Test Report: Jones Island IPS Vibration Survey. United Water

Transcription

1 Test Report: Jones Island IPS Vibration Survey Prepared for United Water by Roseville, CA Report R October 12, 25

2 Table of Contents 1. INTRODUCTION DESCRIPTION OF EQUIPMENT TESTED TEST INSTRUMENTATION INSTRUMENTATION DIAGRAM DESCRIPTION OF DATA ACQUISITION AND STORAGE OPERATING CONDITIONS PUMP 1 VIBRATION PUMP 2 VIBRATION PUMP 3 VIBRATION VIBRATION SEVERITY MOTOR VIBRATION CONCLUSIONS AND RECOMMENDATIONS...21 DynaTech Report R , October 12, 25 Page 2

3 1. Introduction This report presents the results of a vibration survey conducted on the three Ebara pump assemblies in the Inline Pump Station (IPS) at the Jones Island Wastewater Treatment Plant in Milwaukee, WI. The testing was conducted on April 26 and 27, 25 to determine why two of the three pumps discharging to the South Shore Head Tank have excessive vibration after being returned to normal service following replacement of the tank. A series of measurements were performed to map the vibration state on the three pumps, and in particular, the vibration produced at various speeds. DynaTech Engineering, Inc., an independent mechanical engineering consulting firm, was contracted by United Water to conduct the testing. DynaTech was responsible for determining the type and location of instrumentation used on the pump assembly, installing the transducers, and taking the vibration measurements. A complete description of the required vibration testing is contained in DynaTech test plan R prepared for United Water and dated April 18, 25. Acceptance criteria were taken from Hydraulic Institute standards, which specify general guidelines for evaluation of mechanical vibration. 2. Description of Equipment Tested The IPS consists of three large pumps drawing from the water collection tunnel approximately 3 underground and discharging to a pair of head tanks. A schematic of the pumping system is shown in Figure 1. The pumps are large Ebara 42X36 horizontal split case pumps, designated pumps 1, 2, and 3 in the plant. Each 42 pump inlet is connected to a 1 diameter suction tunnel, and discharges through a 36 pipe to the storage tanks. Note that Pump 1 can only discharge to the South Shore Head Tank (SSHT), Pump 2 can discharge to the SSHT and Jones Island Head Tank (JIHT), and Pump 3 can only discharge to the JIHT. Rating for each pump is 5 mgd, the pumps are currently operating at 35-4 mgd at 385 feet head. While the flow is not significant for a pump of this size, the required head is very substantial, roughly 16 psig, necessary to raise the water from the subterrain tunnel to the top of the head tanks. The pumps are driven by a 4 HP motor, controlled by a variable speed drive, with a nominal speed of 575 rpm. Each pump has a single stage impeller with 5 vanes. The motor and pump shafts are supported by fluid film bearings in a straddle mount configuration. DynaTech Report R , October 12, 25 Page 3

4 Figure 1. Layout drawing of the pumped storage configuration at Jones Island 3. Test Instrumentation The transducers used to acquire vibration signals during the survey consisted entirely of accelerometers. The mounting locations were as close as possible to the fixed instrumentation and specified placements defined in the test plan, generally on the bearing housings. Measurement axes were vertical, horizontal, and axial (longitudinal) directions. Vibration signals from the accelerometers were displayed and recorded using a Hewlett-Packard 3567A dynamic signal analyzer (DSA). This is a portable, 4 channel, laboratory quality analyzer with frequency, time, and amplitude domain analysis capabilities. All accelerometers were powered by a current source obtained directly from the HP 3567A analyzer. Based on the operating frequency range of the pumps, which is a maximum of 9.6 Hz, Vibra Metrics model 512 low-frequency industrial accelerometers were used, attached using magnetic holders. These transducers have frequency response characteristics of.1 25 Hz 5 mv/g, and have been calibrated within one year of the test date. Since a maximum frequency of roughly 1 times rotating speed is usually sufficient to measure significant resonant and forced response, these sensors had more than adequate dynamic range. DynaTech Report R , October 12, 25 Page 4

5 3.1 Instrumentation Diagram Accelerometer locations were also consistent with Hydraulic Institute recommendations, as shown in Figure 2, which also specify measurements in 3 directions at the inboard side of the pump. The vibration limit for these pumps, taken from HI Figure (page 17), is.22 inches/sec. Note that the vibration limit is for overall amplitudes, in RMS units. Figure 2. Hydraulic Institute vibration measurement locations and amplitude limits A total of 12 accelerometer positions were used to acquire vibration signals in the survey. These locations are illustrated in Figure 3 through Figure 6, which shows the position and orientation of the measurement sensors, along with a 3-letter identifier. In some of the pictures, the fixed instrumentation can be seen adjacent to the test accelerometers. DynaTech Report R , October 12, 25 Page 5

6 Fixed Outboard Vertical Sensor MOV MOA MOH Figure 3. Motor outboard measurement locations Fixed Inboard Vertical Sensor MIV MIA MIH Figure 4. Motor inboard measurement locations DynaTech Report R , October 12, 25 Page 6

7 Fixed Inboard Vertical Sensor PIV PIH PIA Fixed Inboard Horizontal Sensor Figure 5. Pump inboard measurement locations Fixed Outboard Vertical Sensor POA POV POH Figure 6. Pump outboard measurement locations A description of each measurement location is provided in Table 1. Although axial measurements were initially made on the first pump tested, it was found very little additional information was obtained from this measurement, and as such, it was not used for data acquisition during the majority of the survey. Since the shafts are supported on fluid film bearings, relatively low axial vibration is typical. DynaTech Report R , October 12, 25 Page 7

8 Table 1. Measurement Point Identifier Reference Identifier Description Identifier Description POV Pump Outboard Vertical MIV Motor Inboard Vertical POH Pump Outboard Horizontal MIH Motor Inboard Horizontal POA Pump Outboard Axial MIA Motor Inboard Axial PIV Pump Inboard Vertical MOV Motor Outboard Vertical PIH Pump Inboard Horizontal MOH Motor Outboard Horizontal PIA Pump Inboard Axial MOA Motor Outboard Axial 3.2 Description of Data Acquisition and Storage Using the instrumentation system described in section 3.1, the vibration signatures of the pumps were acquired and stored in the following manner: Signals from the accelerometers were directly acquired using the HP 3567A. For the operating conditions defined in Section 3.3 of this plan, an overall spectrum for each measurement location was produced and stored. A frequency range was set based on the operating speed range, which for these pumps is a maximum of 9 Hz. Normally, data is taken out to a range of approximately ten times operating speed, or a frequency range of 1 Hz. Broader ranges were examined, although no significant data at was observed at higher frequencies. As described in Section 3.3, various operating conditions were evaluated for vibratory response. The conditions were set by plant operators. During acquisition, the pumps were allowed to stabilize before any data was taken. All data at these steadystate points were acquired using 8 RMS averages. Transient data in the form of starts and stops was acquired by using the time capture feature of the DSA. This data was intended to be used to evaluate any suspected resonances, although as described in Section 4., other information was obtained. Data sheets were used to record measurement parameters and results. In addition to the vibration data, discharge pressure read from an analog gauge and speed read from a control panel next to the pump were noted. All of the data obtained during testing was copied from the floppy disks to a notebook computer for backup purposes. DynaTech Report R , October 12, 25 Page 8

9 3.3 Operating Conditions As specified in the test plan, the three pumps were run individually over as complete a range of operating conditions as possible. The specific conditions evaluated are listed in Table 2. Table 2. Operating Conditions Evaluated Test Pump Speed, rpm Pumping Location South Shore Tank South Shore Tank South Shore Tank Jones Island Tank Jones Island Tank Only one pump was operated at a time, since the vibration problem was very evident after the first test. Discharge was directed to both the South Shore Heat Tank (SSHT) and the Jones Island Head Tank (JIHT) with Pump 2 only. Various running speeds were set during the testing, from a minimum of 54 rpm to a maximum of 57 rpm. Because of the narrow flow-head range the pumps must operate in, the speeds are largely dictated by required discharge head (corresponding to which tank is being filled) and the water level in the collection tunnel. It is possible to vary the speed somewhat, although it was found only to be practical with Pump 3, as speed changes with Pumps 1 and 2 discharging to the SSHT would result in a vibration trip. During the testing, the water level in the main tunnel was approximately 3 feet, and changed very little during the relatively short run times (on the order of 3 minutes for each test). Normal conditions for operating pump 1 or pump 2 to the SSHT is following a rain event with tunnel level at 5 feet or greater. Pump performance and vibration are stated to improved with higher tunnel levels. 4. Pump 1 Vibration Initial measurements were made on Pump 1, since it was the primary source of pre-test vibration problems. The fixed pump outboard horizontal vibration sensor was not attached to the housing in order to permit operation, as the vibration during operation would exceed the allowable value of 1 microns overall. The pump manufacturer published allowable vibration is 12 microns, the limit of the vibration monitoring and protection equipment is 1 microns. Consistent with the fixed sensors, the highest amount of vibration was measured in the horizontal direction on the pump. Overall vibration amplitudes taken from the two tests of Pump 1 are listed in Table 3 for the twelve measurement positions. As noted in Section 3.1, the axial measurements were discontinued after the initial set of tests because no new significant data was present. As the table shows, the outboard pump vibration is higher than the inboard, with the inboard motor vibration higher than the outboard. With the motor, the vibration is very low and attempting to improve on these amplitudes would be difficult. DynaTech Report R , October 12, 25 Page 9

10 Table 3. Pump 1 overall vibration (inches/sec RMS) Sensor Overall Vibration (at condition) 57 rpm 565 rpm POV POH POA.153 PIV PIH PIA.99 MIV.21 MIH.31 MIA.26 MOV.14 MOH.16 MOA.22 Figure 7 illustrates a set of representative spectra recorded during operation of Pump 1 discharging to the South Shore Head Tank (SSHT) for the pump outboard and inboard horizontal measurements (POH and PIH, respectively). As the spectra show, the vibration is almost entirely due to a single predominant peak, at 2,85 rpm, which is 5 times running speed, corresponding to the number of impeller vanes. The relatively high magnitude of the 5/rev vane pass vibration causes the overall vibration level, which is monitored by the fixed instrumentation, to exceed the limit of the vibration protection equipment (1 microns). Large vane pass vibration with a pump is indicative of high head-rise operation, which is certainly the case with this machine. Discharge pressure from the pump was approximately 18 psi. The horizontal vibration component on the pump is greater than any other direction because the volute cutwater is oriented in the horizontal plane. The magnitude of all other vibration on Pump 1 was less than.125, which is very acceptable. It was observed, however, that the 5/rev vibration was the dominant frequency component in all measurements, even at the outboard side of the motor. DynaTech Report R , October 12, 25 Page 1

11 MMSD Jones Island Inline Pump Station Pump 1, 57 rpm, SSHT discharge Date: Time: 12:37: PM Location: POH X: 2.85 kcpm Y: Overall:.392 Location: PIH.5 CPM AVG: 8 X: 2.85 kcpm Y:.321 Overall:.329 5kCPM CPM AVG: 8 5kCPM Figure 7. Pump 1 horizontal vibration spectra To put the magnitude of the 5/rev vibration component into perspective, the amplitude of the running speed vibration (at 57 CPM) can just be discerned in Figure 7, particularly in the PIH measurement. Normally, vibration at running speed is the largest component of vibration, associated with unbalance. As this data shows, the balance of the pump is very good. A large vane pass vibratory component also implies a significant amount of radial load is being generated by pressure differences. Since there are no measurements of rotor displacement, it is not known if the load is excessive. The planned replacement of the existing fixed instrumentation to a setup more suitable for fluid film bearings should allow direct measurement of rotor position, which can then be used to determine if enough load capacity is generated by the bearings. Insufficient load capacity can create accelerated wear and excessive heat generation. DynaTech Report R , October 12, 25 Page 11

12 5. Pump 2 Vibration Pump 2 was tested discharging to both the JIHT and SSHT. By doing so, the difference in vibration due to the higher discharge head required by the SSHT could be observed. As with Pump 1, the highest vibration on Pump 2 was in the horizontal direction on the pump housing. Overall measurements found at the two tests on Pump 2 are listed in Table 4. Table 4. Pump 2 overall vibration (inches/sec RMS) Overall Vibration (at condition) Sensor SSHT discharge JIHT discharge POV POH POA.69 PIV PIH PIA.51 MIV.16 MIH.34 MIA.18 MOV.15 MOH.11 MOA.15 Plots of the spectra data from the two horizontal pump measurements are illustrated in Figure 8 for the POH and PIH locations. When flowing to the JIHT, the vibration of Pump 2 was much lower than Pump 1. The operating speed of Pump 2 was also lower than Pump 1, 54 rpm versus 57 rpm. These differences were due to the lower required discharge head flowing to the JIHT, as the discharge pressure was 16 psi. During the initial running of Pump 2, the cone valve controlling the discharge flow was closed, diverting flow to the SSHT, to obtain data with this condition. Unfortunately, during this operation the vibration increased and the pump tripped. DynaTech Report R , October 12, 25 Page 12

13 MMSD Jones Island Inline Pump Station Pump 2, 54 rpm, JIHT discharge Date: Time: 1:19: PM Location: POH X: 2.7 kcpm Y: Overall:.262 Location: PIH.25 CPM AVG: 8 X: 2.7 kcpm Y:.199 Overall:.212 5kCPM CPM AVG: 8 5kCPM Figure 8. Pump 2 horizontal vibration spectra discharging to JIHT A second attempt to obtain data from Pump 2 discharging to the SSHT was successful, obtained by closing the cone valve and starting the pump, rather than trying to switch it. Data from this condition is displayed in Figure 9, for the POH and PIH locations. Compared to Figure 8, the vibration is roughly 7% higher, even greater than Pump 1. It is not known why the fixed instrumentation system allowed the pump to continue to operate at such a high amplitude. DynaTech Report R , October 12, 25 Page 13

14 MMSD Jones Island Inline Pump Station Pump 2, 56 rpm, SSHT discharge Date: Time: 9:11: AM Location: POH X: 2.79 kcpm Y: Overall:.451 Location: PIH.5 CPM AVG: 8 X: 2.79 kcpm Y:.336 Overall:.346 5kCPM CPM AVG: 8 5kCPM Figure 9. Pump 2 horizontal vibration spectra discharging to SSHT To determine if the vane pass vibration was affected by speed change, Pump 2 was slowed to 55 rpm. The vibration increased and the pump tripped due to excessive amplitude. In general, vane pass vibration is proportional to the difference between the operating point and the pump best efficiency point (BEP), and reviewing the performance data, it appears that Pump 2 is running below BEP. By decreasing speed, this moves further away from BEP and creates correspondingly higher vane pass vibration. Additionally, a 3 minute time capture was taken from start to max speed (56 rpm) to check for resonances. Due to the sharpness of the vane pass peak, it could be theorized that the frequency is exciting a resonance. As can be seen in Figure 1, which is a waterfall plot, the 5/rev peak is quite prominent and the only significant vibratory component. The waterfall shows a succession of spectra captured while the pump is starting and reaching steady state. As the pump starts, the magnitude of the 5/rev gradually increases and shows no sign of resonant excitation. In addition, any resonances would also be seen as constant frequency lines in the waterfall, and as Figure 1 displays, there are no such traces. From the characteristics of the waterfall, it can be concluded that the large vane pass amplitude is entirely due to forced vibration. DynaTech Report R , October 12, 25 Page 14

15 MMSD Jones Island Inline Pump Station Pump 2, -56 rpm (start), SSHT discharge Date: Time: 7:8: AM A: Location: POH.5 45 count count 6CPM 6kCPM Figure 1. Waterfall plot of pump 2 startup discharging to SSHT 6. Pump 3 Vibration The vibration from Pump 3 was the lowest of all three pumps, even lower than Pump 2 discharging to the JIHT. Table 5 summarizes the overall vibration of Pump 3 at speeds from 541 to 555 rpm, with the highest vibration generally occurring at the lowest speed. Compared to Pump 2 discharging to the JIHT, the overall amplitudes on Pump 3 were approximately 5% lower. It should be noted that Pump 3 was just recently rebuilt in February 25, whereas Pump 2 was overhauled in February 24 and Pump 1 in February 23. As such, Pump 3 has the least amount of run time since overhaul, and would be expected to have lower vibration. However, the difference in overall vibration between Pumps 2 and 3 cannot totally be attributed to less operation, since the actual run time of Pump 2 is less than 1, hours. DynaTech Report R , October 12, 25 Page 15

16 Table 5. Pump 3 overall vibration (inches/sec RMS) Sensor Overall Vibration (at condition) 541 rpm 545 rpm 555 rpm POV POH PIV PIH MIV.5 MIH.8 MOV.12 MOH.6 The vibratory response of Pump 3 is nearly identical to that of the other two pumps. As illustrated in Figure 11, the predominant 5/rev component is the only significant peak, although the amplitude with Pump 3 is very good. Compared to Pump 2, the discrete peak magnitude at 5/rev is 5% less with Pump 3. The difference in magnitude is surprising, since the required discharge head is equal. A review of information concerning the three pumps disclosed that the orifice plates in the discharge lines are not the same size. On Pumps 1 and 2, the orifice hole diameter is 19.5 inches, and on Pump 3, the orifice is 23.5 inches. As such, there is significantly more flow restriction on Pumps 1 and 2, and this partially explains the difference in vibration with respect to Pump 3 and Pump 2 discharging to the JIHT. However, the vibration of Pump 2 is higher than Pump 1 under the same conditions (discharging to SSHT), so it can be concluded that the vibration of Pump 2 is the highest of all three pumps. The orifice plates were installed in the discharge lines to limit the backflow speed during shutdown. Since there is a 3+ foot column of water in the discharge pipes, backflow is a serious concern. The original check valves created excessive water hammer, and the orifice plates were installed to present a flow restriction for backflow. Obviously, this flow restriction is present during normal operation, and creates additional vane pass vibration since the pumps must deliver higher pressure to compensate for the loss. DynaTech Report R , October 12, 25 Page 16

17 MMSD Jones Island Inline Pump Station Pump 3, 54 rpm, JIHT discharge Date: Time: 9:58: AM Location: POH X: kcpm Y:.68.1 Overall:.89 Location: PIH.1 CPM AVG: 8 X: kcpm Y:.67 Overall:.89 5kCPM CPM AVG: 8 5kCPM Figure 11. Pump 3 horizontal vibration spectra In an effort to determine if the vane pass vibration could be affected by running speed, Pump 3 was manually commanded to run at various speed points. The results of this test are illustrated in Figure 12, which plots the discrete amplitude of the vane pass frequency versus speed. As the plot shows, the normal operating speed of 54 rpm is probably not the best condition for minimum vibration. A running speed of between 545 and 55 rpm is better from a vibration standpoint, probably because the pump is running near the BEP point. As discussed with Pump 1 and Pump 2 results, changing speed when discharging to the SSHT may improve the vane pass vibration. Unfortunately, the extreme sensitivity of the machines to speed changes off nominal prevented a complete mapping of this response. From the results with Pump 3, it can be concluded that these pumps probably obey the relationship with lowest vane pass vibration at best efficiency. The maximum efficiency point depends on tunnel level, so will be at different speeds depending on water elevation. DynaTech Report R , October 12, 25 Page 17

18 .1 Vane Pass Vibration, PIH POH Pump Speed, rpm Figure 12. Pump 3 horizontal vibration vane pass amplitudes versus speed 7. Vibration Severity Vibration severity, or a quantitative assessment of vibratory amplitudes, can be obtained from ISO , which establishes general guidelines for evaluation of mechanical vibration. A summary of the severity grades, by machine class, is reproduced in Table 6 for reference. These pumps would be considered a Class III machine, which is intended for large equipment on rigid foundations. The threshold of unsatisfactory operation is an overall amplitude of.441, which Pump 1 meets when discharging to the SSHT, although Pump 2 does not. The unsatisfactory rating is defined as a condition in which the machine should not be operated for continuous long-term periods, but is acceptable for limited running until repairs can be made. To be considered satisfactory, in which the machine is considered acceptable for unrestricted long-term operation, the overall vibration needs to be less than.177, which is only achieved by Pump 3 during this survey. As the repair history on these pumps corroborates, the high levels of vibration experienced by these machines results in relatively short overall cycles, certainly less than would be expected for industrial equipment of this size. DynaTech Report R , October 12, 25 Page 18

19 Vibration Level (in/sec rms) Table 6. ISO Vibration Severity Ranges Vibration Severity (by machine class) Class I Class II Class III Class IV Good Satisfactory Unsatisfactory Unacceptable Good Satisfactory Unsatisfactory Unacceptable Good Satisfactory Unsatisfactory Unacceptable Good Satisfactory Unsatisfactory Unacceptable Class I Class II Class III Class IV Small machines, up to 15 kw Medium machines, 15 to75 kw Large machines, above 75 kw, on rigid or heavy foundations Large machines, above 75 kw, on flexible foundations It should also be noted that when the pumps are turned off, the reverse flow due to the water column in the discharge pipes creates a very violent event. In fact, this reverse flow is probably the most extreme vibration encountered by DynaTech, considering tests on several hundred machines. An attempt was made to capture the shutdown event, however, the accelerometers overranged even when set to a maximum input level of 5 g s. Such high energy, short term back spinning is considered to have a substantial impact on the service life of these pumps. If at all possible, this backflow should be prevented or at least controlled so as not to create such a violent vibratory event. 8. Motor Vibration It was requested by a motor service company to check the vibration on the motors, particularly at the brush end, since abnormal life and heating have been occurring. As illustrated in Figure 13, a set of vibration spectra from the Pump 1 motor displays the primary frequency components, and once again, the 5/rev from the pump is by far the highest discrete signal. Obviously, if this component could be reduced, the motor vibration would be extremely low, although it is at a relatively low overall amplitude already. DynaTech Report R , October 12, 25 Page 19

20 MMSD Jones Island Inline Pump Station Pump 1, 57 rpm, SSHT discharge Date: Time: 1:47: AM Location: MOV X: kcpm Y:.7.1 Overall:.14 Location: MOH.2 CPM AVG: 8 X: kcpm Y:.12 Overall:.16 5kCPM CPM AVG: 8 5kCPM Figure 13. Pump 1 motor vibration spectra In the horizontal measurement on the motor, a small peak at roughly 23 rpm is present. Looking over the historical data, it was noted that the pump has a calculated critical speed at 237 rpm, and this peak may be the signature of this mode. This resonance is not apparent in the data from the pump due to the excessive magnitude of the 5/rev peak. The amplitude of the resonance is very low, approximately.2, so it appears to be well-damped and not of concern. DynaTech Report R , October 12, 25 Page 2

21 9. Conclusions and Recommendations Based on the results of this testing, the following conclusions can be made regarding the vibration situation with the IPS pumps at the Jones Island Wastewater Treatment Plant: The high vibration being experienced by Pumps 1 and 2 when discharging to the South Shore Head Tank is entirely due to a strong 5 times running speed excitation generated from the 5- vaned impeller in each pump. Because the required discharge head is much greater when discharging to this tank, the vane pass vibration is correspondingly higher. Current overall vibration amplitudes on Pumps 1 and 2 are considered to be unsatisfactory according to ISO classifications. The vibration of Pump 2 is higher than Pump 1 even though Pump 2 was overhauled a year earlier. The vibration of Pump 3 is 5% lower compared to Pump 2 when both are discharging to the Jones Island Head Tank. Some of this reduction may be due to Pump 3 having been recently rebuilt, and a larger orifice in the Pump 3 discharge line. The overall vibration of Pump 3 is considered to be satisfactory by the ISO standards. The characteristic of the vibration on all three pumps is completely unrelated to any effect from replacing the head tanks during 24. The vane pass vibration is solely due to the head demand from the discharge piping, including any flow restrictions. Based on the overhaul history of Pumps 1 and 2, it can be concluded that the service life of these machines is far less than would be expected from equipment of this size, and that the higher vibration observed when returning the pumps back to service is due to operational degradation. Another significant vibration problem with all three pumps is associated with shutdown. Although the current orifice plate limits the backspin speed, it allows a substantial amount of vibration to be produced, along with adding a flow restriction on Pumps 1 and 2. The severe nature of the shutoff event is considered to be the primary reason why the service life of these machines is so limited. Other than the vane pass vibration, no other significant vibratory signatures were observed in any of the data. Typical components such as unbalance and misalignment were virtually non-existent, showing that the general mechanical health of the pumps is good. Startup transients were examined for resonant effects and no appreciable critical speeds were observed. Considering these findings regarding the vibration of the pumps, the following recommendations are offered to improve the vibratory behavior: A different way of limiting backspin is needed to reduce this very violent event when the pumps are turned off. Although it is acknowledged that devising an effective replacement is not simple, it must be emphasized that the severity of the shutdown is seriously impacting operating life. The most helpful approach with respect to shutdown vibration is to limit or divert the volume of water in the discharge pipes that reaches the pumps. Unlike the orifice plates, any alternate solution should not increase the head demand on Pumps 1 and 2. DynaTech Report R , October 12, 25 Page 21

22 Pump 1 be rebuilt as soon as possible, and until this can be done, it should not be run for extended periods. Lessons learned and procedures utilized to recently rebuild Pump 3 should be employed during the overhaul of Pump 1. Pump 2 should not be used to pump into the South Shore Head Tank until it can be rebuilt or tunnel levels are high enough to produce adequately reduced vibration levels. It may be possible to reduce the vibration on all the pumps by running as close as possible to best efficiency. This point will change depending on tunnel water level. It was not possible to investigate this behavior due to the inability to run at speeds above 57 rpm. Perhaps some form of control system can be implemented to set pump operating speed close to BEP, given tunnel water level. The vibration monitoring system currently used does not provide any trend information or break-outs to determine if the amplitudes are correct. Since the system is being planned for replacement, the new system should include these features as a minimum requirement. Furthermore, the new system should not use structural motion sensors, but instead use shaft displacement probes, due to the use of fluid film bearings. The pumps at the IPS are being used in very high discharge head application. Under this condition, the high vane pass vibration is expected. Removing the discharge line orifices and running closer to BEP should improve the vibration, but completely eliminating the vane pass component is not possible. DynaTech Report R , October 12, 25 Page 22