements would cause a decrease in the ZN/P ratio. This would lead to higher bearing friction, and a smaller area supporting the shaft. This would in turn lead to a higher bearing stress and strain.

Conversely, the overall displacement amplitudes measured by the proximity probes mounted on the bearing did not increase because of the oil supply being lost. This was especially noticeable in the Y-direction proximity probe data which showed no change in overall amplitude to the loss of the lubrication, while the X-direction proximity probe data continued to show a decrease in overall amplitudes.

Case 3

Figure 7: Case 3- % overall amplitude change

Figure 7 shows the change in overall vibration amplitudes with the same oil temperature and bearing clearance that were used in Case 2, but with a significant static unbalance added to the rotor. With the higher rotor load caused by the unbalance, the overall dynamic strain amplitudes increased even more significantly than the amplitude change measured in Case 2.

A higher overall dynamic strain amplitude change was expected when compared to Case 2 as the increase in load with the low oil viscosity would lead to a further decrease in the ZN/P ratio causing a further decrease in the film thickness within the thin film lubrication regime.

Again, in this test case, a much lower change in the overall shaft displacement amplitudes were measured by the proximity probes. Even with the decrease in oil viscosity and increase in load, the overall displacement amplitudes measured by the X-direction proximity probe continued to show a decrease after the oil supply to the bearing was lost.

Case 4

Figure 8: Case 4 - % overall amplitude change

As shown in Figure 8, the overall amplitude increase in dynamic strain was higher at nearly 200% in Case 4 compared to the 100% increase measured in Case 1. Due to the lower eccentricity of the shaft in the bearing with a tighter clearance, the initial overall strain amplitude was lower in Case 4 compared to Case 1. Once the oil supply to the bearing was removed, the overall dynamic strain measured with the high clearance and low clearance bearings were similar. This similar change in overall dynamic strain in both test cases, as well as the lower initial strain amplitudes measured in Case 4, caused the high percent change in the Case 4 overall amplitudes.

Case 4 was the first test case in which both proximity probes measured a positive increase in overall shaft displacement amplitudes. However, the overall shaft displacement amplitudes did not reach even twice the initial amplitude even after the oil supply was lost.

Case 5

Figure 9: Case 5 - % overall amplitude change

Figure 9 shows the overall amplitude change in dynamic strain and shaft displacement with hot oil and the same balance quality and bearing clearance as Case 4. Just as the lower oil viscosity caused a larger increase in overall dynamic strain amplitudes in Case 2 compared to Case 1, the same influence of viscosity was seen in Case 5 compared to Case 4. However, the lower oil viscosity of the hot oil led to a smaller change in the overall shaft displacement amplitudes measured by the proximity probes in Case 5 than were seen in Case 4.

Case 6

Figure 10: Case 6 - % Overall Amplitude Change

Figure 10 shows the change in dynamic strain and shaft displacement amplitudes with increased load caused by the added static unbalance to the rotor while maintaining the hot oil temperature and low clearance used in Case 5. This case was the first test case in which the increase in overall shaft displacement amplitudes measured by the proximity probes was similar to the increase in overall dynamic strain. Significantly, the increase in dynamic strain in Case 6 was consistent with the increase measured in Case 5. This shows that with a tighter bearing clearance the change in load on the bearing has less of an effect on the change in overall strain amplitudes when the oil supply was lost.

Case 7

Figure 11: Case 7 - % overall amplitude change

Figure 11 shows the overall amplitude change in dynamic strain and shaft displacement with cold oil and the same balance quality and bearing clearance as Case 6. Just as in Case 6, the higher bearing load due to the static unbalance and lower clearances caused a more significant increase in the overall shaft displacement amplitudes measured by the proximity probes compared to the increases seen in Cases 1 through 5. While a significant increase in overall shaft displacement was seen in this case, a more significant increase in overall dynamic strain amplitudes was measured than the increase in overall shaft displacement.

Conclusion

In all tested cases, the overall dynamic strain measured on the bearing increased significantly following the loss of lubricating oil to the bearing. A consistent increase in overall shaft displacement amplitudes measured by the proximity probes mounted to the bearing was not seen in all test cases. Typical vibration shutdown limits for fluid film bearings protected by proximity probes are set as a high alarm value based on bearing clearance. This can protect the equipment from failure related to many different malfunctions. However, based on the results of the testing on the laboratory rotor, in some test cases overall shaft displacement amplitudes decreased following the loss of oil supply. The data collected in all test cases support that a single dynamic strain sensor such as the Voyager FDS can offer better protection to a bearing that has lost oil supply.