2009 Pump Symposium Case Study
Replacing 7 Vane Impellers with 8
Vane Impellers Leads to Electric Motor
Overload in 9 Stage Pipeline Pump
By Robert Perez and Dewane Fambrough
Enterprise Products Inc.
2009 Pump Symposium Case Study 1
Case Study Abstract
Replacing 7 Vane Impellers with 8 Vane Impellers Leads to Electric Motor
Overload in 9 Stage Pipeline Pump
This case study deals with a west Texas pipeline transfer pump with the following design
specifications:
•Byron Jackson 3x6x9E DVMX
•BEP at 650 gpm, 3560 rpm, Ns=1220
•Originally 9 stages, 2 impellers were removed, utilizing remaining 7 stages
•Directly coupled to a 450 HP electric motor driver
After de-bottlenecking the pipeline to obtain more flow, the motor driver for this pump began
overloading at the new, higher flowrates. Horsepower calculations suggested that the electric
motor driver should have been adequate for the new hydraulic conditions. However, we began to
suspect there was an internal pump design problem once it was found that the field performance
data did not match the original test data. We discovered that the pump was producing 20% more
head than expected at end-of-curve conditions (900 gpm). Upon disassembly, we found that 8 of
the 9 impellers had 8 vanes instead of the expected 7 vanes. (The pump OEM provides both 7 and
8 vane impeller designs.) The OEM’s 4x6x9D 8 vane design produces more head past the best
efficiency point than the corresponding 7 vane design and does not continuously rise to shutoff.
To counteract the additional head producing capability of the 8 vane impellers, we decided to
remove 2 stages. After de-staging, the motor overload condition was resolved.
It is assumed that someone replaced the original 7 vane impellers with 8 vane impellers to obtain
more head and hence more flow with the same pump. This case study clearly illustrates the
danger of changing impeller designs without a thorough hydraulic analysis.
2009 Pump Symposium Case Study 2
Background
• The west Texas pipeline pump is a Byron
Jackson 3x6x9E DVMX driven by a 450 electric
motor operating a 3560 rpm. It has a BEP of
650 gpm and a specific speed of (Ns) 1220
• The pump, which was installed in 1995, transfers
liquid hydrocarbons approximately 14 miles
down a 6” pipeline.
• In 2008, to increase system flow capability,
approximately 9 miles of original 4” pipe were
upgraded to match the remainder of the 6”
pipeline.
2009 Pump Symposium Case Study 3
Photos of the west
Texas pipeline
pump installation
2009 Pump Symposium Case Study 4
Problem
• West Texas pipeline pump driver is rated for 450
hp. Calculations show there should be enough
horsepower for all flows on pump curve.
• However, the motor began tripping on high
amps after the pipeline size was increased.
• A field performance test 2/11/08 showed that the
pump was generating significantly more
pressure head and drawing more horsepower
than expected.
2009 Pump Symposium Case Study 5
2009 Pump Symposium Case Study 6
Comparison with OEM Pump Curve with Field Data Point
2009 Pump Symposium Case Study 7
Horsepower Analysis
Diff.due to
Head
Flow (gpm) eff., pf,
ratio
HPcurve HPelec HP ratio etc.
910 1.35 1.2 15%
464.9 627
910 1.31 1.2 11%
464.9 610
941.5 1.33 1.2 13%
471.4 625
889 465.0 561 1.20 1.2 0
873.6 454.8 555 1.22 1.2 2%
2009 Pump Symposium Case Study 8
Our Thinking After Field Testing
• The excessive power draw observed is
due to either:
– 20% higher than expected head. This may be
due to different impeller geometry (rare 8-
vane) than originally purchased
– Internal pump wear. More than likely, wear at
the impeller wear rings.
– Unknown motor factors such as the actual
power factor and efficiency.
2009 Pump Symposium Case Study 9
Findings After Field Testing
• Upon disassembly of the pump, we
discovered that 8 of 9 impellers had 8
vanes instead of the expected 7 vanes.
(The pump OEM provides both 7 and 8
vane impeller designs.)
• The existing 2” flow control valve and 2”
and 3” piping on the pump discharge
represents a large percentage (55%) of
the total head losses in the yard.
2009 Pump Symposium Case Study 10
How the number of vanes affects pump performance
(Source: Centrifugal Pumps, Design and Application by Lobanoff and Ross,
page 30)
2009 Pump Symposium Case Study 11
How the number of vanes affects pump
performance
• The fewer the vane number the greater
the difference between the BEP head and
the shutoff head
• The greater the number of vanes the
smaller the difference between the BEP
head and the shutoff head.
• As the number of vanes increase beyond
6, curve droop becomes more
pronounced.
2009 Pump Symposium Case Study 12
Example of drooping 8 vane impeller
Drooping portion of curve
2009 Pump Symposium Case Study 13
Path Forward
• Debottleneck existing 2” control valve and
2” and 3” piping on the pump’s discharge.
• Remove two pump stages to reduce
horsepower requirements at required
flows.
2009 Pump Symposium Case Study 14
Navajo Pump Derate Analysis
Expected system effects from removing 2 stages and debottlenecking yard piping
4500
Conclusion: The combination of
debottlenecking and destaging
should not affect delivered flow.
3500
Pump Curves
Existing
Head (feet)
2500
New
1500
System Curves
500
400 500 600 700 800 900 1000
-500
Flow (gpm )
9 stgs 7 stgs Existing piping New piping
2009 Pump Symposium Case Study 15
Results
2009 Pump Symposium Case Study 16
2009 Pump Symposium Case Study 17
Lessons Learned
• It is assumed that someone replaced the
original 7 vane impellers with 8 vane
impellers to obtain more head and hence
more flow with the same pump.
• This case study clearly illustrates the
danger of changing impeller designs
without a thorough hydraulic analysis.
• These effects are more pronounced
beyond BEP.
2009 Pump Symposium Case Study 18
