Piping Load Effect on Shaft Vibration in a Multi-Stage Barrel Type Boiler Feed Pump

Transcription

1 Piping Load Effect on Shaft Vibration in a Multi-Stage Barrel Type Boiler Feed Pump Maki M. Onari - MSI Manager of Turbomachinery Teing Principal Engineer Gary Krafft - HydroTex Dynamics, Inc. Pump Maintenance Speciali Kei Munn - AEP Welsh Plant Maintenance Supervisor

2 Piping Load Effect on Shaft Vibration in a Multi- Stage Barrel Type Boiler Feed Pump Maki M. Onari: Mechanical Solutions, Inc. Manager of Turbomachinery Teing - Principal Engineer: responsible for all MSI Turbomachinery Teing B.S.M.E., Zulia University, Venezuela Rotating Equipment Engineer, PDVSA: responsible for e predictive maintenance of one of e large petrochemical complex in Latin America Co-Auor Pump Vibration Chapter, McGraw-Hill Pump Handbook Member of ASME and e ISO TC108/S2 Standards Committee for Machinery Vibration 2 Kei Munn: AEP Welsh Power Plant Maintenance Supervisor 5 years: responsible for e mechanical maintenance of e HP & LP turbines, boiler feed pump turbines, boiler feed pumps, generators, valves, condensers, HP and LP heaters, water treatment facility, cooling towers and all oer pumps in e electrical process. Welsh Maintenance Planner- 3 years: performed planning work for all mechanical maintenance. Welsh Station Machini- 25 years: performed mechanical maintenance on all plant equipment. Gary Krafft : HydroTex Dynamics, Inc. for 17 years as a technical and sales representative BSME Texas A&M University Regiered Professional Engineer in e State of TX 22 years in e electric power generation indury. He originated TXU's Equipment Repair Group formed in 1982 to improve reliability wi rotating machinery. He has worked in fossil and nuclear plants as well as handled projects for gas pipeline and mining. He was a member of e Pump Symposium Advisory Committee for 5 years co-auoring one tutorial and leading discussion groups.

3 General Information American Electric Power (AEP) Welsh 1 Power Plant in Pittsburg, TX Base load: 1500 MW Coal Fired Power Plant (Welsh 1, 2, and 3 combined) The Turbine Driven Boiler Feed Pump (TDBFP) was inalled in Continuous operation wi one 100% BFP per Unit Type: 4 age Double Case wi Twin Suction and 2 nd age bleed-off Rotation: CCW Viewed from e Driver Speed: 4,000 rpm to 4,860 rpm (66.7 Hz to 81 Hz) Capacity: 9126 GPM Suction Pressure: 216 psig Discharge P.: 3016 psig Total Dynamic H: 7,370 ft Temperature: 375 ºF Specific Gravity: BHP: 17,530 HP (eam turbine) 5-pad tilting pad bearings wi forced lubrication syem Bearing nominal clearance: 7 to 9 mils diametral 3

4 4 age Boiler Feed Pump Welsh 1 - TDBFP Viewed from nor 4

5 Welsh 1 TDBFP Cross-Sectional Drawing 1B Ring 1A Ring Viewed from sou 5

6 Background Welsh 1 plant was built using a De Laval BFP (1974), while Welsh 2 and 3 were built wi Pacific Pumps. All ree plants have e same Weinghouse eam turbine. In May 2009, e pump case was weld repaired and machined on-site to correct potential concentricity and parallelism issues. The train ran wiout vibration concerns for a few mons. In September 2009, elevated shaft vibration levels on e entire train were detected. It was concluded at e vibration was mo likely due to internal rub and coupling alignment. In May 2011, HydroTex (HT) supervised a head in place balance assembly change out. In December 2011, a vibration issue in e coupling area was detected. Vibration apparently had arted rising in late November HT inspected e pump and turbine bearings. The pump inboard X & Y amplitude difference increased after is inspection. 6

7 Background In March 2012, HT took cold and hot alignment readings to determine a new coupling offset target. The new target was implemented in e vertical direction and a little off target in e horizontal direction. The entire machine train came up wi low vibration levels and reduced sound from e turbine exhau hood. After a few days and low load operation, e vibration gradually ep increased, resulting in an X reading of above 3.0 mils pk-pk (pump inboard bearing or IBB), while e Y probe indicated 1.5 mils pk-pk. Vibration increased about 0.5 mils pk-pk every night until 3.0 mils pk-pk was reached at e IBB-X probe. In September 2012, e pump inboard bearing housing was machined by HydroTex Houon shop, wiout improvement on e IBB-X probe magnitude. From December 15, 2012 rough January 5, 2013 e IBB-X probe was indicating 2.5 mils pk-pk (below alarm level of 3.5 mils pk-pk). It appeared at e alignment has been playing a significant role affecting e vibration readings (pump sensitive to e alignment). There was a time when slight adjument of e seal water drain temperature would inantly change e IBB-X from 3.0 to 1.5 mils pk-pk. 7

8 Background In March 2013, HT repositioned e pump IBB housing wiout improvement on IBB-X probe. In September 2013, HT swapped elements and adjued e horizontal rim closer to target value. After art up, IBB-X & Y values were close in magnitude and low. Shaft vibration amplitude remained low and acceptable rough December 19, On December 29, 2013 vibration levels ramped up. In late Dec '13 and early Jan '14 e IBB-X probe indicated over 4.0 mils pk-pk and e Y probe about 2.0 mils pk-pk. The levels were e highe at low load initially, en after several days, switched to being highe at e higher loads. The element at was removed in September 2013 showed egged-shape wear on top of e 1A and 1B wear rings, wi about 2.0 to 3.0 mils of wear. 8

9 TDBFP Hydro Tex Inspection of Element Removed in September 2013 Welsh 1A Ring Suction Diaphragm TIR 3.0 mils wear at e top of e wear ring Welsh 1B Ring Inlet Guide TIR 3.5 mils wear at e top of e wear ring 9

10 Background In January 2014, AEP raised e alarm level to 6.0 mills. Then, e vibration amplitude reached 8.0 mils pk-pk. On January 18, 2014, e plant reported 6.0 mils pk-pk at e IBB-X probe at low speed operation. Then is behavior changed and e high vibration occurred at high load. On February 1, 2014, e plant was shut-down for a short outage and AEP decided to inspect and replace e IBB-X probe. Unfortunately, during e artup, e main transformer from e main turbine-generator failed and e outage was extended for several weeks until March 6, During is extended forced outage MSI was engaged to set-up and perform orough vibration and nozzle load teing. 10

11 Vibration Teing Approach Experimental Modal Analysis (EMA) te to determine e natural frequencies of e pump ructure and e rotor syem. Continuous Monitoring (CM) teing during transient and eady operation to monitor e shaft and bearing vibration amplitude, ructural natural frequencies, and pressure pulsations. In addition, rain gauges were used to monitor forces and moments in ree orogonal directions for e suction and discharge nozzles of e TDBFP. Operating Deflection Shape (ODS) teing during eady operation at full load conditions. 11

12 Experimental Modal Analysis (EMA) Te Results While TDBFP Was Not Operating 12

13 TDBFP OBB Frequency Response Function (FRF) Plot Horizontal Direction FRF Pump Not Running 13

14 TDBFP IBB Frequency Response Function (FRF) Plot Horizontal Direction FRF Pump Not Running 14

15 Experimental Modal Analysis (EMA) Te Results While TDBFP Was 4100 rpm or 200 MW (Accumulative Time-Average Meod) 15

16 TDBFP Structural Natural Frequency at 48 Hz 16

17 TDBFP Structural Natural Frequency at 84.8 Hz 17

18 TDBFP OBB Frequency Response Function (FRF) Plot Horizontal Direction (Pump 4100 rpm) 18

19 TDBFP OBB Frequency Response Function (FRF) Plot IBB-Y Direction (Pump 4100 rpm) 19

20 Continuous Monitoring Te Set Up and Results from March 5 rough 10,

21 TDBFP Pump MSI s Inrumentation Diribution 45 Channels of Data Acquisition Axial T-1 OBB-X OBB-Y IBB-Y T-2 IBB-X E N S W ST-#: Short Travel Proximity Probe T#: Tri-Axial Accelerometer 21

22 TDBFP Pump MSI s Inrumentation Diribution 45 Channels of Data Acquisition OB Sou A PT-discharge PT-suction IB Sou A OB Top A Axial OBB-Y OBB-X T-4 IBB-Y IB Top A OB Nor A T-3 IBB-X IB Nor A IB Nor H LT-#: Long Travel Proximity Probe ST-#: Short Travel Proximity Probe T#: Tri-Axial Accelerometer PT: Pressure Transducer E 22 S N W

23 TDBFP Pipe Strain Gage Diagram B Torque Axial Shear S-N Axial Torque Shear W-E A Bending S-N Bending W-E Orientation of Axis Positive (+) Directions Shown C Suction Pipe OD: 18.0 Suction Pipe Thick: 3/8 Discharge Pipe OD: Discharge Pipe Thick: 1.7 W N S E 23

24 Strain Gage Rosettes Wiring Diagrams Bending Torque Shear Axial 24

25 Strain Gages Tack-Welded on e Suction and Discharge Piping 25

26 Speed Trend Plot From 3/5/14 to 3/10/ rpm 4030 rpm TDBFP Start-up Plant Full Load (525 MW) Plant Low Load (200 MW) (355 MW) Lube Oil On Casing Warm-Up Cold Conditions Turning Gear On 26

27 Suction Pipe Bending & Torque Trend Lube Oil On 20.6k 12.8k 2.3k -6.8k TDBFP Start-up Effects of plant s load changes Cold Conditions Turning Gear On Casing Warm-Up -.5k 27 During warm-up, suction pipe is twiing from sou to ea and bending towards sou-ea. At full load (525 MW) maximum piping load was reached in e same direction. Plant load plays a role in changing e suction bending and especially e torque.

28 Suction Pipe Shear & Axial Trend Plot Lube Oil On 3.0k Effects of plant s load changes 3.2k 0.25k 7.75k TDBFP Start-up 1.4k 18.3k Cold Conditions Turning Gear On Casing Warm-Up 37.4k 19.1k During warm-up, suction pipe is in compression and lightly shearing towards nor-ea. At full load (525 MW) maximum piping load was reached, especially in e axial direction. The MW load of e plant applies significant downwards axial load to e pump casing. 28

29 Discharge Pipe Bending & Torque Trend Lube Oil On 57.5k 12.0k 25.0k 1.25k Effects of plant s load changes Cold Conditions Casing Warm-Up Discharge Bending Gage Damaged during Inallation of Insulation 29 During warm-up, discharge pipe is twiing from sou to ea and bending towards ea (nor-sou bending gage lo prior warm-up). At full load (525 MW) maximum piping load was reached in e same direction. Plant load plays a role in changing e discharge torque and especially e bending load.

30 Discharge Pipe Shear & Axial Trend Plot Lube Oil On Cold Conditions Turning Gear On 1.0k -2.5k -5.0k Casing Warm-Up TDBFP Start-up 17.7k -11.6k -138k Effects of plant s load changes -108k 30.0k 30 During warm-up, discharge pipe is lightly loaded (compared wi full load operation). At full load (525 MW) maximum piping load was reached, especially in e axial direction (compression). The MW load of e plant applies significant downwards axial load to e pump casing. The discharge pipe is under shear in e nor-we direction.

31 Welsh 1 TDBFP Pipe Strain De Laval Maximum Allowable Force & Moments

32 Welsh 1 TDBFP Pipe Strain De Laval Maximum Allowable Force & Moments Suction Discharge UnitsSuction Discharge Units Shear 10 E-W Fx (+E) 10A ft A 2.70 ft Shear 60 N-S Fz (+S) lbf 60B ft lbf B 3.75 ft Axial Fy (+Up) C ft C 3.75 ft to S) Bending -500 N-S Mx (+N 0 to S) o W) Bending E-W Mz (+E to W) ft-lb ft-lb ) Torque My (+Up) Fxt Fxt Resultant force exceeds Fyt Fyt Fr x Fr times De 175 Laval x1000 Calculations Fzt Fzt Mxt -500 Mxt -500 Resultant moment exceeds Myt Myt Mr x1000 Mr 3.2 times De 181 Laval x1000 Calculations Mzt Mzt

33 Welsh 1 TDBFP Pipe Diagram 33

34 Photos of Hangers for e Suction & Discharge Piping 25% loaded 0% loaded Suction Pipe Hanger top leaning we. 25% loaded Discharge Pipe Discharge Pipe 34

35 Turbine Shaft Overall Vibration Trend Plot Lube Oil On TDBFP Start-up Cold Conditions Casing Warm-Up 35 Turning Gear On Turbine rotor overall vibration was measured to be low and acceptable (~1.0 mil pk-pk) for e entire art-up evolution of e te.

36 Pump Shaft Overall Vibration Trend Plot 4860 rpm ~830 rpm 4030 rpm Lube Oil On TDBFP Start-up 4.6 to 1.6 mils pk-pk Cold Conditions Casing Warm-Up 36 Pump rotor overall vibration at e IBB was measured to be elevated (~5.0 mil pk-pk) at full load operation (525 MW). Strong vibration change versus speed effect is observed in is trend plot. ~830 rpm reduction, e vibration at IBB-X was reduced from 4.6 to 1.6 mils pk-pk.

37 Pump Inboard Bearing Orbits Sequence of e IBB orbit plots at different load conditions. 37

38 Pump Inboard Bearing Vibration Lube Oil On TDBFP Start-up Cold Conditions Casing Warm-Up Pump IBB housing overall vibration was measured to be elevated mild (~0.35 in/s RMS in e horizontal direction) at full load operation. 38

39 Pipe Axial Load and Power Lube Oil On TDBFP Start-up Cold Conditions Casing Warm-Up 39 As e plant load was increasing e axial load of e piping again e casing was also increasing (compression load). The change in temperature of e condensate water was playing a significant change in load on e pump casing due to ermal expansion effect.

40 Temperature and Pipe Axial Load Lube Oil On TDBFP Start-up Cold Conditions Casing Warm-Up 40 As e plant load was increasing e axial load of e piping again e casing was also increasing (compression load). The change in temperature of e condensate water was playing a significant change in load on e pump casing due to ermal expansion effect.

41 Pump Shaft Vibration (IBB X) and Pipe Axial Load Lube Oil On TDBFP Start-up Cold Conditions Casing Warm-Up This plot shows e relationship between e IBB-X overall vibration and e casing temperature at e IB and OB ends. 41

42 Operating Deflection Shape (ODS) Te Results while TDBFP was Operating at a Steady High Load (~525 MW) 42

43 Operating Deflection Shape (ODS) Teing 778 Vibration Locations/ Directions 43

44 ODS Animation at 1x rpm (wiout Shaft)

45 ODS Animation at 1x rpm (wi Shaft) 45

46 ODS Animation at 1x rpm (wi Shaft) OBB IBB 46

47 Conclusions The root cause of e high vibration amplitude at e IBB of e pump is due to excessive preload acting on e bearing. Excessive piping rain was measured from e suction and discharge loading. The pump casing was acting as a pipe anchor or support. 2. Diortion of e pump casing was taking place. Egged shape of e wear ring from previous element confirmed at e casing was diorted at high load operation (i.e. high temperature). 3. The close hanger to e pump on e suction piping (double hanger) was lightly or barely loaded (0% to 25% loaded) wi e load indicators at e bottom end of e scale. 4. The close hanger to e pump on e discharge piping was detected to be lightly loaded (only 25% loaded) and e upper end leaned towards we direction. 5. The suction and discharge nozzles were under severe axial or vertical loading (downwards - in compression). Bo nozzles were also under torque in e CCW direction (view from e top). 6. The lateral rocking and e twi modes of e pump on its pedeal were not reacting wi e running speed frequency. The fir lateral mode of e pump rotor was detected above e 1x rpm (174 Hz).

48 Recommendations 1. The loads of e hangers for e suction and discharge piping mu be corrected. The hangers mu be set in eir raight vertical condition. The excessive vertical load was considered e main contributor for e casing deformation and erefore, pre-loading of e IB bearing of e pump. 2. The loads for each hanger should be corrected at cold conditions, taking into account e longitudinal ermal grow of e piping. Once ese issues are corrected, e new baseline alignment approach should be conducted. 3. Conduct a hot alignment for verification purposes. 48

49 Follow- Up Vibration Data After Correcting Pipe Hangers 49 Pump Speed: Load: Pump IBB-X: Pump IBB-Y: 4816 rpm 523 MW 2.2 mils pk-pk 0.9 mils pk-pk