DESIGN AND FABRICATION OF A VERTICAL PUMP MULTIPHASE FLOW

LOOP

A Thesis

by

KLAYTON EDWARD KIRKLAND

Submitted to the Office of Graduate Studies of

Texas A&M University
in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

Approved by:

Chair of Committee, Gerald Morrison

Committee Members, Yassin A. Hassan
Dan McAdams
Head of Department, Jerald A. Caton

December 2012

Major Subject: Mechanical Engineering

Copyright 2012 Klayton Edward Kirkland

 ABSTRACT

A new centrifugal pump has been devised to handle two-phase flow. However, it

requires full scale testing to allow further development. Testing is required to verify

performance and to gain information needed to apply this design in the field. Further,

testing will allow mathematical models to be validated which will allow increased

understanding of the pump’s behavior. To perform this testing, a new facility was

designed and constructed.

This facility consists of a closed flow loop. The pump is supplied by separate air

and water inlet flows that mix just before entering the pump. These flows can be

controlled to give a desired gas volume fraction and overall flow rate. The pump outlet

flows into a tank which separates the fluids allowing them to re-circulate. Operating inlet

pressures of up to three hundred PSIG will be used with a flow rate of twelve hundred

gallons per minute. A two-hundred fifty horsepower electric motor is used to power the

pump.

The loop is equipped with instrumentation to measure temperature, pressure,

flow rate, pump speed, pump shaft horsepower, shaft torque, and shaft axial load. The

pump itself has a clear inlet section and a clear section allowing visualization of the

second stage volute interior as well as numerous pressure taps along the second stage

volute. This instrumentation is sufficient to completely characterize the pump.

Design and construction details are provided as well as a history of the initial

operating experiences and data collected. A discussion of lessons learned is given in the

ii
conclusions. Future projects intended to use this facility are also given. Finally, detailed

design drawings are supplied as well as operating instructions and checklists.

iii
 DEDICATION

This thesis is dedicated to Katelyn Anne Kirkland and Connor James Kirkland.

iv
 ACKNOWLEDGEMENTS

I would like to thank my committee chair, Dr. Morrison for his unwavering

support. I also wish to thank Shell for sponsoring the project which made this thesis

possible.

For encouragement and advice, I wish to thank Glenn Wiltcher, and Todd Grage.

At the Turbomachinery Lab, there are many people to thank including Ray Mathews,

and my colleagues Sahand Pirouzpanah, Emanual Marsis, Ramy Moaness, Nicolas

Carvajal, and Joseph Marchetti. All of you have contributed enormously to this thesis.

Finally, thanks to Karen Vierow for her unwavering support and love.

v
 NOMENCLATURE

GVF Gas Volume Fraction

MVP Multi-Vane Pump

PH Pipe Hanger*

PHA Pipe Hanger Assembly*

PS Pump Stand*

PSA Pump Stand Assembly*

PSIG Gauge Pounds per Square Inch

PSID Differential Pound per Square Inch

SPOOL Pipe Spool*

TS Tank Stand*

TSA Tank Stand Assembly*

VFD Variable Frequency Drive

*These items appear as drawing number prefixes (see Appendix B)

vi
 TABLE OF CONTENTS

Page

ABSTRACT .............................................................................................................. ii

DEDICATION .......................................................................................................... iv

ACKNOWLEDGEMENTS ...................................................................................... v

NOMENCLATURE .................................................................................................. vi

TABLE OF CONTENTS .......................................................................................... vii

LIST OF FIGURES ................................................................................................... ix

LIST OF TABLES .................................................................................................... xi

1 INTRODUCTION .............................................................................................. 1

1.1 Background .......................................................................................... 1

1.2 Problem statement ................................................................................ 4
1.3 Test program ........................................................................................ 5

2 LITERATURE REVIEW AND REQUIREMENTS .......................................... 7

2.1 Previous work ....................................................................................... 7

2.2 Description of MVP pump ................................................................... 8
2.3 Test loop requirements ......................................................................... 13

3 DESCRIPTION OF FACILITY.......................................................................... 15

3.1 Overall configuration ........................................................................... 15

3.2 Piping ................................................................................................... 20
3.3 Pipe supports ........................................................................................ 21
3.4 Cooling loop ......................................................................................... 24
3.5 Separator tank and stand ...................................................................... 25
3.6 Pump stand ........................................................................................... 26
3.7 Pump inlet ............................................................................................ 31
3.8 Pump outlet .......................................................................................... 33
3.9 Seal assembly ....................................................................................... 34
3.10 Coupling ............................................................................................... 36
3.11 Operating instructions .......................................................................... 39

vii
4 CONSTRUCTION AND COMMISSIONING ................................................... 43

4.1 Separator tank and stand ...................................................................... 43

4.2 Piping and control valves ..................................................................... 44
4.3 Motor and stand .................................................................................... 51
4.4 Pump installation .................................................................................. 52
4.5 Initial operating experiences ................................................................ 53
4.6 Initial data collection ............................................................................ 57

5 SUMMARY AND CONCLUSIONS.................................................................. 63

5.1 Summary .............................................................................................. 63

5.2 Conclusions .......................................................................................... 63
5.3 Future work .......................................................................................... 64

REFERENCES .......................................................................................................... 66

APPENDIX A PROCEDURES ................................................................................ 67

APPENDIX B DRAWINGS ..................................................................................... 76

viii
 LIST OF FIGURES

FIGURE Page

1.1 MVP impeller ............................................................................................. 3

2.1 Basic pump assembly ................................................................................. 8

2.2 Pump with pressure taps ............................................................................. 9

2.3 Second stage volute with window slot ....................................................... 10

2.4 Window slot on assembled pump............................................................... 11

2.5 Complete pump assembly .......................................................................... 12

3.1 Two phase flow pump test loop piping and instrumentation diagram….. 16

3.2 Plan and elevation views ............................................................................ 17

3.3 Facility north view ..................................................................................... 18

3.4 Facility south view ..................................................................................... 19

3.5 Valve support stands .................................................................................. 22

3.6 Separator mounted supports ....................................................................... 24

3.7 Separator support stand .............................................................................. 26

3.8 Pump stand ................................................................................................. 27

3.9 Detail of cross brace fittings ....................................................................... 28

3.10 Pump mount ............................................................................................... 29

3.11 Pump head .................................................................................................. 30

3.12 Inlet assembly ............................................................................................. 31

3.13 Inlet window at 1000 PSIG and ambient temperature ............................... 32

3.14 Pump outlet, seal, and coupling cross section ............................................ 33

ix
3.15 Seal assembly ............................................................................................. 35

3.16 Coupling spacer .......................................................................................... 37

3.17 Coupling spacer strain under 3000lb axial load ......................................... 38

3.18 Coupling spacer strain under 4380 lb-ft torque .......................................... 39

4.1 Main support column being raised ............................................................. 46

4.2 Rear of tank showing the temporary crane arrangement ............................ 47

4.3 Temporary crane assembly mounted to top of separator tank.................... 48

4.4 Ready to lift spool seven ............................................................................ 49

4.5 Lifting frame .............................................................................................. 56

4.6 Lifting frame toggle ................................................................................... 57

4.7 Comparison of experimental data to pump supplier data ........................... 58

4.8 Data for 100 Psig inlet pressure at 3600 rpm ............................................. 59

4.9 Data for 100 psig inlet pressure at 3300 rpm ............................................. 60

4.10 Data for 100 psig inlet pressure at 3000 rpm ............................................. 61

x
 LIST OF TABLES

TABLE Page

2.1 Test matrix.................................................................................................. 13

xi
 1 INTRODUCTION

1.1 Background

Centrifugal pumps have been in use for well over a century and are some of the

most common pieces of industrial equipment in existence. It is easy to see why, they are

reliable and efficient. They possess relatively few moving parts and can handle high

flow rates. Finally, the vertical turbine layout lends itself to downhole applications.

They do possess one drawback in that they cannot easily tolerate two phase flow. A

new, proprietary pump design has been developed to address this issue. This has been

named the Multi-Vane Pump, (MVP).

The MVP is a development of the Electric Submersible Pump (ESP), which is a

vertical centrifugal pump with a self-contained, close coupled electric motor. The entire

unit can be immersed in the working fluid as the motor is hermetically sealed, with

power provided by electrical cabling. Immersing the inlet helps ensure that the pump has

sufficient Net Positive Suction Head Available (NPSHA) to prevent pump cavitation. In

centrifugal pumps, cavitation occurs when the pump fluid vaporizes due to the local

pressure falling below the vapor pressure for the fluid. This typically occurs in the eye

(center) of the impeller as the pressure is lowest at this point. When the pressure

increases above the vapor pressure, the vaporized fluid bubble collapses rapidly, often

with enough force to damage the pump. To prevent this, the inlet pressure must be

maintained safely above the vapor pressure. This margin is calculated by the pump

manufacturer and is called the Net Positive Suction Head Required (NPSHR). To

1
prevent cavitation, the NPSHA (determined by the inlet conditions and intake design)

must be greater than the NPSHR (determined by the pump design). [1]

This requirement is a limitation of all centrifugal pumps, but is neatly addressed

by the ESP configuration. The close coupling of the electric motor means that the shaft

very short, reducing the cost and the alignment and dynamic problems associated with a

long power shaft.

Applications for ESPs are numerous, including sewage treatment, fire water, and

general industrial purposes. They are very useful for wells, due to their vertical

configuration and immersive capabilities. Many impellers can be stacked within an ESP

to provide pressure generating capability while still fitting within a well bore.

Furthermore, immersing the inlet in the pumped fluid helps ensure that enough NPSHA

is available to prevent cavitation. For these reasons, ESPs are often found in oilfield

applications. Oil wells often do not possess enough pressure to produce oil without some

work input. In other words, oil must be pumped to the surface. In the past, surface

mounted pump jacks have been used. These have been replaced in many cases by ESP

units.

Oil wells, however, produce more than oil. Wells often will have a mixture of

fluids, including gases. The pump must in that case be capable of handling two-phase

flow. Note that this is a different phenomenon than the cavitation mentioned earlier.

Cavitation is the vaporizing and rapid condensing of a fluid within the pump. The two

phase flow mentioned here is of two or more different species, with gas already present

at the pump inlet. This is a challenge for centrifugal pumps as the gas tends to

accumulate on the low pressure side of the impeller vanes, impeding the flow. For fluids

2
with an appreciable Gas Volume Fraction (GVF), a gas separator can be fitted to the

pump ahead of the inlet. Gas is separated by inducing a vortex in the inlet fluid, which

forces the denser liquid to the perimeter of the separator. The gas is left near the center,

where it can be routed to bypass the pump. [2] This is an inefficient process, and the

MVP intends to improve on this by handling a large GVF within the pump itself due to

its unique impeller design as shown in figure 1.1. The vanes are split to prevent gas

accumulating at the low pressure side of the vanes, which is what normally occurs in

centrifugal pumps in two-phase service. Large balance holes can also be used to break

up the gas buildup. [3]

Figure 1.1 MVP impeller [3]

3
 To apply any pump effectively in the field, its operating characteristics must be

known. The pump must be appropriately sized for the expected conditions. In this

particular case, the MVP impellers are somewhat less efficient than standard impellers.

So, an MVP pump will use a number of MVP type impellers for the first few stages and

then progress to more conventional impellers for the remaining stages. As the pressure

rises from stage to stage, the GVF declines due to compression of the entrained gas,

eventually allowing normal impellers to be used. It is critical to know how the MVP

impellers behave so that the proper number can be determined for an application prior to

pump installation. This is a particularly important point as the cost of an installed ESP

can approach one million dollars.

1.2 Problem Statement

The performance of MVP pump stages needed to be evaluated under varying

operating conditions. These included controlled parameters such as Gas Volume

Fraction, inlet pressure, speed, and flow rate. Other parameters, such as power

consumption, differential pressure, internal pressure, and thrust load are also measured.

MVP performance had not been extensively characterized previous to this work.

Operating characteristics were thus not known to the level of detail desired.

4
1.3 Test Program

A test program using an actual MVP pump was conducted. The MVP impeller

configuration is sufficiently different from normal designs that it was not prudent to

extrapolate data from existing designs.

Concurrently with the construction of the test loop, a Computational Fluid

Dynamics (CFD) model was developed and validated. This CFD model can be used to

predict behavior under conditions beyond those tested as well as give insight for possible

design improvements. Note that development of this model, while part of the overall

MVP research project, is outside the scope of this thesis. It is, in fact, the subject of a

dissertation currently being developed.

This thesis describes the designing and constructing a test loop for a 250 hp,

three stage MVP that can record the inlet and outlet conditions under various flow rates

and pressures. Additionally, the pressure distribution within the pump can also be

determined. This was accomplished by adding a number of pressure taps (about 30) into

the casing on the second stage of the pump. These were placed to record the pressure of

the fluid as it progresses through the stage. A window installed on the side of the pump

opposite the pressure taps allows flow visualization. A large hole was milled into the

pump and a CNC milled clear acrylic plug inserted. An o-ring seals the plug and it is

held in place against the pump pressure by a clamp assembly. This window looks

directly into the second stage volute allowing a detailed view of the internal two-phase

flow characteristics.

The pump tested is a three stage MVP driven by a two hundred fifty horsepower

electric motor. It will be able to handle up to 350 PSIG of inlet pressure at 1000 GPM of

5
water flow and a GVF of up to 60%. The motor will use a variable frequency drive to

allow it to run from 1800 to 3600 rpm as desired.

6
 2 LITERATURE REVIEW AND REQUIREMENTS

2.1 Previous Work

Test facilities are not new for pumps, of course, even in two phase flow.

However, none have the unique capabilities of this one. A full scale pump will be

evaluated at high pressures and flow rates, with many parameters analyzed both external

to the pump and internal. A typical full scale pump test examines the inlet and outlet

conditions only. This facility will have the capability to measure and observe flows

within the pump.

Gruselle, Steimes, and Hendrick, for example, tested a two phase pump system,

but, only at a pressure of less than one quarter bar.[4] Gamboa, and Prado built an ESP

test loop, with an acrylic casing to allow flow visualization. Pump differential pressures

were no higher than 1.2 PSID, whereas the proposed loop is expected to reach 150 PSID

or more. [5]

Manufacturers often test their pumps by use of a deep well. Immersing the test

pump in the well allows for realistic operating conditions. Baker Hughes Centrilift, for

example can test an entire ESP system at elevated temperatures to 240º C and 400 psig.

But, this does not allow very many parameters to be measured, nor does it allow access

to the pump interior for flow visualization. [6]

7
2.2 Description of MVP Pump

The pump used is a three stage vertical turbine pump manufactured by Baker

Hughes Centrilift. Their designation is MVP G470. The stages are stacked together and

held in place by a set of four threaded tie rods. Included with the pump was a coupling.

Each stage has a multi-vane impeller similar to that shown in figure 1.1. The

basic assembly is shown below in figure 2.1, with the pump intake on the bottom and the

outflow near the top. The coupling is shown at the very top.

Figure 2.1 Basic pump assembly

8
 Figure 2.2 Pump with pressure taps

To make the pump suitable for measurement, the manufacturer milled a slot in

the second stage volute and drilled holes on the opposite side for pressure taps. An

acrylic plug was made to fit into the slot forming a window into the second stage. Figure

2.2 above shows the pressure taps in the completed pump. Below (figures 2.3 and 2.4) is

a view of the second stage volute with the window slot milled into it and a view of the

assembled pump with the window slot.

9
Figure 2.3 Second stage volute with window slot

10
Figure 2.4 Window slot on assembled pump

11
 The acrylic plug is held in place with an assembly of square tubes and tie rods in

order to hold it in place under the high pump operating pressure. The plug is sealed by

an o-ring which fits into a groove machined in the plug itself.

The completed pump as installed is shown in figure 2.5. The inlet and outlet

assemblies and the coupling will be described in section three.

Figure 2.5 Complete pump assembly

12
2.3 Test Loop Requirements

The test rig configuration must allow a series of tests to be run on the pump

varying the pump speed, inlet pressure, flow rate and GVF as desired. This will provide

adequate data for pump verification and CFD validation. Table 2.1 shows the test matrix

to be used:

Table 2.1 Test Matrix

Speed (rpm) Inlet Pressure Inlet flow (BPD) Inlet GVF (%)
(PSI)
3600 100 35000 0+
3300 200 30000 5% increments
3000 300 25000 To max possible
2700 400 20000
2400 15000
1800 10000

BPD = Barrels per day, 1 BPD = .02916 GPM

For each speed, the pump was tested at all inlet pressure and flow combinations, starting
at 0% GVF and moving upward until the maximum GVF is obtained.
Data collected from each run included:
Water inlet flow rate, temperature and pressure
Air inlet flow rate, temperature and pressure
Pump inlet temperature and pressure
Pump outlet temperature and pressure
Differential pressure between pump inlet and outlet
Pressures within the pump volute
Pump speed
Shaft torque and axial loading

In addition, future work will include resistive tomography measurement of the

pump inlet flow and second stage volute GFV distributions. The current design allows

13
visual inspection of the pump inlet as well as air and water inlet pipes via sight glasses.

However, no data was taken from these sources.

14
 3 DESCRIPTION OF FACILITY

3.1 Overall Configuration

A closed loop arrangement was used due to the desire to have high inlet

pressures. A non-loop configuration would require a huge reservoir of high pressure air

and water, and would also produce enormous amounts of discharge.

The working fluids are air and water. In actual service, oil is the predominant

liquid. But, its use would complicate the test facility design. The facility uses a heat

exchanger to cool the test fluid (actually a surplus air conditioning unit). Oil cannot be

used in this unit, so an additional heat exchanger would be required to keep the process

fluid separate from the coolant if oil was used as the process fluid. Further, oil may be

difficult to separate from the air due to excessive foaming, making a flow loop difficult

to achieve. There is also the cost of such a large volume of oil and the issue of disposal.

Future work may include viscosity enhancing additives to achieve more oil like behavior

without the aforementioned difficulties.

In schematic form, the facility appears as figure 3.1:

15
Figure 3.1 Two phase flow pump test loop piping and instrumentation diagram, (P&ID)

The pump is supplied with two separate inlets, one for air and one for water as

well as a single outlet. Pneumatically controlled valves are placed at the inlets for the air

and water supply and at the pump outlet. These will allow the control of the inlet

pressure, flow rate, and gas volume fraction. The inlet valves are controlled

automatically via Labview to maintain a specified inlet pressure and liquid flow rate.

The water inlet valve will control the liquid flow rate via a Proportional/ Integral (PI)

controller. The air inlet valve will control the inlet pressure also via a PI controller.

Manually controlling (again via Labview) the outlet valve will change the GVF of the

inlet fluid. Opening the outlet valve will cause a drop in the inlet pressure,

causing the air control valve to open, increasing the GVF. Views of the pump flow loop

are included in Figures 3.2 to 3.4.

16
Figure 3.2 Plan and elevation views

17
Figure 3.3 Facility north view

18
Figure 3.4 Facility south view

19
 The pump itself is mounted on a derrick type stand with the motor on top. This

will allow ready access to the pump and can easily be adapted to accept other pumps of

similar design. The pump will be driven by a standard industrial vertical electric motor.

Using a sealed ESP motor is not required as the motor behavior is not being investigated.

Water is added to the system via the fill port located on the water inlet pipe. A

differential pressure gauge measuring the pressure difference between top and bottom of

the separator allows the water column height to be monitored when the pump is not

running. A fitting in the air inlet pipe allows a commercial portable air compressor to be

attached in order to pressurize the system.

3.2 Piping

The main piping runs are six inch schedule 40 for water and water/air mixtures,

and three inch schedule 40 for air. Flanges are ANSI class 600. These sizes were chosen

to allow for the required flow rates and pressures with the ability to extend the facility’s

capability to handle one-thousand PSIG inlet pressures. The pipe material is stainless

steel in order to minimize corrosion. All piping was fabricated and tested per industry

standards.

Some piping is of different sizing to match existing equipment. In particular, the

separator tank has a pair of ten inch class 300 nozzles for the inlet and outlet.

Beginning at the separator tank outlet, the first pipe is a fabricated elbow/reducer.

The reducer is concentric rather than the preferred eccentric design. The original

configuration called for a forged flanged elbow and separate eccentric reducer. This

proved impossible to procure, and fabricated elbows take more space than standard

20
forged elbows. As the tank and its stand were already installed, the reducer was changed

to a concentric design to compensate. Following the elbow is a butterfly valve. This

allows the downstream piping to be drained without draining the entire tank. An eight

inch strainer is next, with a section of eight inch pipe to clear the tank stand.

After the strainer, the next pipe has a reducer to bring the pipe to six inches, this

reducer is eccentric but, is installed inverted from normal to help make up the shortfall

caused by the fabricated elbow. This piece also has a pair of welded threadolets for

instrumentation, and also for the seal flush and water supply line. The water turbine flow

meter is next, followed by a straight six-inch pipe to the water inlet control valve. The

connection from the control valve to the pump inlet is a six- inch triple ply flexible metal

hose, as is the connection from the pump outlet to the outlet control valve. From the

outlet control valve, a straight run six-inch pipe connects to the final outlet pipe. This has

a reducer to change the pipe size to ten inches to match the tank, and a pair of elbows to

allow the pipe to reach the tangential separator tank inlet.

3.3 Pipe Supports

The prime concern for the pipe support designs was to ensure that the pipe

flanges would be stressed only by the fluid pressure, and would not support any weight

of either the pipes themselves or any fluid. Whenever possible, commercial components

were used for piping supports. The lower pipes i.e. the water inlet and air inlet pipes are

supported by saddle/base assemblies without the need for any custom fabrication. The

control valves, however, required some additional work so that they would remain

upright. This work is in the form of a plate, welded to the end of each saddle support and

21
drilled to match each valve’s flange bolt holes. In this manner, the flange bolts hold the

valve upright. For the smaller three inch air control valve, a pair of plates and L-beams

was fabricated to support it. These details are shown in figure 3.5.

Figure 3.5 Valve support stands

The outlet valve and piping required a higher elevation than was possible with

commercial buyout components. Two separate stands were designed, one for the outlet

control valve and one for the end of the straight run outlet valve. Each consisted of a pair

of vertical tubes welded to base plates. The tubes are supported by a set of turnbuckles

22
and rods and a set of channel irons. A threaded nut and sleeve assembly fits into the top

of each tube and in turn supports a saddle, allowing adjustment for height.

The supports near the separator tank proved particularly challenging to

implement (figure 3.6). After much thought, a column was devised which would be

mounted on the arms of the separator support stand. This would reduce the footprint of

the support assembly and reduce the amount of raw materials needed. The arms of the

support stand are cantilevered and are thus not the most efficient structural design.

However, analysis indicates that additional support is not needed.

The column is in the form of a 3.500” nominal diameter pipe with a welded base

which is bolted to the separator support stand. Braces support the column, and are bolted

to tabs which were field welded in place. The column has short sections of 4.000” pipe

welded in place which support rotating sections of pipe which serve as mounting points

for the supports themselves. The ten inch pipe support is in the form of a crane using a

gaff with a turnbuckle from which is suspended a clevis hanger. This allows two degrees

of freedom to position the hanger. In addition, the crane assembly provided a mounting

point for a chain hoist to lift the ten inch pipe into position.

The top of the column likewise has a mount for the small two inch clevis hanger

supporting the air pipe.

23
 Figure 3.6 Separator mounted supports

3.4 Cooling Loop

A cooling loop was added to the system to keep the water temperature to a

reasonable level. This is critical due to the use of Plexiglas components to allow for flow

visualization. Plexiglas loses strength rapidly with elevated temperatures. Uncooled, the

process temperature rises rapidly allowing for only short data acquisition runs.

The cooling loop is constructed mainly of existing components. Primarily, the

heat exchanger has been salvaged from the Turbomachinery Laboratory air conditioner

24
unit, which was replaced by a newer unit early in the project. A skid mounted

pump/motor assembly also came from a prior project.

The cooling system draws water from near the bottom of the separator tank,

draws it through a filter assembly and then through the heat exchanger. After leaving the

heat exchanger, the water is pumped back to the separator tank near the top, passing

through a turbine flow meter on its way.

3.5 Separator Tank and Stand

The separator tank is a refinery surplus item. It was chosen because of its low

cost and immediate availability. It is made of 304 stainless steel and is rated for 450 psig

working pressure. Capacity is 1760 gallons. The inlet is tangential and the interior is

equipped with a disk shaped sparger to aid in liquid/gas separation. Four mounting tabs

exist, equally spaced around the tank.

The stand was designed to support the tank from the existing tabs. Each tab has

its own a framed support the bottom of which has two foot plates that are grouted and

bolted to the underlying surface. The top of each frame is a long, square tube allowing it

to be bolted to the adjoining frames. An FEA(figure 3.7) analysis was performed on the

stand to ensure it met strength requirements.

25
 Figure 3.7 Separator support stand

Here the stand is shown loaded with the full weight of the tank filled with water

and a one thousand pound force side load. Note that results within one diameter of a hole

edge can show unrealistically high stress levels due to limitations of the FEA software

(Solidworks simulation)

3.6 Pump Stand

The pump is mounted in a derrick (figure 3.8) with the motor on top. This design

allows ready access to the pump and can easily be adapted to accept other pumps of

similar design. A key design feature is the ability to be disassembled as the stand is too

26
large to fit through the doors in the test cell when assembled. Therefore, the stand is

designed to bolt together. The cross members are also threaded to allow some minor

adjustment.

Figure 3.8 Pump stand

The motor base is a welded piece of one half inch steel plate and steel tubing

while the base is welded steel tubing. The other components are all bolted together as

27
 Figure 3.9 Detail of cross brace fittings

shown. The cross braces on any one side can be removed if needed whether or not the

motor or pump is in place. This is needed for removing or installing the pump itself. A

detail of the cross brace threads is shown above (figure 3.9).

In order to mount the pump, a girder assembly (figure 3.10) is bolted to the

bottom of the motor stand after the pump is maneuvered into position. The pump is

attached to the girder assembly via the holes in the mounting blocks at the bottom of the

girder assembly. The bottom of each girder has a block welded into it with a threaded

hole in it. This allows the mounting blocks to be bolted to the girders. Inserting shims

28
between the girder and mounting blocks allows the angular alignment of the pump to be

adjusted. Plates are bolted to the girders to provide additional stiffness and to serve as

coupling guards.

Figure 3.10 Pump mount

Figure 3.11 shows the pump when mounted. The motor base of the pump stand

and some of the plates are not shown so that the mounting arrangement can be seen.

29
 Figure 3.11 Pump head

The top part of the MVP pump is seen in green at the bottom of the picture.

Proximity probes to monitor the shaft position were added after the pump began

operating as will be described under initial operating experiences.

30
3.7 Pump Inlet

Figure 3.12 Inlet assembly

The pump inlet (figure 3.12) was fabricated by a vendor from commercial elbows

and flanges along with several custom machined parts. The double elbow design

provides clearance for the six inch flange while keeping the overall height of the

assembly as low as possible. The air inlet flange simply fits on the side of the inlet with

plenty of room.

The original concept called for a perforated tube to be inserted into the inlet via

the air inlet flange. The idea was to evenly distribute the air and mix it with the water

prior to entering the pump. This proved to be unnecessary and more difficult to model

with CFD. So, it was not implemented.

31
 Between the inlet and the pump is a tubular polycarbonate window allowing

visualization of the flow into the pump (figure 3.13).

Figure 3.13 Inlet window at 1000 PSIG and ambient temperature

This was expected to be a weak point in the design as the rest of the pressure

boundaries are made of steel. However, the window’s ability to withstand the loop

operating pressure is clearly adequate.

32
3.8 Pump Outlet

Figure 3.14 Pump outlet, seal, and coupling cross section

The pump outlet is an assembly that consists of a machined body to which is

welded the outlet flange via a commercially available weldment. The body is bolted to

33
the top of the pump and is sealed to the pump by means of an O-ring. The body also has

a stuffing box machined into the top for a mechanical seal, a bore for a throat bushing,

and a port for instrumentation. Figure 3.14 shows the outlet assembly.

3.9 Seal Assembly

The seal is a Flowserve QB-2000 cartridge commercially available single inside

mounted end face design (figure 3.15). The basic seal is composed of a ring shaped

stationary face against which runs a rotating face. The rotating face is spring loaded to

ensure contact with the stationary face. In addition, hydraulic forces act to force the

faces together when the stuffing box is pressurized. In fact, this particular seal has a

shoulder on the sleeve underneath the rotating face to reduce the hydraulic pressure. In

industry terms, it is balanced.

All of the basic seal components are contained in a cartridge assembly with a

sleeve and gland so that the seal may be easily installed and removed from the pump.

The cartridge also includes a gland bushing (sometimes called a disaster bushing). The

bushing serves to protect the seal components from damage if the shaft should become

misaligned as the sleeve will rub against the bushing before contacting the brittle

stationary face. It also reduces the amount of leakage should the seal fail. A drain port

has also been drilled into the gland (not shown on the figure). It is located at the same

location axially as the flush ports and rotated ninety degrees as seen from the top. The

drain has a thru hole situated between the stationary face and the flange bushing so that

any leakage from the seal will be directed out through the drain port rather than be an

uncontained flow past the bushing. In this assembly, the drain port is plumbed to a short

34
piece of tubing terminating in easy view of the operator. This allows any leakage to be

detected before becoming a serious problem.

Figure 3.15 Seal assembly

This seal configuration is common in industry [7] for is simplicity and reliability.

However, for best functioning it requires the proper environment. Typically, this is done

by providing a seal flush. A number of configurations have been standardized by the

American Petroleum Institute. [8] The most common is the API Plan 11, a small amount

of fluid is drawn from the pump discharge and routed into the seal chamber. The flush

35
cools the fluid and keeps it from vaporizing across the seal faces. In vertical pumps a

Plan 13 is common; a flush line runs from the seal chamber back to the pump inlet rather

than the discharge. This will tend to draw any vapor bubbles out of the seal chamber as

the chamber is typically at the pump discharge pressure. In both cases, the pressure

gradient between the chamber and the discharge or inlet provides the motive force for

the flush. Combining the two gives a plan 14, a modified version of which has been

employed here.

The flush is of particular concern for this project due to the multiphase nature of

the flow. Mechanical seals for liquid service can suffer damage if operated in a vapor

environment. In order to ensure a healthy seal environment, a modified plan 14 is used.

One flush line comes from the water inlet and is run through a pump and regulator ( a

normal plan 14 flush has no pump or regulator) to provide flow into the seal chamber

while another line runs from the chamber to the top of the separator [Link] allow any

vapor to escape as in a plan 13. A close clearance bushing is press fit into the bottom of

the stuffing box to limit the amount of flush flowing into the process fluid. This ensures

a cool, pure liquid environment around the seal.

3.10 Coupling

The coupling assembly, save for the spacer, was provided by Baker Hughes

Centrilift. As supplied, the coupling consisted of three main components. The lowest

piece slides onto the pump shaft and is keyed to transmit torque. The next piece is

threaded to match the pump shaft. Turning this part changes the shaft axial position and

thus sets the impeller spacing for the pump. A spacer bolts to the threaded, and pump

36
pieces. Removing it creates a coupling gap through which the seal can be installed or

removed without disturbing the pump itself. Figures 3.16, 3.17 and 3.18 show the spacer.

Figure 3.16 Coupling spacer

The spacer also serves as the point for torque and axial load measurements.

Strain gauges configured for axial and for torque measurements are mounted on the tube

of the. A pair of Binsfeld transmitters are mounted with the strain gauges to allow real

time reading of the axial load and torque. The spacer has been designed to provide

37
adequate strain response for the expected loadings. Further, a separate test fixture has

been made so that the entire spacer, strain gauge and transmitter assembly can be

calibrated and tested off of the pump assembly.

Figure 3.17 Coupling spacer strain under 3000lb axial load

38
 Figure 3.18 Coupling spacer strain under 4380lb-ft torque

The coupling was sized to provide an even, measurable strain in response to the

expected torque and axial loading.

3.11 Operating Instructions

Checklists for before operation and after operation are given in the appendix. In

this section, those checklists are annotated in italics. Refer to figures 3.3 and 3.4 for

locations of components

Before operation checklist:

39
Piping and flanged connections – check for integrity
Flange bolting should be tight for every nut and stud.
Tubing – check for integrity:
Water supply to inlet pipe
Inlet pipe to seal flush pump
Seal flush pump to seal
Seal flush outlet to top of separator tank
Instrumentation tubing
Tubing is especially prone to damage.
Variable frequency drive – covered
The variable frequency drive is mounted on the north wall of the test cell
adjacent to the pump stand. It is fitted with a sheet steel cover to protect it from
spraying water. The cover must be in place with its top cover and access panel
closed.
Instrumentation wiring – check for integrity
Wires must be fully plugged in to function.
Labview valve control program – running
This is run from the computer at the southwest corner of the control room. When
it is running, the instrumentation can be checked.
Water inlet control valve – open (via Labview)
Outlet control valve – open (via Labview)
Pump Outlet/seal area– clear of debris, loose items etc.
Look into the area around the seal. Tools and loose items can congregate there.
Coupling strain gauge transmitters – connected to batteries.
The battery wires are connected to the transmitters with a miniature screw.
Coupling guards – installed and bolts tightened
IMPORTANT – do NOT run the pump without the coupling guard in place except
at very low speeds (90 RPM or less).
Water drain – closed
This valve drains the water out of the system. It must be closed in order to
pressurize the system.
Air inlet shut off valve – closed
If this valve is open, water could enter the air inlet pipe, possibly damaging the
flowmeter.
Instrument air supply – open
This is located on the wall near the control valves and supplies air to operate the
control valve actuators.
Air supply line sight glass – check for water
There should be no water in the air line. If water is present, see below.
Air drain – open to drain any water, then closed
Use this to drain any water out of the air inlet line as needed.
Separator tank outlet valve – open
This valve must be open to run the pump. It can be closed to drain water out of
the pump without draining the tank.
Water supply valve – open (if needed to fill separator tank)

40
 See below.
Separator tank water level - 10 to 12 feet.
Water level is indicated on the Labview control panel.
Pressure relief valve – closed
Located on top of the tank. It must be manually closed as the return spring is not
strong enough to close it.
Tank air pressure – set (at least 120 psig)
Attach the portable air compressor to the air inlet connection with the high
pressure yellow hose. Run the compressor until the desired pressure is reached.
Disconnect the air compressor when done. Disconnect the hose at the inlet pipe
before disconnecting the hose from the compressor.
Proximity probe power supply – on
The power supply is a box mounted above the window next to the door into the
control room.
Shaft alignment – within .005” each axis
Shaft alignment is displayed on the Labview control panel.
Seal flush pump – on
The switch is located on the pump itself.
Manual transfer switch - In up position for the MVP pump.
Located on the side of the electrical panel next to the VFD.
Safety switch of Variable frequency drive –on
VFD handheld controller – Switch to LOCAL
The handheld controller is located in the control room.
The test loop is now ready to run.
VFD handheld controller- Press the RUN bottom

After Operation Checklist:

Air inlet shut off valve – closed (while pump is running)

This will prevent water from entering the air line.
VFD – Press the STOP bottom
Variable frequency drive’s safety switch – power off
Seal flush pump – off
Air inlet line sight glass – check for water
There should be no water. If water is present, drain it as noted below.
Air drain – open to drain any water, then closed
Proximity probe power supply – off
Instrument air supply – off
Secondary loop’s pump- off
This is the heat exchanger pump.
Secondary loop’s heat exchanger fans- off
The switch for this is located on the heat exchanger on the side furtherest from
the test cell.
Secondary loop’s inlet and outlet valves-off

41
The loop may be kept pressurized in this condition if it is anticipated that it will be run
again shortly, i.e. within a week.
If the system will not be run for an extended time, or if it is necessary to work on any
pressure containing component, the system should be depressurized and drained.

To depressurize:
Instrument air supply – on
This must be on to open the control valve.
Outlet control valve – open (via Labview)
Pressure relief valve – open
Simply pull on the chain to the valve to open. Note, the exiting air will make a
loud noise for several minutes.
Pressure – monitor until pressure is at atmospheric
This can be done via Labview or one of the mechanical pressure gauges.
To drain:
Separator tank outlet valve – open to drain entire water supply, closed to drain only the
pump and piping.
Usually, the tank will not have to be drained, so this valve will remain closed.
Water inlet control valve – open (via Labview)
Water drain – connected to hose leading to storm drain.
The water drain has a quick connect fitting for this purpose.
Water drain – open
When system is drained:
Water drain – closed
Instrument air supply – off (control valves will close)
Separator tank outlet valve – closed
If previously opened.
Pressure relief valve - closed

42
 4 CONSTRUCTION AND COMMISSIONING

4.1 Separator Tank and Stand

The separator tank was the first major item to be installed. The tank itself was

delivered in the summer of 2010. Due to its previous use and long period of outdoor

storage, some preparation work needed to be done. Surface scale, old paint, and

corrosion, was removed by grinding. Also, one of the support tabs was corroded beyond

use and was replaced. These tabs are made of carbon steel as opposed to the stainless

steel employed in the rest of the vessel. When first inspected for use in this facility, the

tank had been lying on its side in an unpaved yard positioned so the corroded tab was in

the dirt. This no doubt was the cause of the excess corrosion. To replace the tab, a

wooden jig was made to locate the tab relative to the other tabs. Then, the tab was

removed with a cutting wheel and the jig used to fit the replacement piece in place. The

other tabs also showed some corrosion, but, not enough to require replacement. Other

than that, the tank was found to be in very good condition despite its age and relative

neglect.

While the tank was being readied, the stand was being fabricated by an outside

contractor. Also, the site was prepared by clearing the area of existing equipment and

preparing forms for grouting. The site was marked as to the location and orientation of

the tank. Wooden forms were then placed in the proper location and sealed with

caulking. A great deal of effort was spent to ensure a watertight seal between the forms

and the concrete site pad. This was shown to be unnecessary due to the grout’s

extremely high viscosity.

43
 When the stand was finally delivered, it was cleaned and painted at the

Turbomachinery Laboratory and assembled inside the main bay using the large overhead

crane located there. The completed assembly frame was moved to location via forklift

and placed over the forms. The frame was supported by shims to keep the weight off of

the forms and to facilitate leveling. With the frame in place, the forms were filled with

grout and eventually, one inch diameter anchor bolts were installed.

With this done, the tank was ready for installation. An outside rigging contractor

brought a crane to emplace the tank. This was done on January 12th, 2011. All openings

were then sealed with plastic covered plywood disks to protect against the weather.

This was the only outside help required for the erection of the facility. All other

installations were done entirely with Turbomachinery Laboratory resources.

4.2 Piping and Control Valves

In general, piping was installed by moving it into place and bolting the flanges in

place. Only when the pipe was in place were the flange bolts loosened or removed to

allow the flange gasket to be installed. Doing the initial installation without the gaskets

prevents them from being crushed due to uneven loading. In all cases, the final flange

bolting was done per industry standards. That is, the bolts are torqued in intermediate

steps to the final value. Each time the bolts are tightened, it was done in an alternating

pattern to ensure even tightening of the flanges.

The largest and most difficult section of pipe was installed first. This was the ten

inch diameter angled pipe that leads directly to the separator tank tangential inlet

(drawing: Spool-7). This is the heaviest pipe section. It also needed to be lifted a

44
considerable distance. Moreover, with other pipes in place, it would be difficult to reach,

which is why it was chosen to be the first piece installed.

The first step was to emplace the tank mounted supports for the pipe. To do this,

a temporary crane was fabricated by bolting a timber across the top of the tank. This

timber was braced with a number of cargo straps to support the required loads. This was

used to raise the main support column into place. Figures 4.1 through 4.4 show this

process.

Once in place, the braces were fitted into place and their tabs welded. The braces

and main support column were then taken down, painted and reinstalled.

With the support in place, a chain hoist was attached and used to lift the top end

of the pipe spool into place. The lower end was lifted with a portable jib crane. With

both ends in place, the upper end flange was bolted to the tank and the lower end’s

support was erected in place, leaving the pipe fully supported.

45
Figure 4.1 Main support column being raised

46
Figure 4.2 Rear of tank showing the temporary crane arrangement

47
Figure 4.3 Temporary crane assembly mounted to top of separator tank

48
Figure 4.4 Ready to lift spool seven

49
 The adjoining pipe section (Spool-6) was then lifted into place with the same jib

crane and bolted into place.

The next part was the outlet control valve. With a weight of eight hundred

pounds, this presented another challenge. Two portable cranes were used. One lifted

each end of the valve. During lifting, the valve was supported on each end by a sling

which went under the body of the valve and was attached to one of the cranes. When the

slings were tensioned prior to lifting, cables were wrapped between the slings and the

stem of the valve to provide stability. Both cranes were then used to lift the valve into

place.

Directly below the tank is attached a large elbow leading to the water inlet

piping. This was installed by maneuvering the elbow underneath the tank. This required

lifting the two hundred pound fitting about three feet to clear the tank support struts and

pulling it underneath the tank. The lifting was provided by an engine hoist and the

pulling by a block and tackle rigged between the elbow and the flange at the bottom of

the tank. Once there, the elbow was lifted into place by using threaded rods as jacking

screws.

The inlet water pipes and valve were emplaced using the portable cranes with

wooden blocks to provide temporary support. When all of the piping was in place, the

permanent supports were fitted into place and welded.

The inlet air pipes were light enough to be fit into place by hand with no special

lifting equipment. The air inlet control valve weighs three hundred pounds. However, it

did not need to be lifted any distance. It was simply lifted upright on its support and

pushed into place.

50
 At this point, it was envisioned to pressure test the tank and piping assembly.

However, this proved to be impossible due to the fact that the control valves were not

designed to hold pressure. They will leak up to five percent of their maximum flow rate

even when fully closed. Pressure testing would have to await the completion of the full

pressure loop.

In the meantime, all of the pipe supports were secured to the floor via wedge

anchors. The actual location of the components were measured and compared to the

design specifications. The inlet water pipe and outlet pipe were seen to diverge at an

angle of about five degrees. Plans called for these pipes to be parallel, and it appears the

discrepancy was caused by the separator tank mounting tabs not being positioned in the

correct location. This required a slight redesign of the flexible outlet hose, resulting in its

having an s-curve shape.

4.3 Motor and Stand

One of the most difficult tasks contemplated in this project was the installation of

the electric motor. This was due to its position on top of the pump stand and lack of

clearance between the top of the motor and the ceiling of the test cell. Only twelve

inches were available, and to get this much requires removing the light fixtures. Clearly,

raising the three-thousand pound motor some eight feet in the air under these constraints

was going to be quite a challenge.

The solution involved the use of a forklift. One was available which was capable

of fitting inside the test cell control room. From this point, the forks could reach into the

test cell through the main doors, although the forklift itself could not fit. Therefore, the

51
motor was bolted to its base outside the test cell and carried via forklift inside the cell.

At this point, the motor/base could be raised high enough to attach the remaining pump

stand components. The stand was now assembled just inside the test cell. But it was

positioned adjacent to the doorway, not where it was needed. The forklift could only

reach just inside the cell. The assembled stand would have to be moved. This was

accomplished by raising the entire structure onto rollers. Temporary crossbeams were

attached to the frame base with bolts that were turned to lift the structure. Each

crossbeam had a roller positioned under it, so that as the bolts were tightened, the frame

was lifted off of the floor to rest solely on the rollers. At this point, the entire assembly

was easily moved into position.

The variable frequency drive (VFD) for the motor was attached to the test cell

wall. Wiring for the motor and VFD was done by a certified electrician.

4.4 Pump Installation

Removing the cross members on one side of the pump stand opened up enough

space to roll the pump into position in the vertical orientation. So, the pump was placed

on a dolly vertically with the outlet assembly bolted into place. Also in place, were the

mounting blocks and bars that provide the transition from the pump to the pump

mounting. Elevated boards were laid across the base of the pump stand to allow the dolly

to be rolled over the structural members that formed the base. With the pump rolled

under the stand, the mounting girders were bolted onto the bottom of the motor base.

Bolt could then be inserted through the pump mounting blocks and into the mounting

girders allowing the pump to be lifted into position by tightening the bolts.

52
 The pump inlet was then bolted to the bottom of the pump. The remaining piping

components were then installed.

Aligning the motor and pump was the next task. This was done per the procedure

outlined in the appendix. It is important to note that the alignment was between the

motor shaft and the pump case. It was not done between the motor shaft and pump shaft

since the rigid coupling provides for this. Initially, no provision was made for

monitoring the alignment during operation.

After that, the seal and seal support system was added. This involved sliding the

seal assembly onto the pump shaft and bolting the flange into place. Then the coupling

was installed. The uppermost part of the coupling was bolted to the motor shaft. The

lowermost part was slid onto the pump shaft and the disk threaded onto the end of the

pump shaft far enough to allow the spool piece to be inserted. The spool was then

inserted and bolted to the uppermost piece. The disk was then rotated until the gap

between it and the bottom of the spool was one sixteenth of an inch and the motor shaft

rotated until the bolt holes were aligned. The spool, disk and lower part were then bolted

together, which raised the pump shaft by one sixteenth of an inch, setting the impeller

clearance.

The tubing for the seal flush was installed using compression fittings. All

pressure and temperature sensors were fitted and wired into the controlling computer.

4.5 Initial Operating Experiences

Initial operation was done with water only. The entire system was filled with

water until the pressure was equal to the water main pressure, 100 psig. All joints and

53
fittings were checked for leaks. The motor was then run at 1800 rpm briefly while

monitoring the pump for any unusual vibrations or other problems. None were detected.

Water was drained until the separator tank was half full and the system pressurized with

the air compressor.

The pump was then run at a series of increasing speeds to give a baseline pump

performance curve. After several hours of operation, the seal began to leak. The pump

was stopped and the system depressurized and drained. A visual inspection revealed that

the seal sleeve had rubbed against the bushing, indicating an out of alignment condition.

Upon further disassembly and inspection, the seal faces were seen to have broken. The

seal was returned to the vendor for repair. Investigation of the pump showed that the

misalignment was caused by the pressure in the inlet line pushing against the bottom of

the pump. To combat this, stabilizer blocks were added to the bottom of the inlet

assembly. These were bolted to the floor and had adjusting screws to hold the pump inlet

in position.

In addition to this, proximity sensors were added to the shaft at the seal. This

allowed continuous monitoring of the alignment while the pump was running. It also

allowed the alignment to be adjusted after the pump was pressurized. With these changes

in place, the system was run with no more alignment problems.

After running with pure water, air was added to the inlet stream. Very low air

flow rates, however, did not supply sufficient pressure to open the check valve that was

in the air inlet line. Pressure would build up until the valve opened and let a bubble of air

into the pump. This lowered the pressure enough to close the check valve. Pressure

would build up and the process repeat, resulting in a surging of inlet gas volume fraction

54
rather than a steady state condition. To prevent this, the check valve was removed and a

manual valve substituted. This valve was opened by hand after the system was running

and enough air pressure existed to keep water from back flowing into the air supply

pipes. These pipes were also modified by adding a sight glass to verify that no water was

in the line and a drain to remove any water that accumulated.

Operations were continued long enough to obtain a complete set of data for the

100 psig and 200 psig inlet pressure conditions, and were continuing onto higher

pressures when the motor suffered a critical failure. This failure was apparently caused

by an aftermarket shaft sleeve that was installed inside the motor at some point prior to

its acquisition by the Turbomachinery Lab.

Removal and reinstallation of the motor was a difficult task due to its three

thousand pound weight and the limited clearance between the motor and the ceiling. A

pair of overhead gantry cranes was erected inside the test cell, one on each side of the

motor. Each gantry had a chain hoist that was connected to the pump via a frame

assembly.

Due to the limited space, the chain hoist hooks could not be used. Instead, the

hooks were removed from the chain and the chain passed through the frame and a toggle

bolted to the chain. The resulting system allowed the absolute maximum amount of

space for lifting the pump. This is shown in figures 4.5 and 4.6.

55
Figure 4.5 Lifting frame

56
 Figure 4.6 Lifting frame toggle

4.6 Initial Data Collection

My colleague Sahand Pirouzpanah is in charge of pump testing and

instrumentation. Complete details of the flow loop instrumentation and testing

procedures will be available in his dissertation. This is expected to be completed in

December of 2012. All of the graphs in this section (figures 4.7 through 4.10) are

courtesy of Mr. Pirouzpanah.

The pump manufacturer supplied reference data for the differential pressure and

flow rate for pure water. Initial test loop data was plotted against the manufacturer

supplied data as shown:

57
 GVF=0, Pure Water
250
Pump Differential Pressure(psi)

200

150 Centrilift Data-

3500RPM
100
Turbolab's
50 Experimental
Results-3600 RPM
0
0 1000 2000 3000
Water Flow Rate(GPM)

Figure 4.7 Comparison of Experimental data to pump supplier data

The data are in close agreement with the experimental data being slightly higher.

This can be attributed to the slightly higher pump speed. Some data for the one hundred

PSIG inlet pressure for varying GVF are shown below. Data collection for higher

pressures is continuing at the present time.

58
Figure 4.8 Data for 100 Psig inlet pressure at 3600 rpm

59
Figure 4.9 Data for 100 psig inlet pressure at 3300 rpm

60
Figure 4.10 Data for 100 psig inlet pressure at 3000 rpm

61
 In general, the curves look similar to that which would be expected for normal,

pure liquid pumps. The pressure generated decreases with increasing flow rate, while

power consumption grows larger. Efficiency peaks at around 1400 GPM at 3600 rpm.

This is higher than the 3300 and 3000 rpm case, where efficiency reaches its maximum

at 1000 GPM.

Increasing the GVF lowers efficiency, power output, and differential pressure.

This would be expected as energy is being expended in compression of the gas rather

than increasing the head of the liquid.

Changing from a pure liquid flow to multiphase flow does have a startling effect

on the differential pressure. For liquid, the differential pressure decreases almost in a

linear fashion with flow rate. When gas is added, however, the differential pressure

peaks at just over 800 GPM. This effect grows more pronounced at slower speeds and at

lower GVF.

62
 5 SUMMARY AND CONCLUSIONS

5.1 Summary

A two-phased flow pump test loop has been constructed to test a three stage

vertical pump. The loop is capable of handling flows of up to twelve hundred gallons per

minute at three hundred PSIG inlet pressure.

The loop was run initially with only water. Pure water (i.e. 0%bGVF) data is

very close to manufacturer supplied information.

Operating data was then taken for one hundred PSIG inlet pressure at speeds

from 3000 to 3600 rpm. In each case the gas volume flow rates were varied from zero to

the maximum possible. In these cases, the maximum attainable GVF was 30% to 35%.

Data for higher inlet pressures are presently being collected.

5.2 Conclusions

Throughout construction, many difficulties were encountered and solved. Many

of these can serve as valuable lessons learned when engaged in future projects.

The most significant problem has been maintaining the alignment between the

pump and motor. This alignment must be closely monitored for this type of installation.

For future projects, a stiffer motor base and pump mount should be considered as well as

a means to secure the inlet of the pump to prevent movement due to pipe stresses. An

alternate design concept would be to provide a gimbaled mount for the motor and pump

so that they move together. Such a design would be more difficult to achieve than the

current one. But, it may be worthwhile to consider for future projects.

63
 A lesser problem has been sealing around the second stage pump window plug.

Initially, a flat gasket was used. But, this proved inadequate. An o-ring gives a much

better seal as long as a proper o-ring groove can be used.

Sight glasses were not part of the original design. One was added to each inlet

pipe to observe the flow. In the case of the water pipe, it allows operators to verify that

the separator is functioning properly and supplying only water through the water inlet.

The air inlet line sight glass is used to verify that there is no back flow of water from the

pump. This could cause a problem if water gets to the air flow meter. Water will damage

the flow meter.

In specifying flanged connections, it should be noted that ‘slip on’ flanges are

actually welded into place while ‘lap joint’ flanges are free to rotate for installation

reasons.

Another industry standard is the fact that control valves do not seal entirely when

closed but will leak about five percent of the maximum flow rate. They can be made to

seal completely if this is specified when the valve is ordered.

5.3 Future Work

Electrical Resistive Tomography (ERT) will be used to analyze the flow. ERT

uses a series of electrodes on the perimeter of a pressure vessel that make contact with

the process fluid. An electric current between pairs of electrodes and the voltage

difference between the remaining pairs is measured. This is repeated for each pair of

electrodes and produces a map of the electrical conductivity of the fluid. Since

conductivity will vary from different species or phases, a map of the flow conditions can

64
be made. In this case, a map of the air and water phases of the inlet or interior pump flow

can be seen.

A set of sixteen electrodes has been mounted onto the pump inlet window to

allow ERT measurements to be made at the inlet. A new insert to fit into the second

stage volute has also been made. This insert has its own electrodes on the interior surface

which will allow data collection of the second stage volute flow. ERT testing is expected

to begin soon.

In addition to ERT testing on this pump, a different pump of similar

configuration is likely to be tested in the near future. This is an ASM series 1025P ARC

model by the same manufacturer. This pump is slightly longer than the existing pump so

that the inlet configuration will have to be modified.

Another project which will use some of this loop is a polycarbonate pump. A

pump will be made with a casing of polycarbonate. This will allow unprecedented visual

access to the operating pump. An additional set of valves and tee fittings will be added to

the loop to allow the new pump to use the existing separator tank, control valves and

piping.

Finally, an erosion pump project has been initiated which will measure the

effects of entrained sand on pump internal components. Although this does not use the

same facilities, the design of the pump stand is identical to that for this project with

minor changes due to different pump and motor dimensions.

65
 REFERENCES

[1] Heald, C. C. editor, 1988, “Cameron Hydraulic Data,” Ingersoll-Rand Company pg.
1-42.

[2] Baker Hughes Centrilift, 2009, Claremore, OK , “Centrilift Submersible Pump

Handbook Ninth Edition,” pg.47.

[3] Burliegh, Lawrence, 2009, “Pushing the Limit: Taking ESP Systems to 250º C,”
Baker Hughes Centrilift, Claremore, OK pg. 3.

[4] Gruselle, F, Steimes, J., and Hendrick, P., 2010, “Study of a Two-Phase Flow Pump
and Separator System,” Proceedings of ASME Turbo Expo 2010: Power for Land, Sea,
and Air (GT2010) pg. 5.

[5] Gamboa, J. , and Prado, M., 2010, “Visualization Study of Performance Breakdown
in Two-Phase Performance of an Electric Submersible Pump,” 26th International Pump
Users Symposium Proceedings pg. 4.

[6] Burliegh, Lawrence, 2009, “Pushing the Limit: Taking ESP Systems to 250º C,”
Baker Hughes Centrilift, Claremore, Oklahoma pp. 1-3.

[7] Hydraulic Institute, 2006, “Mechanical Seals for Pumps: Application Guidelines,”
Hydraulic Institute pg.39.

[8] Hydraulic Institute, 2006, “Mechanical Seals for Pumps: Application Guidelines,”
Hydraulic Institute pp. 139-220.

66
APPENDIX A

PROCEDURES

67
A-1 Alignment procedure

The pump is equipped with two proximity probes that allow the alignment to be

monitored even when the pump is running. They check the axial alignment of the pump

shaft to the seal housing. They do not directly measure angular alignment.

The proximity probes must be calibrated before use. Once calibrated, they do not

need to be recalibrated unless they are disturbed. Calibration involves aligning the pump

using ‘traditional’ means. i.e. dial indicators. This is a lengthy process. So, do not disturb

the proximity probes unless absolutely necessary!

A-1.1 Calibration (do this only if the proximity probes have been disturbed)

To align the pump using dial indicators, first depressurize the system. Failure to

do this could cause injury and damage to the equipment. Then remove the coupling

guard nearest to the control room door. Then remove the coupling spacer and bottom

components of the coupling. It is not necessary to remove the coupling piece that is

attached to the motor shaft. Now remove the seal by disconnecting the tubing fittings

and replacing the setting plates into the seal sleeve grooves. (Refer to the seal assembly

drawing) Loosen the drive collar set screws and the flange nuts. The seal should slide up

the shaft and can be removed through the coupling gap.

The dial indicator holder can now be attached to the motor shaft and the dial

indicator set to read the top surface of the pump outlet. This will indicate the angular

alignment of the pump. Set the dial on the indicator to zero and carefully rotate the

motor shaft through one revolution by turning the motor coupling. Avoid using the dial

68
indicator holder to turn the motor. Ensure that the dial indicator probe travels freely over

the pump outlet surface. Ideally, the indicator should read zero for the entire shaft

revolution. If it does not, the alignment can be adjusted using shims at the pump mount.

To determine the needed shims, set the dial indicator to zero when it is positioned near

one of the four mounting girders. Rotate the motor shaft and record the dial readings at

each girder. Compare each reading to the one that is opposite. The shim thickness is

nearly equal to the difference between the two values and is placed under the girder

where the dial indicator value is higher. Example: the girders are on the northeast (NE),

southeast (SE), southwest (SW) and northwest (NW) of the pump. Rotate the dial

indicator until it is on the northwest and set it to zero. Rotating clockwise, we get the

following values: NW: 0, NE, +.010”, SE +.015”, SW -.020”. Southeast is directly

opposite northwest and it is a higher value (+.015 as opposed to 0) so a .015” thick shim

should be tried at that point (since +.015 – 0 = +.015). Similarly, a .030” shim should be

tried under the northeast since +.010 – (-.020) = .030.

To add a shim, loosen all of the 7/8” main attach bolts sequentially enough to

allow the shim to be slipped into position. Then tighten all of the bolts and check the

alignment. It may require several iterations to get the alignment close to zero. The dial

readings should be no more than .002” when the pump is properly aligned. Once the

pump is aligned angularly, the axial alignment should be set.

Remove the dial indicator and its holder and slide the seal back into place. Bolt

the flange into place without the proximity probes or probe mounts. Reinstall the

coupling and set the impeller height. Tighten the seal drive collar set screws and remove

69
the setting plates. Do not throw the setting plates away! Put them in the box the seal

came in. Do not install the tubing or tubing fittings into the gland yet.

Mount the dial indicator holder on the coupling and set the dial indicator to read

the outside diameter of the seal gland near the top of the seal. Rotate the shaft through

one revolution and ensure that the indicator probe slides freely over the gland surface.

Rotate the shaft until the indicator is even with one girder, set the reading to zero and

rotate the shaft through one revolution, recording the readings at each corner. The axial

shaft position is controlled by moving the motor using the positioning blocks located on

top of the motor base. Each is equipped with two small threaded bolts which can be

turned to move the motor away from that block. The motor should be moved a distance

of one half the difference between two opposite sides in the direction of the higher

number. That is away from the lower number. To move the motor, loosen all four main

motor mount studs holding the motor onto the base plate. Then loosen all of the

positioning blocks bolts to allow the motor to move. Move the dial indicator until it is

even with the block the motor is to move away from. Set the dial to read the negative

value of the distance that the motor is to be moved and slowly and evenl y tighten the

two positioning block bolts until the dial indicator reads zero. Repeat for the other

direction if needed. Then tighten the main motor mount studs. Recheck the alignment.

Repeat the process as necessary until no more than .002” runout is indicated.

Remove the dial indicator and its assembly. Attach the proximity probe mounts

and the probes onto the seal gland. Adjust the probes until the tips are about .050” from

the coupling. Connect the probes to the converter box mounted near the door. Turn on

70
the box and refer to the Labview valve control program. Set the alignment readings in

Labview to read zero. The pump is now aligned and calibrated.

A-1.2 Alignment

After the proximity probes have been calibrated, the pump can be aligned

without dismantling the coupling or seal. Simply loosen the main pump mounting bolts

and move the motor as needed referring to the Labview readouts to check the alignment.

Note that the readings may shift some when the mounting studs are retightened. The

stabilizer blocks at the bottom of the pump inlet assembly can be used to tweak the

alignment after the studs have been tightened. Also, with some experience, it is possible

to anticipate how the alignment will shift when the studs are tightened and to

compensate accordingly.

A-2 Before operation checklist

Piping and flanged connections – check for integrity

Tubing – check for integrity:

Water supply to inlet pipe

Inlet pipe to seal flush pump

Seal flush pump to seal

Seal flush outlet to top of separator tank

Instrumentation tubing

Variable frequency drive - covered

71
Instrumentation wiring – check for integrity

Labview valve control program – running

Water inlet control valve – open (via Labview)

Outlet control valve – open (via Labview)

Pump Outlet/seal area– clear of debris, loose items etc.

Coupling strain gauge transmitters – connected to batteries.

Coupling guards – installed and bolts tightened

Water drain – closed

Air inlet shut off valve - closed

Instrument air supply - open

Air supply line sight glass – check for water

Air drain – open to drain any water, then closed

Separator tank outlet valve – open

Water supply valve – open (if needed to fill separator tank)

Separator tank water level - 10 to 12 feet.

Pressure relief valve - closed

Tank air pressure – set (at least 120 psig)

Proximity probe power supply - on

Shaft alignment – within .005” each axis

Seal flush pump - on

Manual transfer switch - In up position for the MVP pump.

Safety switch of Variable frequency drive –on

72
VFD handheld controller – Switch to LOCAL

The test loop is now ready to run.

VFD handheld controller- Press the RUN bottom

To fill water tank, ensure pressure in tank is below 100 psig. The shop water

supply is about 100 psig, so if the tank pressure is higher, no water will flow into the test

loop. Then open the Separator tank outlet valve and the water supply fill valve. Monitor

the water level. Do not allow it to climb higher than 14 feet. Turn the water supply fill

valve off to read the water level gauge. It will not read properly with the water supply

flowing. When the proper level is reached, close the supply valve.

Tank pressure – check, should be at least 120 psig to operate. Higher values will

be needed depending on the test being run. Pressure can be decreased by cracking open

the air drain valve. Pressure is increased by use of the air compressor. To use the air

compressor, open the drain on the air compressor to drain any moisture. Close the drain.

Plug in the compressor and attach compressor to the air supply line via the high pressure

yellow air hose. Turn on the air compressor and monitor the tank pressure until the

desired pressure is achieved. Note, the water tank level gauge will read erroneously

when the air compressor is running. When the tank is at the right pressure, turn the air

compressor off and disconnect the high pressure air hose starting with the end attached

to the air line, then the end attached to the compressor.

To adjust the speed before and during the operation the RPM of the VFD drive

can be set by going to U1-01 in the handheld controller.

73
A-2.1 Running Secondary loop (Heat Exchanger)

Inlet valve (bottom valve) and outlet valve (top valve) before running the secondary loop

pump-Open

Safety switch behind the secondary-loop’s VFD drive (outside the test-cell) - Switch on.

Safety switch next to the secondary-loop’s pump and VFD drive –Switch on.

Frequency of the secondary-loop’s VFD drive- Set to 60Hz.

VFD drive- Switch to LOCAL then press the RUN bottom to turn on

Fans of the heat exchanger – Turn on

Secondary-loop’s water flow rate - Check (via LabView)

A-3 After operation checklist

Air inlet shut off valve – closed (while pump is running)

VFD – Press the STOP bottom

Variable frequency drive’s safety switch – power off

Seal flush pump – off

Air inlet line sight glass – check for water

Air drain – open to drain any water, then closed

Proximity probe power supply – off

Instrument air supply – off

Secondary loop’s pump- off

Secondary loop’s heat exchanger fans- off

Secondary loop’s Inlet and outlet valves-off

74
The loop may be kept pressurized in this condition if it is anticipated that it will be run

again shortly, i.e. within a week.

If the system will not be run for an extended time, or if it is necessary to work on any

pressure containing component, the system should be depressurized and drained.

A-3.1 To depressurize:

Instrument air supply - on

Outlet control valve – open (via Labview)

Pressure relief valve – open

Pressure – monitor until pressure is at atmospheric

A-3.2 To drain:

Separator tank outlet valve – open to drain entire water supply, closed to drain only the

pump and piping.

Water inlet control valve – open (via Labview)

Water drain – connected to hose leading to storm drain

Water drain – open

When system is drained:

Water drain – closed

Instrument air supply – off (control valves will close)

Separator tank outlet valve – closed

Pressure relief valve - closed

75
APPENDIX B

DRAWINGS

76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
`

155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176