ESP II Module Outline

Presentation Material
1) Applications Engineering
2) ESP Applications Theory
3) Special Applications
4) Monitoring and Analysis
5) ESP Diagnostic Exercise
6) Trend Analysis Exercises
7) ESP Troubleshooting

Reference Material

1) Artificial Lift White Paper

2) Borets Vortex Gas Separator
Borets Case Study in Gassy Environment
 2014

Applications Engineering Overview

 Application Theory

 Gather Good Data

 Know Your Well
 Good Design
 Proper Operations
 Thorough Monitoring
 Good Failure Analysis
 Review of Levels of Data

 Absolutely Necessary:
• Without this data we can not even get close.
 Better:
• This data will give us a fairly good design.
 Best:
• With this data we should be able to “nail” the design.
Absolutely Necessary “We Can Make it Work”

• Well Name • Oil Cut

• Casing Size & Weight • GOR / GLR
• Tubing Size & Weight • Expected Pump Intake
Pressure
• Reservoir Temperature
• Tubing Pressure
• Deviated Hole
• Expected Rate
• Pump Set Depth
• VSD or Switchboard
• Specific Gravities (Oil,
Water, Gas)
 Better Data “A Much Better Design”

• Total Depth • SBHP or Well Inflow

• Casing Tools & Depths • Pressure Datum
• Producing Intervals • Casing Pressure
• Deviation Survey • Best Inflow
Performance
• Current Well Test
• Corresponding PBHP
• Corresponding Tubing
PSI • Bubble Point
• Gas Impurities (CO2,
H2S, N2)
Best Data “We Can Zero-In on Our Design”

• Producing Zones • Best PVT Correlations

• Percent Gas Up Tubing • PVT Data
• Percent Fluid Up • Viscosity Values
Casing
• Wellhead Temperature
• Test Vessel Pressure
and Temperature
• Best Friction Loss
Correlations (Tubing
and Wellbore)
Important Things to Know About Your Well

 Solids Production  Fracture Close

Pressure
 Scaling Tendency
 Cyclic Nature
 Paraffin / Asphaltenes
 Corrosion Problems
 Water Coning
 Chemical Treatments
 Sizing Tools

 A good design is a result of gathering good well data,

understanding the well and equipment, and applying the
proper sizing tools.

 It’s important to understand the correct correlations.

 Utilization of Proper Software

• Each Vendor is now either using their own software or well into the
development of their own software
– SLB = Avocet (Design Pro)
– CL = Autograph
– WGESP = IHS SubPump but moving to their own
– BWFT = IHS SubPump but moving to their own
– Others = SubPump
• Oil companies will “muscle” their vendors to provide this software to
key individuals
• Training on the software is essential - - or the time to simply learn it.
 Proper Operation is Essential

 Clean Power  Minimum Cycling

 Good Grounding  Careful VSD Changes

 Wellbore  Tubing & Casing

Management Backpressure

 Chemical Program
A Long Runlife Means the ESP Must be Monitored and
“Optimized”

 Collect Data

 Evaluate Data

 Recommendations and Follow - Up

 Monitoring

 Amp Charts  Tubing & Casing

Pressures
 Electrical Readings
 Fluid Levels / PIP
 Speed Settings
 Pump Tests
 Treatment Histories
 Temperature Readings
 Production Data
 Failure Frequencies
 Optimization

 Gas Interference  Over / Under

Producing
 Trash in Pump
 TDH & Theoretical
 Worn Pump
Rates
 Formation Damage
 Treating Problems

 Power Problems
 Failure Analysis (DIFA)

 More on this tomorrow - - but it is invaluable if you want to

extend runlives
 Don’t depend on the vendor to tell you everything
 You need someone “experienced” to do this
 If you don’t have someone experienced - - get the
training.
 ESP Application Guideline

 A fundamentally sound ESP guideline is essential

 Each BU has unique situations and will need certain modifications to

the guideline
– For example what’s good in Canada may not work in the North
Sea.
 What’s Good for ESPs?

 Large Casing

 High Water Cut

 Low Gas Fraction

 Low BHT

 Good Power System

 No Scaling Tendencies

 No Sour Gas

 BUT - - we don’t always get what we like or want!!

 Small Wellbores

 338 Series Pumps

 Typically limited to 1200 bpd (Can extend with VSD)

 375 Series Motors

 Typically limited to  125 BHP (5 Tandem Units)

 375 Series Seals

 Typically limited to 2400 lbs thrust

• Check with the specific vendor for the exact limits.

 Poor Power Quality

 Effects Include
 Shortened equipment life
 Lost Production (in my view there is no such thing as
“deferred” production)

 Solutions
 Surge Protection
 Filtering (VSD harmonies)
 Surface Power Line Geometry
 Good Grounding Scheme
 Reduce Water Cut

 Effects Include
 Reduced Motor Cooling
 Possible Higher Viscosity
 Possible Increased Gas
 Probable Lower (but could be higher) TDH

 Solutions
 Shrouds
 De-rate Motor BHP
 De-rate Pump Rate and Head Capacities
 Scale

 Effects Include
 Plugged Pumps or Tubing

 Solutions
 Scale Inhibitor
 Coatings
 Change Operating Conditions
 High Gas Fractions (GVF)

 Effects Include
 Reduced Pump Performance
 Reduced Motor Cooling
 Reduced TDH

 Solutions
 Gas Separators
 De-rate Pump Rate Capacity
 Tapered Pumps
 Shrouds (including methods to draw below perforations)
 De-rate Motors BHP
 Sour Gas

 Effects Include
 Reduced Equipment Runtimes

 Solutions
 Corrosion Inhibitors
 Equipment Coatings
 Equipment Metallurgy
 Lead Cables
 High BHT

 Effects Include
 Shortened Motor, Seal, Cable &
Instrumentation runtimes

 Solutions
 HT Motor, Seal & Cable
 Shrouds
 De-rate Motor BHP
Thank you for your attention!

24
 2014

ESP Applications Theory

 Total Dynamic Head

 When we begin to design an ESP system for a well we

start by calculating how much LIFT the pump
must produce at the FLOWRATE that we wish to
achieve.

 The amount of LIFT the pump produces is usually

calculated in FEET. The LIFT can also be calculated in
Meters or even in the Pressure that must be
developed to lift fluid to the wellhead.

 The LIFT that we calculate is known as the Total

Dynamic Head that the pump must produce.
 Total Dynamic Head

TDH is a sum of three basic components:

1) The Net Vertical Lift or net distance which the fluid

must be lifted

2) Wellhead Pressure – converted to feet of head

3) Friction loss in the tubing – converted to feet of head

 TDH Components
Wellhead
Wellhead Pressure
3
Ground Level
Flow

1
2 Net Vertical Lift
Total Friction Loss

Producing Fluid Level

Pump Set Depth

Flow Flow
 Total Dynamic Head

 First we will look at the "net vertical lift".

 This is the vertical distance through which the fluid

must be lifted to get to the surface.

 The energy required to lift the fluid can be determined

by the equation:
Work (energy) = mgh

 Where m is the mass of the fluid, g is the acceleration

due to gravity and h is the height the fluid is lifted.
 Net Vertical Lift
Wellhead
Note that the vertical lift
only depends on where
the fluid level is. Flow
Ground Level

From the Net Lift stand

point, it makes no 1
difference where the Net Vertical Lift
pump is set.

Producing Fluid Level

Pump Set Depth

Flow Flow
 Net Vertical Lift
Wellhead

Ground Level

Note that even

though the pump is
much lower, the net lift Net Vertical Lift
does not change.

Producing Fluid Level

Pump Set Depth

Pump Set Depth

Flow Flow
 Net Vertical Lift

Why is the vertical lift independent of where the pump is

set?

Because if you were to turn the pump off, the fluid level
in the tubing would be the same as the level in the
annulus.

Remember the old rule that water seeks its own level?

A better way to look at it is to note that the fluid level

above the pump is providing pressure at the pump
intake. This is energy which the pump will not have to
supply since it is already there.
 Net Vertical Lift

What if, instead of lifting the fluid vertically, we

move it sideways? How much extra work would
we need to do?

None!

If Work (energy) = mgh,

The h is zero if we move sideways so the
work must be zero.
 Net Vertical Lift

What about deviated wells?

 Net Lift Net Lift

PFL

Regardless of
where the pump is
set or the angle,
the vertical lift will
not change.
 Total Dynamic Head

For the purposes of this example, we will

assume we are given a fluid level of 4000 feet
from surface (vertical distance).
Net Vertical Lift = 4000 ft 1

Remember if the well is deviated, the total

measured distance from surface could be much
greater but, since the work done in moving the
fluid sideways is zero, only the vertical distance
matters.
 Total Dynamic Head

We have therefore calculated the Net Vertical Lift

that the pump needs to generate at one
particular flowrate.

Is the Net Vertical Lift the same at all flowrates

that the pump is capable of producing?
 Total Dynamic Head

To answer this question we need to think about

what happens to the fluid level in the well as the
pump produces at higher and lower rates.

As the fluid level changes with flowrate

and the Net Vertical Lift is measured from
the Dynamic Fluid level to surface it
follows that the Net Vertical Lift must
change as the flowrate changes.

Let’s consider the term DRAWDOWN

 Drawdown

Q1 Q2 Q3
Ground Level

L1
Q1 Q1 L2
L3
 Net Vertical Lift (NVL) v Q Curve

Net Vertical Lift v Q (Flow rate)

NV Lift

Q
 Friction Loss

What is friction?
 Friction Loss

Friction is an energy loss (we actually measure

it as a pressure loss) due to viscous shear of
the flowing fluid.

In a fluid, molecules are free to move past each

other but there may be a little resistance. This
resistance is due to shear forces which must be
overcome.

The higher the shear forces, the higher the

viscosity.
 Friction Loss

In a single phase fluid, most of the liquid is

moving along together so there is not much shear
in the liquid itself and this friction can usually be
ignored.

No
Worries
Excuse me

Sorry

Certainly
 Friction Loss

The walls of the pipe, however, will tend to "stick"

to the fluid so shear forces between the pipe and
the fluid can be quite large and increase as the
velocity of the fluid increases.
I want out
of here!
Hey!

Velocity
Profile
(Laminar Flow)

Ouch!
 Friction Loss

The amount of friction present can be represented

by a "friction factor" - f . There are several
equations available to calculate f but we will not
review those here. If we know f we can calculate
the pressure loss from the following:

f v 2 Where P = pressure loss

= fluid density
P  v = fluid velocity
2 gc d gc = gravity constant
d = pipe diameter
 Friction Loss

What is the friction loss at zero flow?

 Friction Loss

At zero flow, v = 0 so the P due to friction is

zero.

f v 2
P 
2 gc d
 Friction Loss

Assume the flow rate is not zero but is some

constant value. What happens to the friction as
the pipe diameter increases?

f v 2
P 
2 gc d
 Friction Loss

As the pipe diameter increases, the P

decreases as can be seen in the equation. But
something else also happens. What is it?

f v 2
P 
2 gc d
 Friction Loss

As the pipe diameter increases, the velocity, v,

decreases by the square of the diameter change
so it is reduced drastically.

These two factors make an increase in pipe diameter

have a large impact on decreasing the frictional
pressure losses.
f v 2
P 
2 gc d
To put it simply we have less frictional
resistance to overcome in a larger diameter
tubing
 Friction Loss

So how do we calculate friction loss?

 Friction Loss

Fortunately, there are many charts available for

determining friction as we do not need to use
these equations. A very useful chart for our
purposes follows:
Friction Loss
 Friction Loss

This is how to use the chart:

Say, for example, we have a total tubing length of

6500 feet and we want to produce 5000 bpd. We
have both 2 3/8" tubing and 2 7/8" tubing in
stock. What will the friction be?
 Friction Loss

540

200

2-3/8” 2-7/8”

First, find the flow rate on

the X axis and move up
to the correct tubing

5000
 Friction Loss

Say for example we have 6500' of tubing:

For 2 3/8", Friction = 540*6.5 = 3510 feet of loss

For 2 7/8", Friction = 200*6.5 = 1300 feet of loss 2

You can see a small increase in tubing size makes

a big difference in friction loss. What about trying 3
1/2" tubing?
 Friction Loss

2-3/8” 2-7/8” 3-1/2”

73

5000
 Friction Loss

Since we have 6500' of tubing:

For 3 1/2", Friction = 73*6.5 = 475 feet of loss

If we can use 3 1/2" tubing, this will allow us to

use a smaller pump and motor which will reduce
cost
 Friction Loss

Is bigger tubing always better?

No

If a well is producing a lot of sand, the higher

fluid velocity in small tubing helps carry the sand
to the surface. If the tubing gets too large, the
lower velocity can allow sand to accumulate on
top of the pump discharge and eventually plug it
up.
 Friction Loss

However, in general, friction is something which

cannot be avoided but you should make every
attempt to minimize it.
 Friction Loss v Q Curve

Friction Loss v Q (Flow rate)

Friction

Q
 Wellhead Pressure

Wellhead pressure is sometimes called

"Surface Pressure", "Back Pressure" or even
"Flowline Pressure".

Actually the most accurate term is "Tubing

Discharge Pressure" since this is the pressure
at the discharge of the tubing from the well.
 Wellhead Pressure

"Back Pressure" is also a good term since it

implies the correct location in the discharge of the
tubing string.

"Flowline Pressure" can actually be a much lower

pressure if a surface choke is being used to cut
back the flow rate from the well.

"Surface Pressure" is just ambiguous.

All these terms are used interchangeably in the

industry. Just be aware that they are not all the
same
 Wellhead Pressure

Technically there are two "Wellhead

Pressures“

What are they?

 Casing Annulus Tubing Discharge
or or
Casing Head Pressure Tubing Head Pressure
 Wellhead Pressure

Up to this point, we have been calculating

everything in terms of "feet". This is very
convenient when sizing a pump.

WHY?
 Wellhead Pressure

Because pump curves are usually plotted in terms

of feet of head which makes it easy to calculate the
number of stages.
 Wellhead Pressure

Say for example we are told that we need a

wellhead pressure of 200 psi.

100 to 200 psi is an average number and usually

gives enough pressure to get the fluid through
the flowline into a separator or some other piece
of equipment. If the flowline is very long or has
to go up a high hill, a larger pressure could be
needed.
 Wellhead Pressure

Now that we know what wellhead pressure is, what

do we do with it?
 Wellhead Pressure

Up to this point, we have determined a vertical

lift in feet and calculated a friction loss in feet.

We cannot add "psi" to "feet".

The easiest way to do this is to convert the

wellhead pressure to feet but in order to do this,
we need to know the specific gravity.
 Wellhead Pressure

Most of the time the specific gravity of the well

fluid is known but this time we do not know what
it is.

Let's say, for this example, that we have been

told that the water cut is 60%. (This means that
the other 40% of the fluid is oil).
 Wellhead Pressure

This is the equation to convert from psi to feet but

we still need to know the specific gravity.

Wellhead Pressure*2.31
Wellhead "Feet" = ----------------------------------
[Link].
 Wellhead Pressure

The best thing to do is to try and make a good

guess.

How do we estimate a gravity ?

 Wellhead Pressure

We will need to guess a gravity for both the oil

and the water.

Water is pretty easy. Fresh water has a specific

gravity of about 1.00. However, well water is
usually not fresh - it may be brackish (salty) or
contain minerals or other dissolved solids.

This will make the gravity higher.

 Wellhead Pressure

Water gravities usually vary from about 1.03 (sea

water) to 1.10 although they can go as high as
1.20 in very heavy water.

Unless we have some knowledge of the water in

this particular case, let's assume an "average"
value of 1.05.
 Wellhead Pressure

Now consider the oil (remember the well is only

60% water). Oil gravities can vary a great deal from
one reservoir to another.

Additionally, oil gravities are usually referred to in

"API" gravity rather than specific gravity.

This is similar to measuring temperature in Celsius

and Fahrenheit. 37°C is the exact same as 98.6°F.
They are just two different ways of saying the same
thing.
 Wellhead Pressure

Petroleum Engineers prefer to use the API gravity

because it is a larger number and easier to "get a
feel for". The equations for converting from one
unit to the other are:

141.5
[Link]. =
131.5 + API

141.5
API =  131.5
[Link].
 Wellhead Pressure

What is the specific gravity of an oil with API=28?

What is the API of fresh water ([Link].=1.0)?

141.5
[Link]. =
131.5 + API

141.5
API =  131.5
[Link].
 Wellhead Pressure

What is the specific gravity of an oil with API=28?

141.5
[Link]. = = 0.887
131.5 + 28
 Wellhead Pressure

What is the API of fresh water ([Link].=1.0)?

141.5
API =  131.5 = 10.0
1.0
 Wellhead Pressure

Now we know how to convert, but we still do not

know what value to use.

Oil API gravities can vary all the way from 6 or

7 on the heavy end to as much as 60 or 70 on
the light end.

Remember that high API gravity means lower

specific gravity so - Higher is Lighter
 Wellhead Pressure

This large spread of numbers is why engineers

prefer the API units.

It is much easier to "conceptualize" the difference

between a 20 API oil and a 40 API oil than to
imagine the difference between a specific gravity
of 0.934 and 0.825 although they mean the same
thing.
 Wellhead Pressure

An API of 6 is not very common and usually

makes you think of paving roads (A very sticky
dense crude oil).

An API of 60 is also not very common and makes

you think of pulling your car up to the well and
filling the gas tank directly.
 Wellhead Pressure

If we need to make a "guess" as we do in this case,

an API in the range of 25 to 35 is a good safe guess.

For this example, let's use an API of 30.0 - but we

now need to convert it to specific gravity.
 Wellhead Pressure

This means that we are assuming the oil specific

gravity is 0.876.

141.5
[Link]. = = 0.876
131.5 + 30
 Wellhead Pressure

Now we have determined that the water gravity is

1.07 and the oil gravity is 0.876.

Remember that the well fluid is only 60% water and

40% oil.

What we need to do is get a "composite" gravity for

the combined fluid. The best way to do this is
simply to take an "arithmetic average".
 Wellhead Pressure

This sounds worse than it is. To get the arithmetic

average, we use the following formula:

Sp. Gr. = ( fw   w ) ( fo   o )
Where :
f w is the water fraction
 w is the water specific gravity
f o is the oil fraction
 o is the oil specific gravity
 Wellhead Pressure

For our case, the specific gravity is 0.992

Sp. Gr. = (fw x w) + (fo x o)

Sp. Gr. = (0.60 x 1.07) + (0.40 x 0.876) = 0.992

 Wellhead Pressure

Now that we have calculated a specific gravity,

let's convert the wellhead pressure from [psi] to
feet.

Wellhead Pressure*2.31
Wellhead "Feet" = ----------------------------------
[Link].
 Wellhead Pressure

Using our actual numbers:

200*2.31
Wellhead "Feet" = ------------- = 465 feet
0.992
3
 Well Head Pressure v Q (Flow rate)

WHP

Q
 Total Dynamic Head

The TDH will be the sum of lift, friction and wellhead

pressure.

We will assume 2 7/8" tubing since it was in

inventory:
 Total Dynamic Head
Wellhead 3
Wellhead Pressure = 465 feet
Ground Level Flow

1
2 Net Vertical Lift
Total Friction Loss =
1300'

Producing Fluid Level

4000
Pump Set Depth +1300
+ 465
5,765 ft of TDH

Flow Flow
 Total Dynamic Head

So we would need to design a pump with enough

stages to produce 5765 feet of head.

By the way, what would happen if our guess on

the fluid specific gravity was really bad (i.e. what
if the gravity were really only 0.82 instead of
0.992?).
 Total Dynamic Head

If this were the case, the wellhead "feet" would have

been 563 feet instead of 465 which means we were
92 feet short in our calculations.

We would need a pump to deliver 5857 feet of TDH

rather than one for 5765 feet.
 Total Dynamic Head

This looks like we would have "under sized" the

pump - and we did - but the difference between
5857 and 5765 is only 1.6%.

A 1.6% error in sizing a pump is not worth worrying

about since the well data you are working with in
the first place is usually lucky to be within 15% to
20% accuracy.
 Total Dynamic Head

A pump with a very high fluid level and a very high

wellhead pressure would make the wellhead
requirement much more significant so some
"engineering judgment" may be required in this
case.

Of course we could always request fluid samples of

the produced oil and water and have the actual
Specific Gravities calculate in the Laboratory – this
however takes time and if we do not have time on
our side we need to make some good engineering
estimates.
 TDH v Q Curve

To determine this we need to review the three

curves we have developed so far:

 1. Net Lift
 2. Friction and
 3. Wellhead Pressure

The TDH curve will be the mathematical sum of

these three curves
Net Vertical Lift v Q (Flow rate)

NV Lift

Q
Well Head Pressure Curve v Q (Flow rate)

WHP

Q
Friction Loss v Q (Flow rate)

Friction

Q
TDH (System) v Q (Flow rate)

TDH

Q
TDH Curve (System) v Q (Flow rate)

The TDH curve is also sometimes known as

the

“System Resistance Curve”

 TDH Curve (System) v Q (Flow rate)

TDH

Q
 Pump Curve (Head) v Q (Flow rate)

TDH
Pump curve

Q
 TDH (System Curve) v Pump Curve

TDH

TDH
Pump curve

Q
 TDH (System Curve) v Pump Curve

TDH

H operating

TDH
Pump curve

Q Q operating
 ESP System Sizing

INFORMATION REQUIRED AND WHY?

 ESP System Sizing

Other factors to consider

Now that we know how to understand TDH

(System) Curves and how they relate to
Estimated Pump Performance curves we shall
review some of the parameters relating to the
Oil well that need to be considered when
sizing ESP’s
.
 ESP System Sizing

Casing Diameter and Production Tubing

Available

 To calculate the friction loss from pump to

surface

 To ensure that there is enough room to install

the ESP cable
 ESP System Sizing

Is The Well Deviated or Horizontal

 To ensure that the equipment can be installed

through any deviated sections (doglegs)

 To ensure the correct selection of Protector

configuration
 ESP System Sizing

Downhole Restrictions

 Y –Tool
 Casing patch
 Sand control
 Wireline / junk
 ESP System Sizing

Fluid Properties

 Crude API, Water Cut, S.G. of Water, GOR

 Is there a corrosion inhibitor injection
 Viscosity
 Emulsions
 Abrasives
 ESP System Sizing

General Well Data

 Perforated interval and depth

 SBHP and Productivity Index
 FBHP and Q (at two different flowrates)
 Bottom Hole Temperature
 Well head pressure required
 ESP System Sizing

How is the Reservoir Expected to Change in the

Future

 Pressure decline (Pressure maintenance)

 Water cut increase
 Gas break through
 Effect of other artificial lift schemes in the
area.
 ESP System Sizing

Surface Considerations

 Power supply available – Ministry / Utility /

Genset (Voltage, Frequency, Quality)

 Surface piping and footprint

 Wellhead types available

 ESP System Sizing

Bigger Diameter Is Almost Always Better

 Usually More Efficient

 Usually Less
Expensive

 Better in Gas and

Viscosity

 Handles Higher HP
 ESP Motor Application

Motor Application Guidelines

 ESP Motor Application

With all the motor choices available, which

motor should we use for a given application?
 ESP Motor Application

Motor Series

 In general, just as in pumps, the bigger the

better.

 Larger diameter motors are less expensive to

purchase.

 They can also go to higher horsepower

without the need for tandem connections
which will enhance reliability — simpler
systems are better.
 ESP Motor Application

 After choosing the Motor Series and Type, we

will know if it will be necessary for a single or
tandem motor to match the HP requirement of
the pump.

 Now we should look at Volts and Amps

options available to us for a given HP size.
 ESP Motor Application

For a given horsepower motor, the product of the

volts and amps will be essentially constant.
For example for a particular Hp we could have

A 1000V, 50 amp motor

A 2000V, 25 amp motor

A 500V, 100 amp motor.

In other words KVA is constant.

 ESP Motor Application

So why have more than one voltage

available?
 ESP Motor Application

ESP Providers have more than one motor

voltage available to try and accommodate their
customers requests for voltages that are
compatible with the power distribution voltages
in their oilfields.
 ESP Motor Application

We also supply different voltage motors

because for any design we must consider the
load on the power cable.

Higher voltage motors allow the use of

smaller cable sizes and can give us less
voltage loss in the power cable

Lower voltage motors means higher current

and this results in higher voltage lost in the
power cable.
 ESP Motor Application

A point to note:

High voltage motors (single motors) are no

more or less efficient than low voltage
motors but there are advantages to using
them.

Even though the motor efficiency does not

change, the overall system efficiency will be
less if lower voltage, higher amperage
motors are used because of the increased
losses experienced in the power cable.
 ESP Motor Application

Now we understand why various voltages are

available, but why are such odd voltages
available?

Surface motors, for example, are rated at

460V, 4160V, 2300V, etc. These surface
motors are made to "standard voltages".

So why do the ESP motor voltages turn out to

be such strange numbers?
 ESP Motor Application

The problem again is the voltage drop in the

cable. ESP motors have to contend with a
very long length of power cable which surface
motors do not.

If we have a 460V surface supply, we would

probably only want to have about a downhole
motor designed with about 430 volts required
at the motor terminals.
 ESP Motor Application

We are also in effect limited by “standard”

surface equipment.

Motor control panels come in certain voltage

ranges such as 600V, 1000V, 1500V, 2400V.

Motor voltages are designed, assuming a

length of cable, such that the total voltage
(motor plus cable loss) will fall just below
one of the panel ratings.
 ESP Motor Application

Remember that we have decided the higher

the motor voltage is the better.

But we must also check that the motor

voltage we select is not so high that after we
add on the cable voltage loss we exceed the
control panel rating.
 ESP Motor Application

In reality, there are dozens of potential

winding configurations for any particular
motor Hp but the number is intentionally
limited to three or four to try to cover the
majority of applications with the minimum
amount of inventory.

Occasionally an application requires a special

winding. An ESP Provider’s Engineering
group should be consulted to see if it is
possible - not every voltage can be hit exactly.
 ESP Motor Application

Tandem motors

When putting more than one motor together in

tandem combinations, always keep the sections
the same Hp and voltage. For example a 300 Hp
540 motor should be made of two 150 Hp motors
rather than something like a 200 Hp and a 100 Hp.

Note: It is possible to mismatch motors if the wire

size is the same in each winding but this should
be avoided and never done without consulting
an ESP Provider’s Engineering group.
 ESP Motor Application

So what happens when we put two

motors together?
 ESP Motor Application

With two motors we double the Hp (add the

two Hp's together). We also double the
voltage. The amperage remains the same.

With three motors we triple the Hp and

voltage but the amperage still does not
change.

For example, a 140 Hp, 1299V, 69.5A UT motor

coupled to a 140 Hp, 1299V, 69.5 CT motor
would give us a 280 Hp, 2598V, 69.5A motor.
 ESP Motor Application

Always take care when adding motors

together so that the total voltage does not
exceed the system limits - do not try to put
3500 volts on a 3 kV cable.

Surface controllers, transformers, wellhead

feedthrough mandrels will all have voltage
limits we need to be concerned with.
 ESP Motor Application

A comment on the use of UT motors

A UT motor should not normally be used as a

single section unless a pressure sensing
instrument is being run since tandem motors
are not as efficient as single motors.

Why not?
 ESP Motor Application
The reason for this is that the upper tandem
motor has no Y-splice and drops one-half turn of
wire per phase which the single motor does not
have to do.

This loss of copper lowers the efficiency. Higher

voltage tandems will not be penalized as much as
lower voltage units because the one wire being
left out is much smaller minimizing the effect.

A tandem motor can easily have a 1% loss in

efficiency over a single. Remember that motor
performance curves are based on singles - not
tandems
 ESP Motor Application

Always double check the actual operating

temperature of the motor you have selected. We
can compare the actual temperatures with the
recommended linits and check to see if any
operating parameters have been exceeded.
Once the motor has been selected, using the
software like Catapult / SubPUMP we can get the
actual motor performance, velocity and mainly
the motor Operating temperature.
 2014

Special Applications
Special ESP Applications

• Abrasives
• Corrosive Fluids
• Scale Formation
• Gas
• Viscosity
 Abrasive Fluids Production

Mechanical Failure Modes

The mechanical failures of an ESP can

generally be classified as:
•Radial Instability
•Downthrust
•Erosion/Corrosion
•Vibration

Abrasives will accelerate these conditions

 Abrasive Fluids Production

In applications of ESP systems producing from fractured

completions and unconsolidated formations, the
production fluid can carry with it unconsolidated sand
from the production zones.

The effects of producing abrasive fluids on ESP

components and surface installations will be noticed as a
degradation of the ESP’s performance due to stage wear.

Sand is hard granulated material (up to 0.070 in.

diameter).
Typical Abrasive Composition

COMPONENT PERCENTAGE

SiO2 59.07
Al2O3 15.22
Fe2O3 3.1
FeO 3.71
CaO 5.1
MgO 3.45
Na2O 3.71
 Stage Metallurgy

Standard Alloy: Type 1 Ni-Resist

– Austinetic Cast Iron
•Cast Iron Structure
•Corrosion Resistance Similar to 302/304 SS
– Composition
•Nickel 15%
•Copper 6%
•Chromium 2%
•Silicon 2%
– Hardness
•120-160 Brinell
 Radial Wear Effects

• Tight clearances begin to open up

• A loss of hydraulic efficiency
• Eccentric rotation causes higher
bearing loads
• Excessive vibration leads to other
problems
Radial Flow Stage Problem Areas

Impeller Diffuser

Flow Path

Erosion Erosion/Wear Wear

 Mixed Flow Stage Problem Areas

Impeller Diffuser

Flow Path

Erosion Wear/Erosion Wear

 Journal Bearing Standard Pump Configuration

Shaft/Sleeve Impeller Shaft/Sleeve

Journal Hub/Diffuser Journal
Bearing Bore Bearing
Journal
Bearing
 Abrasive Fluids Production

Once it is determined that the produced fluid

contains sand or abrasives then consideration
should be given to the following four alternatives:

a) Attack the problem in the bottom of the well,

control the sand.
 Abrasive Fluids Production

Once it is determined that the produced fluid contains

sand or abrasives then consideration should be given
to the following four alternatives:

a) Attack the problem in the bottom of the well, control

the sand.

b) Produce well fluids under present conditions

 Abrasive Fluids Production

Once it is determined that the produced fluid contains

sand or abrasives then consideration should be given
to the following four alternatives:

a) Attack the problem in the bottom of the well, control

the sand.

b) Produce well fluids under present conditions

c) Change the lift method

 Abrasive Fluids Production

Once it is determined that the produced fluid contains sand or

abrasives then consideration should be given to the following
four alternatives:

a) Attack the problem in the bottom of the well, control the

sand – gravel pack.

b) Produce well fluids under present conditions – using

operating procedures, harder metallurgy stages and non
metallic such as Teflon (PTFE) and Impreglon coatings.

c) Change the lift method.

d) Re-Complete or Abandon the well.

 Abrasive Fluids Production

Depending on the well conditions, Borets has a number of alloy

Ni-Resist stages:

• Spherical Graphite Ni-Resist

• Significant increase in wear and corrosion resistance
• Increased hardness and durability due to the graphite
microstructure

• High Manganese Ni-Resist

• Cost competitive to similar wear and corrosion materials
• Preservation of austenite microstructure for increased wear and
corrosion resistance
• Co-Resist
• Excellent wear resistance due to the nitride layer formed after
hard surfacing
• Powder Metal - Powder form sintering powder metal process
• 20% increased wear resistance
• 10 – 15% reduction in size due to higher lift per stage
• High dynamic balance results in reduced vibration and wear
 Abrasive Fluids Production

If the fluid being pumped did not have sand or

formation solids (glauconite) made of very hard
particles, and instead had much softer particles than
the stage material.

Would abrasion effects be present?

 Abrasive Fluids Production

The effects of abrasion on pump stages are present

because fluids with sand containing quartz particles
or materials with a hardness greater than Ni-Resist
(metallurgy of the stages) are being pumped.

The most severe wear occurs with small diameter

formation solids as the wear occurs between the
shaft and impellor / diffuser
 Abrasive Fluids Production

It is not necessary for the particles to be

harder than the stage material to have
abrasion wear.

A large amount of a very soft material like

calcium carbonate or barium sulphate can
produce the same or even worse damage to
the pump than a small amount of very hard
abrasives like quartz.
 Abrasive Fluids Production

The presence of any particle can produce abrasion.

The hardness of the abrasive material determines

the time it will take for abrasion’s effects to be
present.
 Corrosive Fluids Production

Corrosion effects not only the submersible

components but also the production tubing and
casing as well.

It reduces the re-usability of the tubing, cable,

surface flow lines and generally all ESP
equipment.

By definition of corrosion we are referring to the

alteration or destruction of one or more metallic
components due to chemical reaction with the
liquids and gas in the well.
 Corrosive Fluids Production

Standard Carbon steel pump, motor, protector

housings are susceptible to CO2 and H2S
corrosion especially in areas of highest
temperature - motors, protector thrust bearing
 Corrosive Fluids Production

Several methods are used to treat and/or

prevent corrosion:

• Chemical inhibition treatment

• Installation of sacrificial anodes (magnesium)
below the motor,
• Coatings with resistant materials (metallic
and non metallic such as Teflon (PTFE) and
Impreglon.
• Special corrosion resistant materials for
equipment housings.
 Corrosive Fluids Production

The use of manufactured alloy and coated

equipment will give good results.

• These alloys may consist of 9Cr1Mo (9%

Chromium and 1% Molybdenum
• 316 Stainless steel housings to improve general
corrosion resistance.
• Monel (Monel is a group of nickel alloys,
primarily composed of nickel (up to 67%) and
copper, with small amounts of iron, manganese,
carbon, and silicon) coated housings
 Corrosive Fluids Production

If H2s (Hydrogen Sulphide) and/or CO2 are

present in the well the use of alloy equipment
should be considered after applying the
following guide:
 Corrosive Fluids Production

If H2s (Hydrogen Sulphide) and/or CO2 are

present in the well the use of alloy equipment
should be considered after applying the
following guide:

a) H2S partial pressure is 10 PSIA or more.

 Corrosive Fluids Production

If H2s (Hydrogen Sulphide) and/or CO2 are

present in the well the use of alloy equipment
should only be considered after applying the
following guide:

a) H2S partial pressure is 10 PSIA or more.

b) CO2 partial pressure is 7 PSIA or more.
 Corrosive Fluids Production

If H2s (Hydrogen Sulphide) and/or CO2 are

present in the well the use of alloy equipment
should only be considered after applying the
following guide:

a) H2S partial pressure is 10 PSIA or more.

b) CO2 partial pressure is 7 PSIA or more.
c) H2S constitutes 15% or more of the gas
volume at any pressure.
 Corrosive Fluids Production

If H2s (Hydrogen Sulphide) and/or CO2 are

present in the well the use of alloy equipment
should only be considered after applying the
following guide:

• H2S partial pressure is 10 PSIA or more.

• CO2 partial pressure is 7 PSIA or more.
• H2S constitutes 15% or more of the gas
volume at any pressure.
• The produced fluid pH is 6 or less.
 Corrosive Fluids Production

This partial pressure in a) and b) is calculated

as follows:

%Gas x FBHP
Pp =
100
 Corrosive Fluids Production

Although not commonly used, cathodic protection if

properly designed can prevent corrosion of the ESP
equipment.
 Scale Applications

Scale is a deposit formed by chemical reaction

or temperature and pressure changes on
surfaces in contact with water.

Calcium carbonate, magnesium carbonate,

calcium sulfate and barium sulphate are the
most common.
 Scale Applications

When scale is formed, the deposits are found in

the tubing’s internal surface, well head, surface
lines, manifolds, well perforations; and it also
seriously affects the performance of three basic
ESP components:

INTAKE
 Scale Applications

When scale is formed, the deposits are found in

the tubing’s internal surface, well head, surface
lines, manifolds, well perforations; and it also
seriously affects the performance of three basic
ESP components:

INTAKE

PUMP
 Scale Applications

When scale is formed, the deposits are found in

the tubing’s internal surface, well head, surface
lines, manifolds, well perforations; and it also
seriously affects the performance of three basic
ESP components:

INTAKE

PUMP

MOTOR
 Scale Applications

The most common preventive method for

producing fluids with a tendency to form
scale is continuous injection of anti-scale
chemicals by drip into the annulus or by
using a capillary tube and a mandrel
injection valve.

It is important to note that scale inhibition

chemicals have a low pH. This must be
stabilized as annulus drip applications
can be very corrosive to tubing and
casing.
 Scale Applications

For the anti-scale treatment to be effective on all

ESP components care must be taken to ensure
that the chemical injection point is located
below the motor.

This can be accomplished by a power cable

incorporating a capillary tube.
 Scale Applications

Capillary Tube for

corrosion inhibitor injection
 High Gas Applications

Bubble Point Pressure (Pb) and Gas Oil Ratio (GOR)

are the key factors in determining amount of free gas
produced. The free gas, not necessarily the gas
fraction, is what affects ESP pump operation and the
pump setting depth.

Producing below the Bubble Point Pressure will

result in gas that is in solution being released
into the fluid being produced.

How much free gas is released depends on the GOR

and how much the pressure drops below the Bubble
Point
 High Gas Applications

The simplest solution, even though not always

possible, is to design the pump setting depth in
such a way that the pump intake pressure (PIP)
will be at least 50 psi above Bubble Point
Pressure.

This keeps the gas in solution and it is taken

through the pump in a liquid phase

With the PIP set in this way, the pump will be

sized to handle a flow equal to the product of the
surface volume desired times the fluid’s formation
volume factor.
 High Gas Applications

The basic problem in the production of high gas

oil ratio wells is the interference of the free gas
in the first stages of the pump, causing the
stages to “gas lock”. (If properly sized, solution
gas does not present a problem for the pump.)

Gas locking takes place when the gas

completely fills the impeller and diffuser fluid
passages, blocking the fluid entrance.
 High Gas Applications

If a packer is not used - and depending on the

gas amount - a static or rotary gas separator can
be used.

Now, with the introduction of fluid conditioning or

gas handling devices, it is possible to handle gas
even in wells equipped with a packer, or in
installations without a packer and with extremely
high GORs.
 High Gas Applications

The Vapro, a multiphase pump, is a modified

multiple stage centrifugal device. Its operation
disperses the gas liquid mix and compresses the
associated gas.

This mixture can be handled by the pump with

improved performance and without gas locking
problems.
 High Gas Applications

The Vapro application should generally be

considered in wells with:

• 20 - 30% or more free gas (by volume) at the

pump intake or

• With vapor-liquid ratios (VLR) of 0.25 bbl/bbl or

greater at the pump intake.
High Gas Applications
 High Gas Applications

The Vapro can also be installed in tandem with

rotary gas separators in wells with very high gas-oil
ratios.

The Vapro is available in Ni-Resist.

As the gas passes through the pump there is an

additional bonus of a Gas Lift effect realised in the
discharge tubing as gas comes out of solution.
 PRODUCTION OF VISCOUS FLUIDS

As the liquids Viscosity increases the centrifugal

pump is faced with a more difficult task to perform.
Visualize viscosity and pumping issues as
mayonaise or peanut butter which are primarily
emulsion.

 Significant reduction will be noticed in the Capacity

(Q) and Head (H) produced
 Increased horsepower requirements
 Dramatic reduction will be noticed in the pump
efficiency
 PRODUCTION OF VISCOUS FLUIDS

For a typical Pump operating at BEP

(Best Efficiency Point) if viscosity increases
from zero to 15000 SSU (3300 cSt)

 Capacity can reduce by as much as 30%

 Total Head can reduce by as much as 25%
 Efficiency can reduce by as much as 70%
 BHP can increase by almost double
PRODUCTION OF VISCOUS FLUIDS

Viscosity Correction Factors for Centrifugal

Pumps are published by “Hydraulics Institute
Standards” however these are typically for single
stage surface pumps.

Submersible Pump companies have developed

their own Correction Factors and these are built
into their Computer Sizing Software such as
SubPump
 PRODUCTION OF VISCOUS FLUIDS

Techniques that have been considered to produce highly

viscous wells with ESP’s include:

 Most viscous applications are designed by over sizing

motors and increasing the number of stages

 Injecting water back into the well to lighten the

fluid. This is inefficient because injected fluid is
constantly re-circulated (Power wasted) and
must be separated (Surface facility required).

 Injecting water via a special nozzle at the intake.

This “coats” the viscous fluid and the pump
performs to catalogue / test curve.
Thank you for your attention!

50
 2014

Monitoring and Analysis

 The Traditional Way

• Fluid Shots to determine fluid level and calculate bottom hole

pressures

• ESP under-load and over-load settings

• Shut down production for Static logs

• Basic down hole pressure readings

What Does a Petroleum Engineer Require?

 Amps?

Sensitive Insensitive
Broken shaft Watercut
Stuck Pump Flowrate
Interest for
Gas Pump performance
Petroleum
Shut in (zero flow) THP
Engineers
Voltage change Reservoir Pressure
& Productivity
Index
 Amps Insensitive

 Amps constant and THP dropping…what’s happening?

4 JMD
 Pressure Shows More

 A tubing leak (recirculation)

 Two Pressures Even Better

 Increasing watercut

50% 90%

Pi const Pd
Pwf
 What Pressures Tell Us

Controlled by Used to
validate

Pd THP, Fluid density, Friction Watercut

(Flowrate dependent)

Pi Pd, Pump performance, Flowrate,

Reservoir pressure, Well Pr and or PI
productivity, Fluid density

Pd - Pump flowrate vs head curve Pump Wear

Pi Flowrate

Recirculation
 Modelling Can Tell Us

 Provide preliminary indication that says:

0 2,500
Head (50 Hz)
Measured data
Head (60 Hz)
200 Res. pressure Head (70 Hz)
Head w/corr. factors
2,000 Head at op. freq.
400
Range (min/max)
Op. Point
True Vertical Depth (m TVD)

600

Total head (m)

1,500
800

1,000
1,000

1,200

1,400 500

1,600
“ weP have a problem!”
RES

0
1,800
0 200 400 600 800 1,000 1,200
0 50 100 150
Pressure (bara) 3
Average pump flowrate (rm /day)
 Requires Analysis Skills

 Pump not providing correct head

(lost stages or wear)? 0 2,500
Head (50 Hz)
Measured data

 Water cut wrong?

Head (60 Hz)
200 Res. pressure Head (70 Hz)
Head w/corr. factors
2,000 Head at op. freq.
400

 THP wrong?
Range (min/max)
Op. Point

True Vertical Depth (m TVD)

600

Total head (m)

1,500

 Flowrate measurement wrong?

800

1,000
1,000

 High viscosity? 1,200

 …..which is it?
1,400 500

1,600 PRES

0
1,800
0 200 400 600 800 1,000 1,200
0 50 100 150
Pressure (bara) 3
Average pump flowrate (rm /day)
Production Loss Example
 Production Loss Example

9,000
 Production down from 3600 bbls/day to zero Head (50 Hz)
Head (60 Hz)
8,000 Head (70 Hz)
(from pump dP). Amps no change. Head (w/ CFs)
7,000 Head at op. freq.
 ESP protected by trips (Tm) Range (min/max)
Op. Point
6,000
 Well analysed – inflow problem identified

Total head (feet)

5,000

 Acid job performed

4,000

 Production after acid job 7800 bbls/day 3,000

 ESP saved 2,000

1,000

0
0 5,000 10,000 15,000 20,000
Average pump flowrate (rb/day)
 Consequences
 Pay attention to pressure and rate ……
 Typical Downhole Gauge System

Transformer
MDT Power Cable

Choke

120 Vdc
ISP
 Typical Gauge Specifications

Specifications
Measurement Accuracy Resolution
Pressure 0.1 % 0.1 Bar/psi
Temperature 1% 0.1 C
Vibration 1% 0.001 g
Current Leakage 0.05 % 0.001 Ma
 Downhole Gauge Parameters

 Intake Pressure & Discharge Pressure

 Motor Winding Temperature
 Intake Temperature
 Vibration
 Current Leakage
 Flow (or Discharge Temperature)
 Typical Monitoring Data

Time Pi Pd Ti Tm Vib Cl

9:56:34 3029 4414 125.6 133.8 0.23 0.1

9:56:55 3027 4426 125.8 134 0.253 0.102

9:57:16 3026 4410 125.8 134.1 0.249 0.105

9:57:36 3031 4418 125.5 133.8 0.24 0.101

9:57:57 3024 4408 125.8 134 0.231 0.105

9:58:17 3023 4404 125.5 133.8 0.246 0.102

9:58:38 3033 4420 125.8 134.1 0.251 0.099

9:58:58 3029 4410 125.6 134 0.262 0.096

 Gauge Pressure Measurements

Intake Pressure

 Measured using a 400 Bar sealed strain gauge pressure

transducer. The voltage output from the transducer is
near linear with pressure.
 High pressure option with a 690 Bar transducer
 Gauge Pressure Measurements

Discharge Pressure

 Measured in the same way as intake pressure but with a

1/4 inch sealed control line extending to above the pump
discharge
 Gauge Temperature Measurements

Intake Temperature

 This is measured by utilizing the semi-conductor

temperature sensor which is both linear and accurate in
temperature measurement. The voltage output varies with
temperature. The temperature sensor is calibrated in an
oven.
 Gauge Temperature Measurements

Motor Temperature

 Either motor winding with a thermocouple fitted in the

motor winding, or motor oil with the thermocouple fitted
into the oil at base of the motor.
 Measured using thermocouples and works by means of a
temperature difference system.
 For absolute temperature the Intake temperature is used
as the reference. The temperature sensor is calibrated in
an oven.
 Gauge Current Leakage Measurements

VSD
ISP
Current Leakage
Choke
 The MultiSensor system measures Transformer

leakage to earth through the ESP

system and also when the
Cable
MultiSensor is isolated from the rest
of the system. This is referred to as
active and passive current leakage
Motor

MDT
 Gauge Vibration Measurements

Vibration

 Measured by means of a calibrated accelerometer fitted

inside the MultiSensor
 Vibration Spectrum from ESP Motor

Average Pump Signature Y Axis

3.500
3.000
2.500
A m plitude

2.000
1.500
1.000
0.500
0.000
.6
.1
.6
.1
.6
.1

10 6
1

11 1
12 6
14 1
15 6
16 1
17 6
19 1
6
.
3.

3.
5.
8.
0.
3.
5.
8.
0.
15
28
40
53
65
78
90

Frequency (Hz)

average pump signature

23 JRC
 Rapid Sample Rate

 Rapid Sample Rate RSR allows for updates at 1 second

intervals of Pi only

Cf Pi RSR
Pd Q
Cz Pi Tm Vi
Ti

+120 Vdc

0 Vdc
 What Can We Do With This Data?

 Monitoring
 Trending
 Protection
 Validation and Analysis
 Use the analysis results for optimisation
 PRESSURE (PSI)

0
500
1000
1500
2000
2500
3000
03/11/1996 15:50
03/11/1996 16:16
03/11/1996 16:42

40 Hz
40 Hz
03/11/1996 17:09

Start up 1
03/11/1996 17:35 Start up 2
03/11/1996 18:01 40 Hz
03/11/1996 18:27
03/11/1996 19:23
Start up 4

40 Hz
03/11/1996 20:13

Intake Pressure
Start up 3
03/11/1996 20:40
03/11/1996 21:06
46 Hz

03/11/1996 21:32
48 Hz

03/11/1996 21:58
03/11/1996 22:24
03/11/1996 22:51
Choke 18%

03/11/1996 23:17
03/11/1996 23:43
50 Hz 14%choke

04/11/1996 00:09
04/11/1996 00:36
50 hz choke 40%

Shutdown
Choke 20%

04/11/1996 01:02

Discharge Pressure
04/11/1996 01:28
04/11/1996 01:54
04/11/1996 02:21
40 Hz

04/11/1996 02:47
Start up 5

04/11/1996 03:13
Choke gradually
increased to 40%

04/11/1996 03:39
04/11/1996 04:05
04/11/1996 04:32
04/11/1996 04:58
Motor Temp
Data Trending Events

04/11/1996 05:24
Choke 20%
Frequency increased
to 50 HZ,choke to 25%

04/11/1996 05:50
04/11/1996 06:17
reduced to 15%
Choke gradually

04/11/1996 06:43
04/11/1996 07:09
choke 27%

04/11/1996 07:35
Intake Temp
160
170
180
190
200
210
220
230

TEMPERATURE (F)
 Proactive Analysis Example
Proactive Analysis Critical Delta P
1700
required to 40
produce flow for
1650
motor cooling 35
1600

30

1550
Pump Delta Pressure (psia)

25
1500
Estimated

Vibration (g)
1450
Point of 20
failure
1400
15

1350

10

1300

5
1250

1200 0
0.25 125.25 250.25 375.25 500.25 625.25 750.25 875.25 1000.25 1125.25 1250.25 1375.25
Time

deltaP(psi) Vibration (g)

 Pressure Protection

 Intake Pressure
– Low Trip – protect against Pump Off
• Low Fluid Level
• Gas Breakout
 Discharge Pressure
– High Trip – protect against Shut In
• Closed Valves
• Heavy Fluid Slugs
 Delta Pressure
– Low / High Trip – Upthrust / Downthrust
• High Flow Conditions
• Low Flow Conditions
 Temperature Protection

 Intake Temperature
– High Trip – protect excessive Intake Temp
• Recirculation

 Motor Winding Temperature

– High Trip – protect excessive Motor Temp
• Low Flow
• High Load
• Anything that will cause the motor to heat
 Vibration and Current Leakage

 Vibration
– High Alarm – Submersible Pump Mechanical Damage
• High Solids Production
• Warning of Mechanical Wear
• Resonance ‘Frequency Fine Tuning’

 Current Leakage
– High Alarm – Electrical System Deterioration
• High Pump Heat (MLE)
• Resistance Breakdown
• Phase to Ground Warning
Getting it Right
Motor Cooling & Pump Off Protection
Getting it Wrong
$1.5m lost… and avoidable!
 Initial data modelled & Match gradient above pump Match across the pump
compared with measured using fluid properties using flow and inflow calcs
data – discrepancies found

Resulting data validates fluid

properties and running conditions.

This allows accurate analysis for

Optimisation.
 ESP Optimization

 Process of validation with measured downhole data.

 Increased production through advanced diagnosis.
• More efficient operating point, better use of power resources.
• Better understanding of well performance for subsequent pump design
0 3,000
From inf low Head (50 Hz )
Meas ured data Head (60 Hz )
500 Res . pres s ure Head (70 Hz )
2,500 Head (w / CFs )
Head at op. f req.
Range (min/max )
1,000
Op. Point
True Vertical Depth (ft TVD)

2,000

Total head (feet)

1,500

1,500
2,000

1,000
2,500

500
3,000
PRE S

3,500 0
0 200 400 600 800 1,000 1,200 0 10,000 20,000 30,000 40,000
Pr e s s ur e (ps ig) Ave r age pum p flow r ate (r b/day)
 Improved Pump Sizing

• 7500 bpd pump

Before Upsize • Over range (up thrust)
• FBHP 908psi

1800 stb/d Net Oil

0 12,000
From inflow Head (50 Hz)
Measured data Head (60 Hz)
500
Res. pressure Head (70 Hz)
10,000 Head (w / CFs)
1,000 Head at op. freq.
Range (min/max)
1,500 Op. Point
True Vertical Depth (ft TVD)

8,000

Total head (feet)

2,000

2,500 6,000

3,000

4,000
3,500

4,000
2,000
4,500
PRES
5,000 0
0 500 1,000 1,500 2,000 0 5,000 10,000 15,000 20,000
Pressure (psig) Average pum p flow rate (rb/day)
 Improved Pump Sizing
• 20,000 bpd pump
After Upsize • In range
• FBHP at optimum 650psi
2300 stb/d Net Oil
0 12,000
From inflow Head (50 Hz)
Measured data Head (60 Hz)
500
Res. pressure Head (70 Hz)
10,000 Head (w / CFs)
1,000 Head at op. freq.
Range (min/max)
1,500 Op. Point
True Vertical Depth (ft TVD)

8,000

Total head (feet)

2,000

2,500 6,000

3,000

4,000
3,500

4,000
2,000
4,500
PRES
5,000 0
0 500 1,000 1,500 2,000 2,500 0 10,000 20,000 30,000 40,000
Pressure (psig) Average pum p flow rate (rb/day)
 Highlighting Pump Wear
Using pump Delta P and Flow to validate pump performance
0 14,000
From inflow Head (50 Hz)
Measured data Head (60 Hz)
1,000
Res. pressure Head (70 Hz)
12,000
Head (w / CFs)
2,000 Head at op. freq.
Range (min/max)
3,000 10,000 Op. Point
True Vertical Depth (ft TVD)

Total head (feet)

4,000
8,000

5,000

6,000
6,000

7,000
4,000

8,000
2,000
9,000
PRES

10,000 0
0 1,000 2,000 3,000 4,000 5,000 0 2,000 4,000 6,000 8,000 10,000 12,000
Pressure (psig) Average pum p flow rate (rb/day)

Production Loss = 650 bopd

 Gauge Monitoring Systems

 Submersible Pump Gauge System

 PCP Gauge System
 ESP PCP Gauge System
 Beam Pump Gauge System
 Gas Lift Gauge System
 Jet Pump Gauge System

38 JMD
 ESP PCP Gauge System

 Intake Pressure
 Discharge Pressure
 Intake Temperature
 Motor Temperature
 Vibration
 Current Leakage
(earth fault monitoring)

39 JMD
 PCP Gauge System

 Intake Pressure
 Discharge Pressure
 Intake Temperature
 Vibration
 Current Leakage
(cable to surface integrity)
 Beam Pump Gauge System

 Intake Pressure
 Intake Temperature
 Vibration
 Current Leakage
(cable to surface integrity)
 Gas Lift Gauge System

 Annulus Pressure
 Tubing Pressure
 Downhole Temperature
 Vibration
 Current Leakage
(cable to surface integrity)
Process of diagnosis

Wellhead

1. ABOVE THE PUMP

Pdischarge = WHP + Pgravity +

Depth Pfriction
1. ACROSS THE PUMP
Pintake = Pdischarge - Ppump
Pump
1. BELOW THE PUMP
Pwf = Pintake + Pgravity + Pfriction
Perforations Pwf = Pr – (Q / PI)

Pressure

43 JMD
Thank you for your attention!

45
 2014

Diagnostic Exercises
ESP Diagnosis
Exercise.
Draw
i) Gradient plot
ii) Pump curve
iii) Trend over time of P Discharge & P intake
iv) Amp Chart

for the following cases

1. Shut in
2. Frequency change 50 –60 hz
3. Broken Shaft
4. Increasing water cut from 0% -> 40% -> 90 %
5. Pump wear

44 JMD
 ESP Diagnostic Exercise

i) Gradient plot
ii) Pump curve
iii) Trend over time of P Discharge and P intake
iv) Amp Chart

For the following cases

1. Shut in
2. Frequency change 50 –60 hz
3. Broken Shaft
4. Increasing water cut from 0% -> 40% -> 90 %
5. Pump wear
Thank you for your attention!

3
 2014

Trend Analysis Exercises

 PRESSURE (PSI)

0
500
1000
1500
2000
2500
3000
03/11/1996 15:50
03/11/1996 16:16
03/11/1996 16:42

40 Hz
40 Hz
03/11/1996 17:09

Start up 1
03/11/1996 17:35 Start up 2
03/11/1996 18:01 40 Hz
03/11/1996 18:27
03/11/1996 19:23
Start up 4

40 Hz
03/11/1996 20:13

Intake Pressure
Start up 3
03/11/1996 20:40
03/11/1996 21:06
46 Hz

03/11/1996 21:32
48 Hz

03/11/1996 21:58
03/11/1996 22:24
03/11/1996 22:51
Choke 18%

03/11/1996 23:17
03/11/1996 23:43
50 Hz 14%choke

04/11/1996 00:09
04/11/1996 00:36
50 hz choke 40%

Shutdown
Choke 20%

04/11/1996 01:02

Discharge Pressure
04/11/1996 01:28
04/11/1996 01:54
04/11/1996 02:21
40 Hz
Data Trending

04/11/1996 02:47
Start up 5

04/11/1996 03:13
Choke gradually
increased to 40%

04/11/1996 03:39
04/11/1996 04:05
04/11/1996 04:32
04/11/1996 04:58
Motor Temp

04/11/1996 05:24
Choke 20%
Frequency increased
to 50 HZ,choke to 25%

04/11/1996 05:50
04/11/1996 06:17
reduced to 15%
Choke gradually

04/11/1996 06:43
04/11/1996 07:09
choke 27%

04/11/1996 07:35
Intake Temp
160
170
180
190
200
210
220
230

TEMPERATURE (F)
Trend Exercise One
Trend Exercise Two
 Trend Exercise Three
Monitoring and Protection

100 3

90
2.5

Temperature, Vibration,Current Leakage

80
2

70
Pressure

1.5

60

1
50

0.5
40

30 0
01/12/1999 01/12/1999 01/12/1999 01/12/1999 01/12/1999 01/12/1999 01/12/1999 01/12/1999 01/12/1999 01/12/1999 01/12/1999
14:24 14:38 14:52 15:07 15:21 15:36 15:50 16:04 16:19 16:33 16:48
Time

Pi(bar) Ti(C) Tm(C) Vibration(g) Cl passive(mA)

5 JMD
 Trend Exercise Four
Monitoring and Protection

250 3

2.5

Temperature, Vibration,Current Leakage

200

150
Pressure

1.5

100

50
0.5

0 0
01/12/1999 01/12/1999 01/12/1999 01/12/1999 01/12/1999 01/12/1999 01/12/1999 01/12/1999 01/12/1999 01/12/1999 01/12/1999
14:24 14:38 14:52 15:07 15:21 15:36 15:50 16:04 16:19 16:33 16:48
Time

Pi(bar) Ti(C) Tm(C) Pd(bar) Vibration(g) Cl passive(mA)

 Trend Exercise Five
Monitoring and Protection

260 3.0
250
240
230
220
2.5
210
200

Vibration g & Current Leakage mA

190
Pressure Bar & Temperature C

180
170 2.0
160
150
140
130
1.5
120
110
100
90
80 1.0
70
60
50
40
0.5
30
20
10
0
-10 0.0
22-Dec-98 22-Dec-98 23-Dec-98 23-Dec-98 24-Dec-98 24-Dec-98 25-Dec-98 25-Dec-98 26-Dec-98 26-Dec-98
Time

Pi(Bar) Pd(Bar) Ti(C) Tm(C) Vibration(g) Cl passive(mA)

 Trend Exercise Six
Monitoring and protection

160 120

140
100

120

Temperature (C), Vibration (g*10)

80
100
Pressure (Bar)

80 60

60
40

40

20
20

0 0
10/7/00 00:00:00 10/7/00 04:48:00 10/7/00 09:36:00 10/7/00 14:24:00 10/7/00 19:12:00 11/7/00 00:00:00 11/7/00 04:48:00 11/7/00 09:36:00
Date and Time

Pi(psi) Pd(psi) Cl passive(mA) Cl active(mA) Ti(C) Tm(C) Vibration(g*100)

 Proactive Analysis Example
Proactive Analysis Critical Delta P
1700
required to 40
1650
produce flow
for motor 35
1600 cooling
30

1550
Pump Delta Pressure (psia)

25
1500
Estimated

Vibration (g)
1450
Point of 20
failure
1400
15

1350

10

1300

5
1250

1200 0
0.25 125.25 250.25 375.25 500.25 625.25 750.25 875.25 1000.25 1125.25 1250.25 1375.25
Time

deltaP(psi) Vibration (g)

Thank you for your attention!

10
 2014

ESP Troubleshooting
 ESP Troubleshooting

 Become familiar with equipment

 Keep the following data to help analyze problems

– Operating data from time unit was installed
– Any reasons for shutdown
– Inspection report on prior runs
– Amp charts

2
 ESP Troubleshooting

 Motor Failure Causes

– Worn pump
– Transient voltages and lightning
– Seal section failure
– Well conditions
– Corrosion
– Electrical system
– Faulty equipment

3
 ESP Troubleshooting

 Pump Failure Causes

– Pump operated outside operating range

– Sand and abrasives
– Plugged pump intake
– Twisted or broken shaft
– Corrosion of pump housing

4
 ESP Troubleshooting

 Seal Failure Causes

– Excessive down or up thrust

– Vibration from wear in pump
– Improper handling causing broken or cracked seal
– Heat buildup due to insufficient fluid velocity past unit
– Misapplication

5
 ESP Troubleshooting

 Cable Failure Causes

– Mechanical damage
– High temperature
– Gas
– Excessive current
– High voltage transients
– MLC damage during installation

6
 ESP Troubleshooting

 Surface Failure Causes

– Transformers
• High transient voltages or overloading
– Motor Controller
• Transient voltages, dust and sour gas
– Surface Cable
• Transient voltages, mashed
– Wellhead
• Power feedthrough

7
 ESP Amp Chart Analysis

– Records motor input amperage

– Review along with production data

8
 ESP Amp Chart Analysis

 Normal

– Symmetrical, smooth curve

– May be above or below NP
GRAPHIC CONTROLS CORPORATION
amps BUFFALO, NEW YORK

NORMAL

NOON
LOCATION

– Full scale startup spike is TRANS. RATIO CIRCUIT

normal DATE ON TIME M.

DATE OFF TIME M.

BR 3800

9
 ESP Amp Chart Analysis

 Power Fluctuation

– Smooth curve
– Small fluctuations cause
by primary power voltage GRAPHIC CONTROLS CORPORATION
BUFFALO, NEW YORK

NOON
POWER FLUCTUATIONS
fluctuations LOCATION

TRANS. RATIO CIRCUIT

– Investigate and try to DATE ON TIME M.

eliminate DATE OFF TIME

BR 3800
M.

10
 Gas Locking
Gradual decrease
- Here is a chart which shows 5 6 AM 7 8 Amperage
4 60 9

startup followed by a gradual 3

50 becomes

10
40

decrease in amperage until 30

20 erratic

11
10
about 7:30 AM where the line

NOON
1
becomes very ragged. At about

MIDNIGHT

1
8:30 AM, the amperage takes a

2
drastic drop and becomes very

11

3
smooth.

10
4
9 5

8
7 6 PM
 ESP Amp Chart Analysis

 Pump Off

– Similar to gas locking B C

except no free gas A
breakout to cause D
fluctuations GRAPHIC CONTROLS CORPORATION
BUFFALO, NEW YORK

– This unit is too large for PUMP OFF

NOON
LOCATION

application TRANS. RATIO CIRCUIT

– Remedial action same as DATE ON

DATE OFF
TIME

TIME M.
M.

gas locking, except add BR 3800

possible well stimulation

– If unit is lowered, check
design because of
increased TDH

12
 ESP Amp Chart Analysis

 False Starts

– Same as pump off

except auto-restart GRAPHIC CONTROLS CORPORATION
BUFFALO, NEW YORK

delay is not long enough FALSE STARTS

NOON
LOCATION

TRANS. RATIO CIRCUIT

DATE ON TIME M.

DATE OFF TIME M.

BR 3800

13
 ESP Amp Chart Analysis

Excessive Cycling

–Similar to fluid pump-off except

running times shorter and cycles
more frequent
–Unit may be is oversized
GRAPHIC CONTROLS CORPORATION
BUFFALO, NEW YORK

NOON
•Check fluid level immediately EXCESSIVE CYCLING
LOCATION

TRANS. RATIO CIRCUIT

after shut-down
•If fluid above pump check DATE ON

DATE OFF
TIME

TIME M.
M.

check for abnormally high BR 3800

tubing pressure

–Plugged discharge line will

result in reduced production rate
accompanied by drop in
amperage
–If discharge pressure is
reasonable or low, check for a
hole in tubing near surface 14
 ESP Amp Chart Analysis

 Gassy Conditions

– Fluctuation is caused by
entrained and free gas in
fluid production
GRAPHIC CONTROLS CORPORATION
BUFFALO, NEW YORK
– Usually accompanied by a “GASSY”

NOON
LOCATION

reduction in total liquid TRANS. RATIO CIRCUIT

production DATE ON TIME M.

DATE OFF TIME M.

BR 3800

15
 ESP Amp Chart Analysis

 Undercurrent Load

– Unit starts, runs a few seconds,

and shuts down on underload
– Generally caused by fluid lacking GRAPHIC CONTROLS CORPORATION

sufficient density or volume to BUFFALO, NEW YORK

NOON
UNDERCURRENT
LOCATION LOAD
load motor TRANS. RATIO CIRCUIT

– If fluid level is above pump, DATE ON

DATE OFF
TIME

TIME M.
M.

undercurrent setting can be BR 3800

adjusted
– Can also be caused by timing
relay not blocking undercurrent
during start

16
 ESP Amp Chart Analysis

 Under Current Below No Load

– Normal start followed by decline

to no-load current followed by
overload fault
– Unit is oversized for application
GRAPHIC CONTROLS CORPORATION

– Unit pumps well down to where BUFFALO, NEW YORK

NOON
UC BELOW NO LOAD
undercurrent relay should stop LOCATION

TRANS. RATIO CIRCUIT

unit DATE ON TIME M.

– Relay either inoperative or set DATE OFF TIME M.

below no load current

BR 3800

– With low or no fluid flow motor

ran at idle until heat build up
caused a motor or cable burn

17
 ESP Amp Chart Analysis

 Tank Level Controls

– Unit is controlled by tank

switch
– Normal operation, but restart GRAPHIC CONTROLS CORPORATION
BUFFALO, NEW YORK

time is too short - minimum CONTROLS

NOON
LOCATION

30 minutes TRANS. RATIO CIRCUIT

DATE ON TIME M.

DATE OFF TIME M.

BR 3800

18
 ESP Amp Chart Analysis

 Overload

– Normal start followed by normal C

operation followed by gradual B
A
increase to overload
– Unit does not automatically start GRAPHIC CONTROLS CORPORATION

after overload so cause can be BUFFALO, NEW YORK

OVERLOAD

NOON
LOCATION

corrected TRANS. RATIO CIRCUIT

– Overload probably caused by DATE ON TIME M.

increase in fluid density or DATE OFF TIME

BR 3800
M.

viscosity
• heavy brine, mud, sand,
motor overheat, worn
equipment

19
 ESP Amp Chart Analysis

 Debris

– Unit started, pumped

erratically, then stabilized
and pumped normally
GRAPHIC CONTROLS CORPORATION

– Expected from clean up of BUFFALO, NEW YORK

DEBRIS

NOON
LOCATION

well debris such as scale, TRANS. RATIO CIRCUIT

loose sand or heavy brine DATE ON TIME M.

DATE OFF TIME M.

– Not unusual but, avoid if BR 3800

possible
– Make sure motor is sized
to lift heavy fluids

20
 ESP Amp Chart Analysis

 Excessive Manual Restarts

– Normal operation until

power fluctuations cause
overload GRAPHIC CONTROLS CORPORATION
BUFFALO, NEW YORK

– Multiple manual restarts EXCESS STARTS

NOON
LOCATION

were attempted TRANS. RATIO CIRCUIT

– Never attempt more than DATE ON

DATE OFF
TIME

TIME M.
M.

one restart without BR 3800

determining cause of
failure

21
 ESP Amp Chart Analysis

 Emulsion, Surface Load or

Generator

– Normal Start followed by a periodic

gradual increase in current followed
by a sudden drop to normal current. GRAPHIC CONTROLS CORPORATION
BUFFALO, NEW YORK

– Possible Causes: EMULSION

NOON
LOCATION

• Emulsion formation in the pump TRANS. RATIO CIRCUIT

which periodically clears. DATE ON TIME M.

DATE OFF TIME M.

Emulsion breaker is required. BR 3800

• Decreasing surface voltage due to

other heavy equipment on the line
which is cycling on and off.
• Unit on a generator if the regulator
is faulty causing fluctuation in
voltage or speed changes resulting
in changing power frequency.

22
 ESP Amp Chart Analysis

 Erratic Loading Conditions

– Unpredictably varying plot

– Fluctuation in fluid gravity or
large changes in surface
pressure usually cause this
GRAPHIC CONTROLS CORPORATION

– Unit finally stopped on BUFFALO, NEW YORK

ERRATIC

NOON
overload and will not LOCATION

TRANS. RATIO CIRCUIT

automatically restart
DATE ON TIME M.

– Should not be restarted until DATE OFF TIME M.

cause if found BR 3800

– Overload could be caused by

worn or locked pump, burned
motor, cable fault or blown
fuses

23
 ESP Amp Chart Analysis

 Mystery Loading

– The motor appears to load up

during one period and then
loses some load during
another period. GRAPHIC CONTROLS CORPORATION
BUFFALO, NEW YORK

MYSTERY

NOON
– Could be the result of normal LOCATION

TRANS. RATIO CIRCUIT

well producing characteristics DATE ON TIME M.

or something as simple as the DATE OFF TIME M.

chart paper deforming due to BR 3800

the change in temperature

from day to night.

24
TO SUMMARISE THE TROUBLE
SHOOTING PROCESS
 Trouble Shooting

Sometimes we are faced with trying to guess what is

wrong with a pump (or well) based on very limited data.

The following illustrations are taken from actual

experiences and, although not comprehensive, may be
helpful in diagnosing problem wells.
 Trouble Shooting

Say we have a new unit in a well which was designed to

produce 3000 BPD. It did for a few days but then is only
making 2500 BPD. We look at the amp charts and find
that the amperage is lower than we expect.

What could possibly be wrong with the unit?

 Trouble Shooting

Was the pump O.K. in the first place?

The first thing to do is determine if we have an actual

performance test for the pump. Not all pumps match the
catalog curve exactly and if this particular pump tested
to the bottom tolerance on the head-capacity curve, it
may be performing exactly as it should.

Well we called the factory for a curve and found that it

tested pretty close to the catalog curve.

What else could be wrong?

 Trouble Shooting

The fact that the pump was performing to the 3000 BPD
requirement for a couple of days and then the production
dropped may lead us to suspect that the well data was
bad.

When the unit is first started, the fluid level in the well is
high due to the static reservoir pressure elevating it.

When we run that unit, we will pump the level down to the
producing level -- but this takes time.
 Trouble Shooting

One other very useful piece of information would be

the intake pressure (or a fluid level if we are taking
sonic shots).

If the fluid level is down, we might suspect bad well

data.
 Trouble Shooting

On the other hand, if the fluid level is

up, we could have several possibilities.

Note on Sonic logs

Be careful, however, when interpreting

sonic levels because if the well is very
gassy, it is possible to have several
hundred feet of low density foam at the
top and the sonic log cannot tell what
the density of the fluid is -- it only
knows where it is.
 Trouble Shooting

We have determined that our fluid level is an

accurate estimate and it is still high.

One possibility is that the pump is partially

plugged internally. This will decrease the
volumetric capability of the pump and restrict the
flow. Since we are not pulling the well as hard, the
fluid level would come up.

In this case, whether the amperage goes up or

down will depend on the shape of the particular
pump's BHP curve.
 Trouble Shooting

Another possibility is that we have a tubing leak

and the pump is recirculating fluid. As far as the
pump is concerned the total TDH requirement is
reduced so the pump is producing more fluid.

Some of it makes it to the surface and the rest

recirculates back to the pump intake. We are again
pulling less on the well (the reservoir is only giving
up what we produce to the surface.

Whether amperage goes up or down again

depends on the shape of the BHP curve.
 Trouble Shooting

So a plugged pump and a tubing leak can look very

much alike. We can pressure up the tubing to
check this without pulling the unit.

It is still possible to misinterpret this as check

valves can leak.
 Trouble Shooting

By the way, it is possible to plug a pump several

places. Partial plugging in the stages themselves
may be inconvenient but not a major problem. We
may even be able to remove this with acid
depending on what type of plugging it is.

If the pump intake were to actually to plug off, the

pump would try to produce the oil out of the
protector and the motor and would most likely fail
very quickly.
 Trouble Shooting

Another possibility for low production with higher

than expected current and fluid level is viscosity.

If we are producing a viscous fluid and we did not

account for this in the design, we would have
undersized the pump for the application.
 Trouble Shooting

If everything checks out okay and our pump

meets the published curve and we still have low
production, we should look at the true rotational
speed of the pump.

Catalog curves are based on 3500 RPM (even

viscosity corrections correct from the 3500 RPM
curve).

If the motor is not turning at that speed, we will

get less production.
 Trouble Shooting

What about the case where the production decline

is not so steep but diminishes gradually over a
few months or even years?

This could be a case of a "worn out" pump or

declining well productivity or both.

If the well productivity is down and the company

decides to workover the well when the pump is
pulled it would be a good idea to review the
expected well productivity before simply replacing
the unit for another of the same size.
 Trouble Shooting

A well workover may enhance the productivity

very much and may even improve it beyond what
it was initially (for the first pump).

If this happens, a resizing of the SPS equipment

may be necessary.

The well could require a larger volume pump if a

higher flow rate is desired and/or fewer stages.
 Trouble Shooting

What happens if we have looked at every

reasonable possibility and accounted for gas,
viscosity, tubing leaks, etc. and know our well
data is very close and we are still too low on flow
and head.

There are a couple of other possibilities we may

have overlooked.

What are they?

 Trouble Shooting

What if we have 540 equipment in a 7" 29 ppf

casing or 456 equipment in a 5.5" 20 ppf casing as
an example?

Depending on the horsepower of the motor, we may

have a restriction downhole that we have not really
accounted for.
1.4 x (BPD)2
1000
Loss in feet per foot of motor length = ___________________________

( I.D. - O.D. ) 3 x ( I.D. + O.D. )2

BPD = Flow Rate in Barrels per Day

I.D. = Casing Inside Diameter in Inches
O.D. = Motor Outside Diameter in Inches
 Trouble Shooting

In higher flow rate -- higher horsepower

applications, even though the motor section is
relatively short in comparison to the total well
depth, we can have a very high pressure drop
across the motor.

If we do not have enough pump stages to

compensate for this, we will be short on head and
this may be very significant.
 Trouble Shooting

To summarize
INVESTIGATE EVERYTHING

EVEN THOUGH IT MAY NOT APPEAR TO BE

RELEVANT TO THE LOSS OF PRODUCTION OR
FAILURE OF THE ESP.

AS A PRACTICAL EXAMPLE:
 Trouble Shooting

We have run across one example of a major oil

company in a very warm climate where most of the
pumps were run on generators.

The Production Department was responsible for

sizing and operating the pumps but the Electrical
Department handled all the generators.
 Trouble Shooting

The Production Department was baffled that

production was down (reduced) and went through
a field wide test to see if the units were operating
in reverse rotation.

Since the units were actually hooked up by the

Engineering Department they were deeply
offended by the implication (by the Production
Department) that they did it wrong and felt that the
ESP provider had initiated this suggestion.
 Trouble Shooting

It turned out that ENGINEERING DEPARTMENT was

slowing down the generators from 55 Hz in the Winter
months to 50 Hz during the Summer without bothering to
tell anybody.

ENGINEERING did not realize this would affect production

and simply wanted to protect the Gen Set.
All the Production Department could tell was that
production was way off and they placed all the blame on
faulty equipment from the ESP supplier.

INVESTIGATE EVERYTHING BEFORE DECIDING

WHAT HAS CAUSED THE FAILURE.
Thank you for your attention!

47
REFERENCE MATERIAL
Report by: Distributed by:

Artificial Lift: Making Your

Electrical Submersible
Pumps Talk To You
Author
Sandy Williams, Director, Artificial Lift Performance Ltd

Disclaimer:
Whilst every effort has been made to ensure the accuracy of the information contained in this publication, neither Offshore Network Ltd
nor any of its affiliates past, present or future warrants its accuracy or will, regardless of its or their negligence, assume liability for any
foreseeable or unforeseeable use made thereof, which liability is hereby excluded. Consequently, such use is at the recipient’s own risk on the
basis that any use by the recipient constitutes agreement to the terms of this disclaimer. The recipient is obliged to inform any subsequent
recipient of such terms. Any reproduction, distribution or public use of this report requires prior written permission from Offshore Network Ltd.

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You

ABSTRACT
Using a diagnostic technique incorporating the use of a gradient traverse plot combined with basic
physics and validation principles, this paper will show how an Electrical Submersible Pump (ESP) can
be designed, installed and operated to increase reliability and extend run-life.

Since interactions between the wellbore and an ESP are complex, a method is required to assist
in the analysis of such interactions. Case histories will demonstrate application of the technique
and highlight the practical benefits where well intervention is costly and must be eliminated or
minimised.

Our industry understands how to design, commission and operate surface pumps to attain reliability.
By applying the same laws of physics and increasing our understanding of ESP performance we can
improve ESP reliability. A discussion of the Gradient Traverse plot analysis technique will highlight
the hierarchy of measured parameters (pressures, temperatures, amps etc) and show how these
parameters can be used to validate well and ESP performance. The technique will then be applied to
interpret complex examples of ESP operating problems and demonstrate how measured parameters
can be used to understand the ESP and well interactions and prevent ESP failures.

This technique facilitates a ‘holistic’ approach to the design, performance monitoring and failure
analysis of an ESP as a system. Consideration of all aspects of well design (well inflow, fluid
properties, completion design, desired rates, cost) is critical to understanding the operating
environment of an ESP. By implementing a technique that allows validation of well and reservoir data
and ESP performance, a better understanding of the system is developed and ‘fit for purpose’ design
can be implemented resulting in a more reliable system, thereby reducing well intervention costs.

INTRODUCTION
The subject of this paper is ‘a technique to make your ESP talk to you’; when designing and
operating ESPs the first thing that you need to do is give the ESP a ‘mouth’, then you can attempt to
understand its language and let it talk to you.

Traditionally, the ESP ‘mouth’ was an amp chart and a fluid level shot on the annulus. Interpretation
of the data then told you something about motor load and pump intake pressure. Today’s ESP
technology is such that the prudent operator can obtain ESP operating parameters by using a
downhole sensor. The dilemma when faced with the wealth of data that comes from such a sensor
is ‘which parameters are most important?’. By recognising which variables respond most quickly to a
change in operating conditions, you can give your ESP the ability to talk to you.

wellhead pressure, wellhead temperature, choke setting, total flow rate, oil
Surface
rate, gas rate, water rate, produced gas oil ratio, density ofproduced fluids,
Parameters:
frequency, amps, volts.

pump discharge pressure, pump intake pressure, pump intake temperature,

Downhole
pump discharge temperature, motor oil temperature, motor winding
Parameters:
temperature, vibration, current leakage, flowrate.

Table 1 – Summary of operating parameters for an ESP well.

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 2
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You

Table 1 identifies some of the parameters that an operator may routinely collect on an ESP well.

To focus on which parameters are the most important, the purpose of the ESP needs to be
recognised - to assist a well to produce fluid (hopefully oil) to surface. As such, the ESP should be
considered a part of the well system rather than a pump in isolation. Flow occurs in the well as a
result of reservoir drawdown; an ESP assists the well to flow (or to flow at higher rates) by increasing
the drawdown. Consequently, a change in well behaviour will manifest itself as a pressure change in
the wellbore. This paper will show how pressure information from a downhole sensor on an ESP can
be used to match and diagnose well and ESP performance.

THE TECHNIQUE
The technique utilises the gradient traverse plot to present wellbore pressure and depth information
in a format that makes it easy to grasp the controlling factors in a well. By combining well inflow and
the physics of fluid dynamics it is possible to understand the theory behind the shape of the gradient
traverse plot. The theory can then be applied as part of a step-by-step process to validate ESP
performance, well information and reservoir data.

The Gradient Traverse Plot

Figure 1 – Typical trend plot of data from an ESP start-up.

The graph in figure 1 shows a typical trend plot of data obtained during the first 24 hours runlife
of an ESP. From the plot it can be seen that increasing frequency causes a greater drawdown at
the pump intake, discharge pressure stays relatively constant. Motor oil and intake temperatures
increase as warmer fluids flow from the reservoir into the wellbore. The first few hours of data
for a well are important as these provide the signature of how the ESP interacts in the producing
wellbore. Changes in these parameters will represent a change in well or pump performance. By
observing trends over time, changes can be identified and the well analysed to determine what has
caused the change. The trend plot is essentially the life story or ‘movie’ of the pump system in the
wellbore.

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 3
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You

Figure 2 – Gradient traverse plot for a naturally

flowing well. Flowrate 1700 BFPD, Watercut
0%, Reservoir pressure 3500 psia, Wellhead
pressure 100 psig.

To analyse the well when a change occurs

takes a little more work but can be done fairly
easily with most production optimisation
software. The key to understanding what
is happening in the wellbore is to develop
an understanding of the pressure regime
in the wellbore. The pressure regime can
then be plotted on a graph of pressure vs.
depth, known as the gradient traverse plot
(or the ‘snapshot’ of what is happening at
any moment in time). To best explain the
gradient traverse plot it is easiest to consider
a naturally flowing well (see figure 2).
This plot shows wellhead pressure (WHP),
bottomhole flowing pressure and reservoir
pressure. The plot was produced for a rate of 1700 BOPD with 0% watercut, a reservoir pressure of
3500 psia and a wellhead pressure of 100 psig. It can be observed that the curve of the pressure vs.
depth line is fairly straight below a depth of 6000 ft. The change in the slope of the line from 6000
ft to surface is due to a reduction in fluid density due to gas break out (the bubble point was approx.
1600 psia).

Figure 3 – Gradient traverse plot for a naturally

flowing well. Flowrate 565 BFPD, Watercut 0%,
Reservoir pressure 3500 psia, Wellhead pressure
200 psig.

Using the gradient traverse plot it is relatively

easy to demonstrate graphically the effect of
a change in an operating parameter. Figure 3
shows the same well but this time the plot is
produced for an increased wellhead pressure
of 200 psig. The effect of this change is to
reduce production to 565 BOPD. Looking at
the plot shows that the line has moved to
the right; increased wellhead pressure has
resulted in a higher bottomhole pressure (less
drawdown) which means the well will flow
less.

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 4
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You
Figure 4 – Gradient traverse plot for a well killed
due to increased watercut.
Flowrate 0 BFPD, Watercut 12%, Reservoir
pressure 3500 psia, Wellhead pressure 200 psig.

Figure 4 shows the gradient traverse plot for

the same well but represents the case where
the well is producing water as well as oil. For
the case shown, the water-cut is 12%. Notice
also that the plot of pressure vs. depth is more
of a straight-line i.e. the effect of the gas is
not as significant. The hydrostatic of the fluid
column is such that the bottomhole pressure
is approximately equivalent to reservoir
pressure, there is no drawdown and the well
can no longer flow.

A quick review of these plots shows that

pressure changes in the well bore will move
the curve across the chart whereas a change in
density of fluids causes a change in the slope
of the line…and now for the good bit …we
get to put an ESP in the well!

Figure 5 – Gradient traverse plot for an ESP

lifted well. Flowrate 3330 BFPD, Watercut 12%,
Reservoir pressure 3500 psia, Wellhead pressure
200 psig.

Figure 5 shows the same well producing

at 12% watercut but with the assistance
of the ESP the well can now produce at a
rate of 3330 BFPD. Take a close look at the
gradient plot, observe how the plot to a
depth of 10000ft is almost the same as the
case represented in figure 4 (there is a slight
difference due to frictional pressure drop).
Now, notice the effect of the ESP (at a depth
of 10000 ft) - the head produced by the
pump causes the pressure at the pump intake
to be reduced. The line representing flow
below the pump is shifted to the left causing
a much lower bottomhole flowing pressure
and allowing the well to flow at significantly
higher rates than the when the well was
flowing naturally. Examination of the gradient
traverse plot for a well with an ESP installed
illustrates a fundamental principle - ESPs suck
rather than push!

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 5
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You
The useful thing about the gradient traverse plot is that it can be used to draw a picture of what’s
happening in the wellbore. By presenting the data in this format it is easy to understand what the
controlling issues are that govern productivity of the well. You may even stand a chance of being able
to explain it to your supervisor!

Figure 6 – Figs 2,3,4 & 5 presented as inflow-outflow performance.

Figure 6 shows the same information as the gradient traverse plots but presented in the more
traditional inflow-outflow relationship. Note the unusual (but characteristic) shape of the ESP lift
curve.

Theory

To make the gradient traverse plot useful we need to do something more meaningful than draw
nice plots. A good starting point is to consider the physics behind the shape of the curve. Examining
the gradient traverse plot for a well fitted with an ESP we observe that there are three points that
characterise the shape of the curve; wellhead pressure, discharge pressure and intake pressure. In
most cases the reservoir pressure will also be known. The relationship between the parameters can
be defined as follows

Above the pump:

Pd = WHP + ∆Pg above pump + ∆Pf above pump + ∆Pa above pump

Across the pump:

Pi = Pd - ∆Ppump

And below the pump

Pwf = Pi + ∆Pg below pump + ∆Pf below pump + ∆Pa below pump

Also from an inflow perspective we can say

Pwf = Pr – (Q / PI)
Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 6
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You
Where
Pd = discharge pressure (psi)
Pi = intake pressure (psi)
WHP = wellhead pressure (psi)
∆Pg = hydrostatic head (gravity pressure loss in psi)
∆Pf = frictional pressure loss (psi)
∆Pa = acceleration pressure loss (psi)
∆Ppump = (Head in feet x fluid specific gravity/2.31) (psi)
Pwf = bottomhole flowing pressure (psi)
Pr = reservoir pressure (psi)
Q = well flowrate in stock tank barrels liquid (stbl/day)
PI = productivity index (stbl/day/psi)

By using the above relationships to understand the output from your production optimisation
software it is possible to calculate water cut, mixture density, pump head, bottomhole flowing
pressure and flowrate.

For oil wells the acceleration term is negligible and so can be ignored in the majority of cases
(particularly on wells that are ESP lifted). Also, in oil wells the gravitational term comprises a
minimum of 80% of the pressure change in the wellbore and in some cases can be as high as 99%,
the frictional term therefore becomes the secondary factor rather than the governing parameter.
Since software is available to calculate the pressure drops in the wellbore we do not need to concern
ourselves, for the purpose of this paper, with how to calculate the values manually. Suffice it to know
that gravity forces are the big bit.

Table 2 – Comparison of friction and gravity effects on the ESP produced well.

Table 2 provides the density and friction terms used to produce figure 6 and illustrates, in this case,
that the gravity term is 95% of the pressure loss whereas friction is only 5%. Note: For viscous crudes,
emulsion-forming fluids or in the case of high rate and narrow tubing the frictional pressure drop can
become as high as 20%.

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 7
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You
Since flow in the wellbore occurs due to pressure effects, it is evident that the parameters that
respond most quickly to tell us what is happening in the wellbore are discharge and intake pressure at
the pump. Historically, amps have been used to analyse ESP performance - in actual fact what they are
measuring is motor load and change in motor conditions in response to change in pressure or flow in
the wellbore. Why not just analyse pressures from the start?

Validation

In the majority of cases the hierarchy of parameters for giving an indication of what is happening to the
ESP system is:

• Discharge and Intake pressure,

• Intake and Motor Temperature,
• Amps and Flow.

None of these parameters used in isolation will tell you what is going on in the ESP but when used
in conjunction with the gradient traverse plot (and the previously stated relationships) it becomes
possible to analyse and determine what is happening in the wellbore by applying the following process:

1. Work top (WHP) down

2. Match pump discharge
3. Match across pump (gives pump intake pressure)
4. Calculate Pwf
5. Compare to well inflow
Consider flow in the well above the pump. We have already developed the relationship for what
happens above the pump:

Pd= WHP +∆Pg + DPf

The known parameters, from measurement, should be WHP and discharge pressure, if these two
parameters are known then the beginning and end point for the curve above the pump are known. The
density of the fluid and frictional forces in the tubing affects the shape and slope of the curve. By 5
using the output from our production software we can match our wellhead and discharge pressures to
determine water cut, tubing gas oil ratio (GOR), frictional pressure drop, specific gravity of the fluid at
the pump discharge and in the process we validate (principally) our fluid property (PVT) and watercut
information.

Similarly, across the pump

Pi = Pd - ∆P

The known parameters are discharge pressure and intake pressure. The DP produced by the pump is
affected by frequency, flow, number of stages, specific gravity of fluid and pump efficiency. Normally
the frequency and flow are known (but not always). The DP value can be converted to head and
plotted on the pump curve. When the operating point fits the pump curve then we have validated our
data, when the operating point does not fit the curve then we must look at frequency, flow, number
of stages (broken shaft), gas effects (or specific gravity of fluid), pump wear, viscosity and emulsion
effects. Note: The operating point on the pump curve should always be the volume downhole (i.e.
surface volume with formation volume factors applied to correct to a downhole volume).

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 8
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You
And below the pump
Pwf = Pi + ∆Pf + ∆Pg

Since the PVT and watercut should already have been validated above the pump, it should be
straightforward to model down from the pump intake to the reservoir interval and calculate a
bottomhole pressure (note that the friction term below an ESP is usually small due to the larger
casing diameter.) The bottomhole pressure should also correspond with that calculated from the
inflow performance relationship for the given flow rate. Knowledge of the bottomhole pressure
allows you to determine PI or reservoir pressure (one has to be known and the other can be
calculated).

• Validate PVT
• Watercut
Above the pump • Fluid specific gravity at pump discharge
• Tubing GOR
• Measure friction effect

• Validate Q, frequency, number stages

• Measure viscosity and emulsion effects
Across the pump • Fluid specific gravity at the pump intake
• % free gas at the pump intake
• Obtain operating point for pump curve

• Calculate bottomhole Pwf

Below the pump
• Obtain PI or Pr

Table 3 – Summary of calculated / derived parameters that can be obtained from the gradient traverse
plot.

Table 3 provides a summary of the parameters that can be derived using the gradient traverse
plot. By focusing on the wellbore rather than just the pump and motor, it is possible to determine
more than “is the pump happy?”; the information can be used to perform production allocation
and prediction of future well performance including production optimisation. Additionally, this
information can be used for intelligent control systems or to define well specific alarm and trip
settings to assist in failure prevention.

Using the analogy of pump in a refinery or platform topsides, facilities engineers use flow indication
in conjunction with pressure and temperature readings as the main indicators of what is happening
to a pump (of course the motor would also be protected using current trips). Why, when we put a
pump downhole, do we try and diagnose the performance of the pumping system using a current
(amps) reading? By putting comprehensive instrumentation on a downhole pump and using the data
from the pump we can listen to our pump talking to us.

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 9
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You

THE APPLICATION
Consideration of the following examples and case histories should help to explain how the technique
is applied in practice to analyse and predict the performance of the well and ESP.

Case 1 - Failure Analysis or Failure Prevention – Your choice!

Figure 7 – Trend analysis plot for an ESP start-up and failure in less than 2 hours.

The ESP in this well failed in less than 2 hours. Why? The trend analysis plot (figure 7) shows a typical
pattern of the ESP being started and intake pressure reducing, the well is brought on against a
closed choke and so a rise in discharge pressure is observed. Choke changes at surface cause some
changes in the discharge pressure reading but essentially the discharge pressure remains stable;
changes in frequency cause the intake pressure to decline further (to 54 Hz). It can be observed
from the plot of pump dP (Pd minus Pi) that from 54 Hz onwards the pump dP is declining. Normally
a declining pump dP over time is indicative of higher flow (i.e. the operating point on a head curve
moves from left to right), in this case the declining dP is not due to increasing flow. At around 17:40
the frequency is increased to 55Hz, the pump dP continues to decline and the motor temperature
starts to increase. The ESP failed eventually on downhole over-current.

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 10
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You

Figure 8 – Gradient traverse plot for Match #2.

Figure 8 shows the gradient traverse plot for the point shown as Match #2 on the trend plot
(figure7). From the gradient traverse plot it can be seen that even with the pump operating the
bottomhole pressure is 150 psi higher than reservoir pressure. The well is unable to flow. Plotting
the head on the pump curve (calculated from measured dP) also confirms that the pump is in a
no flow condition. At this point everything is operating ‘normally’, even though the well is not
flowing, the pump head developed conforms to the head curve at the operating frequency. (Note:
A downhole flowmeter was installed in this well and was reading zero flow. The flowmeter was
deemed to be not working correctly.) A reduction in bottomhole pressure below reservoir pressure
to create a flowing condition could have been achieved in two ways; by reducing the wellhead
pressure by 200 psi or by increasing the frequency to 58Hz. Neither of these actions was taken.

Figure 9 – Gradient traverse plot for Match #3.

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 11
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You

Figure 9 shows the gradient traverse plot shown as Match 3 on the data trend plot after the
frequency had been increased to 55 Hz. The plot shows that the pump head has decreased and no
longer conforms to the expected head curve for the operating frequency indicating damage to the
pump impellers / diffusers. The ESP ran a further 30 minutes before failing on over-current.

At no time during the life of this pump was enough reservoir drawdown achieved to allow the well
to flow; this was the fundamental cause of failure. This analysis was performed as a failure analysis
which could so easily have been failure prevention if the appropriate tools had been used during
startup of the well.

Case 2 – Validation of PVT and PI

This example illustrates the extent of information that can be gained by an analysis of the data
collected from monitoring an ESP system. Apart from diagnosing pump performance, the data has
been used to infer fluid properties, well flowrate, watercut and bottomhole pressure.

This well is a recompletion to an upper reservoir zone. The zone had been tested during a drill stem
test in an offset well as part of the original field discovery. No PVT data existed for this zone. The ESP
system for this well was designed using PVT for the lower reservoir. By using the trend analysis plot
in combination with the gradient traverse plot it was possible to highlight that hydrocarbon (PVT)
properties were very different from those that had been used for the ESP design.

Matching the discharge pressure (above the pump) for a known WHP and watercut proved to be
unsuccessful using the PVT for the lower zone. Validation indicated that to match the discharge
pressure a light but dead oil fluid (32º API) with unusually low GOR
(6 scf/stb) and bubble point (50 psia) would have to be present.

Figure 10 – Trend analysis plot.

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 12
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You
PI was calculated from the stabilised flow period on the 14-Jun-99 as shown in the trend plot in figure
10. Watercut (5%) was calculated from the pump discharge pressure and known wellhead pressure,
and flowrate was derived from the pump performance curve (1900 stbl/day) thus allowing accurate
calculation of flowing bottomhole pressure. With a given reservoir pressure of 2404 psia, the PI was
calculated to be 1.0 stbl/day/psi . This PI was used to cross-correlate flowrates in earlier time periods.

Figure 11 – Gradient traverse plot for Match #1.

Match 1 Match 2

WHP (barg) 6 6

Pd (psia) 2335 2100

Pi (psia) 962 814

Total flowrate (stbl/d) 1370 1550

Watercut (%) 80 15

Pwf (psia) 1060 922

Table 4 – Summary of parameter changes as the well cleans up.

The two match points clearly show the well rapidly unloading from 80% down to 15% watercut with
an associated decrease in bottomhole flowing pressures and hence increase in flowrates (from 1370
to 1550 stbl/day). The gradient traverse plot shown in figure 11 represents match point 1. Table 4
summarises the key parameter changes between the two match points. The pump is operating in
downthrust throughout the start-up period. This indicates that the PI used to design the ESP was higher
than the actual productivity of the well. For the next completion of this well a smaller pump will most
likely be run. This example shows how the data obtained from an analysis can be used as part of an
iterative process to validate fluid properties and well inflow performance. This information can then be
used to design future ESP installations.

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 13
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You
Case 3 – Viscosity Correction

This example illustrates how knowledge of the pump intake and discharge pressures can be used to
measure the effect of high viscosity on pump performance and flow in the tubing.

Figure 12 – Reduction in pump performance with high viscosity.

Many papers have been published concerning the effects of viscosity and emulsions on pump
performance. The published material shows that the normal pump affinity laws do not apply for
viscous crudes and the ability of a pump to generate head and handle throughput are reduced.
Figure 12 shows a plot of this effect, produced from tests performed using high viscosity dead
crudes.

The prediction of the effect of viscosity (or emulsions) on pump performance is very difficult. Any
form of predictive analysis or modelling makes assumptions about the change in viscosity of the
fluid (with heating and change in pressure) as it passes through the pump. Viscosity correlations are
available for flow in the tubing but they are not valid for flow through the pump; attempts to model
or calculate these viscosity effects are liable to miscalculate pump dP (head).

To design and size a pump for use in heavy crudes, correction factors are used for head and flowrate.
These correction factors are based on laboratory tests that have been performed using selected
pumps in a variety of dead crudes (typical correction factors are 0.8 on head and 0.5 on flow). The
results of these tests are then extrapolated to other pumps to assign correction factors so that
viscosity effects can be accounted for in pump design. This is a good approach where no existing
data exists for pump performance in a specific field. However, for any operating ESP system the
actual viscosity effect can be determined. Measured values of Pd and Pi give an exact pump dP and
hence direct measurement of the viscosity effect at the operating condition.

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 14
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You

Figure 13 – Reduction in pump performance with high viscosity based on measured parameters.

If WHP, Pd and Pi are measured we can prepare a gradient traverse plot and plot an operating point
on the head curve for the specific operating frequency (see figure 13). The gradient traverse plot
allows us to:

• validate tubing viscosity effect (above the pump);

• determine the pump correction factor (across the pump) and;
• validate well inflow performance (below the pump).

This technique is particularly valuable in viscous crudes as the tubing viscosity effect determination
can be used to refine flow correlations and the empirically derived pump correction factors can
be used for design of future pumps. Well inflow performance is used for history matching and
production profile prediction.

Case 4 – Fluid level validation and well performance prediction

Figure 14 – Fluid shot indicates very low pump

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 15
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You
In many land operations where ESPs are relatively inexpensive and can be worked over easily, no
sensors are run downhole. Operators analyse such wells by performing a fluid level measurement
and converting fluid level to pump intake pressure. In this case, the gradient traverse plot was used
to validate well performance data and determine that the fluid level measurement was in error.

The well was known to be producing approx. 5000 stbl/d with a watercut of 54% operating at a
wellhead pressure of 160 psi. The pump installed in the well was direct on line and operating at a
frequency of 60Hz. The indicated bottomhole pressure from a fluid level measurement was 2100
psig. Using production optimisation software the discharge pressure was calculated to be 2306 psig
based on well PVT, watercut and WHP. Plotting this information on the gradient traverse plot (figure
14) shows a very low pump dP (200 psi). This would only be possible if either the flowrate is 30%
higher or the pump is damaged (30% effective stages) and well PI is wrong. One of the initial input
measurements has to be incorrect!

Figure 15 – With Corrected Pump Intake Pressure. (1580 psig vs 2100 psig from fluid level measurement)

The operator was confident that the metered production from the well was accurate, the calculated
discharge value for the produced fluid PVT should be within +/- 5% and since the pump was relatively
new it was unlikely that 70 % of the stages were inoperative. By a process of elimination the pump
intake pressure value determined from the fluid level measurement had to be wrong. Using a pump
intake pressure of 1580 psig the data was re-examined (figure 15). The revised intake pressure is
consistent with inflow performance assumptions of reservoir pressure of 4300 psig and a PI of 15
stbl/day/psi.

It can also be observed that the pump is operating in upthrust. Using the gradient traverse plot in
a predictive mode it is possible to determine what would have to be done to operate the pump
‘inrange’. Since the pump is direct-on-line it is not possible to alter frequency so wellhead pressure
is the only means of control. Increasing the wellhead pressure to 250 psi will reduce the flowrate to
bring the pump in range. Operating at a lower well flowrate will improve ESP efficiency and improve
pump runlife, although whether or not the operator is prepared to accept lower production rates for
longer ESP runlife requires a cost benefit analysis.

Note: In this case the fluid level measurement was shown to be 33% in error.

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 16
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You
Case 5 – Production Allocation

Measured parameters can be used throughout the life of the ESP to derive watercut, flowrate,
bottomhole flowing pressure and well productivity index (or reservoir pressure) to assist with
production allocation and reservoir management.

The technique was applied to a well where it was not possible to perform well tests on a regular
basis. By using measured values for WHP, Pi, Pd and frequency, derived values were obtained
for watercut, flowrate, Pwf and PI. Table 5 tabulates the derived parameters for this well and
shows how the technique was used to monitor well clean-up following a workover and to provide
production allocation information for a 6-month period after the watercut had stabilised.

Table 5 - An example of production allocation using measured parameters.

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 17
Artificial Lift: Making Your Electrical Submersible Pumps Talk To You

BENEFITS / CONCLUSIONS
In the above examples an attempt has been made to focus on understanding ‘normal’ ESP
performance and how being able to understand what is happening can prevent failures, determine
what is occurring in the wellbore (reservoir, pump, tubing), provide information for ESP design and
aid control and operation of the system. None of these examples used amps to perform diagnosis;
the wells were analysed purely from a hydraulic standpoint (although amps should be used to verify
that the electrical indications match the hydraulic interpretation).

The key to using this technique is to understand what is happening in your well when you first turn
on your ESP. By understanding present and past performance we can quickly identify and analyse
any change in performance. Using this same technique it is possible to diagnose problems such as
plugging at the pump intake, gas locking, tubing leaks, broken pump shafts, pump wear or wrong
PVT (design information). Additionally the method can be extended to predict future well behaviour
based on sensitivity type analyses for watercut, GOR or reservoir pressure.

Understanding your wellbore from a holistic standpoint can help prevent premature ESP failure (as
shown in case 1), be used to set appropriate alarms and trips on ESP control and protection systems
(calculated specifically for each well) and gives the operator the tools to improve pump runlife.

This technique is inexpensive to implement and does not require any new equipment, software
or tools. It simply requires a little education and an understanding of which parameters tell us the
most about what is happening in a well. Most operators already collect the data required to perform
this type of interpretation. In many cases so much data is being collected that people do not know
where to begin analysing the data. This paper identifies the parameters and the technique that
should be your starting point in attempting to understand well performance in ESP lifted (and other
forms of artificial lift) wells. In essence this technique is no different from traditional nodal analysis
techniques, the difference is that the way the information is presented is easy to understand. With
a little practice you too can use this technique to let your ESP talk to you ……..Are you prepared to
listen?

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 18
Borets Vortex Gas Separator
GSVB5-250

2013
 Borets Gas Handling Units Performance
Capabilities
Maximum intake free gas content , %*

Vortex gas separator GSVB5-250 + Gas 90%

handler MNFB type 3

Vortex gas separator GSVB5-250 72%

Gas handler MNFB type 3 70%

Rotary gas separator + disperser 68%

Tandem rotary gas separator 68%

Rotary gas separator 57%

Disperser 55%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90%

* Tests were performed at Borets facilities in accordance with Borets procedures, at 2910 rpm

Borets offers two types of gas handling units, depending on the solids content in the formation liquid:
 Standard – solids content is up to 0.5 g/l;
 Wear-resistant – solids content is up to 1g/l.
2
 GSVB5-250
Properties and Characteristics

 The operating principle is based on using the free vortex for the gas separation.
 Vortex Gas Separator doesn’t have a separating rotor, what differs it from centrifugal
gas separator.

GSVB5-250 vortex gas separator main parameters with 1000 kg/m3 density water

Parameter Value

Maximum single-phase fluid flow rate, m3/day 250

Separation factor, % 60-80*
Head, m 0.4
Power, kW
- at maximum thrust load 0.9
- transferred 95
Rotation speed, rpm 3000

* In accordance with test results performed at Gubkin Oil and Gas Russian State University at 2910-4500 rpm.
3
 Borets Vortex Gas Separator Performance
test
Tests were performed at Gubkin Oil and Gas Russian State University.
Correlation of intake gas content and separation factor with initial liquid flow at various rpm.
βrg* = 25%.

4
 Borets Various Gas Separators
Endurance tests at 2910 rpm
Tests were performed at Gubkin Oil and Gas Russian State University.
part wear part wear

Part

Rotary gas separator Vortex gas separator

Description
at 2910 rpm at 2910 rpm
Liner wear speed, micron/hour* 0.375 0.125
Projected operational life till complete liner
560 1600
cut, days
* Solids content is 200 mg/l
5
 GSVB5-250
Endurance tests at 3600 rpm
Tests were performed at Gubkin Oil and Gas Russian State University.
Vortex gas separator component part wear

Gas separator GSVB5-250 estimated run-life could reach up to 3500 days with
solids content up to 200mg/l (in accordance with test results performed at Gubkin
Oil and Gas Russian State University at 2910-4500 rpm).
6
 GSVB5-250 Supply Dynamics and Field
Trials

Est.

 Totally 120 vortex gas separators were supplied to different oil companies.
 5 units of GSVB5-250 gas separators are under field trials at Rosneft oil company.
 As of March, 2013 the MTBF totaled 312 days.
 Current MTBF is 436 days.

No failure caused by Borets equipment occurred during field trials

7
 GSVB5-250 Failure Analysis Results

 MTBF totaled 436 days at 402 mg/l solids content;

 Failure analysis results :
– free shaft rotation;
– no working parts, liner, base and head wear;
– bearing sleeves minor wear.

Borets GSVB5-250 vortex gas separators trial results were recognized as successful.

8
 GSVB5-250 Vortex Gas Separator Benefits

 Increased reliability due to the absence of separating rotor and correctly

selected gas separator design type and materials reduce the flow
hydraulic resistance and as a result – wear decrease.

 Vortex gas separator wear-resistance is three times more than rotary gas
separator.

 Operates with formation fluid solids content up to 1g/l at 3500 rpm

 Operates with formation fluid solids content up to 0.5g/l at 3500-4500 rpm

 Separation quality of vortex gas separator is higher than rotary gas

separator.

9
OFFSHORE NETWORK WOULD LOVE TO HEAR FROM YOU…

BIOGRAPHY:
SANDY WILLIAMS
Sandy Williams has worked in the Petroleum Industry for 24 years. He is the founder of Artificial Lift
Performance Ltd (ALP), a company which specializes in helping operators get more oil - a skill which
he explains to the layman as being ‘like Viagra for oil wells’.

Sandy worked 9 years for Amoco, then Phoenix and Schlumberger Artificial Lift before becoming
a consultant focused on artificial lift and production optimization. He has worked and lived in the
USA, Oman, Venezuela, Ecuador, Colombia and has taught over 150 courses related to production
optimization and artificial lift and is fluent in Spanish.

He started ALP 10 years ago, a company focused on assisting operators: improve the application of
artificial lift; reduce productions costs and; increase staff competency. ALP also offers a number of
software products to help operators manage artificial lift. ALP is currently assisting operators on
projects in the UK North Sea, Colombia, Oman, Qatar and the USA.

Offshore Network Ltd. is an independent business intelligence & conference provider catering specifically to the
offshore oil & gas industry. We exist to facilitate a safe and efficient future for the exploration and production of
oil & gas around the globe. We do this by uniting the most influential figures in the industry to challenge the status
quo and share cutting edge innovations. This all happens at our industry leading conferences and through our
original content.

If you would like to contribute to this discussion or are interested in taking part in a future Q&A or article,
please contact Offshore Network today:

info@[Link] Join the Offshore Engineering Group on LinkedIn

UK Telephone: (0)203 411 9937

Follow Offshore Network on Twitter
US Telephone: 713-5706-576

[Link]

Offshore Network Limited, Registered in England and Wales, Company Registered Number 8702032 PAGE 19
 BORETS COMPANY

case study: BORETS EQUIPMENT OPERATION

IN GASSY ENVIRONMENT

On the majority of the oil fields in Russia oil is produced in harsh operating environ-

ments wich effect and reduce efficiency of artificial lift methods.

Borets is focusing on new equipment development and improvement of serial equip-

ment that increase the pumping system efficiency through the whole range of challenging

conditions.

Borets engineers have developed vortex gas separator and new design of gas han-

dler to provide stable operation in the conditions with up to 90% of free gas.

Gas handlers application results

More than 400 Borets gas handlers were supplied to the different regions of

Russia. During the field trials with the Rosneft Company, on January 2012 the MTBF

reached 652 days, with a maximum current MTBF of 707 days, there was no failure caused

by gas handler.
 BORETS COMPANY

Vortex gas separators application results

Vortex gas separators GSVB-250 field trial was performed together with Rosneft Company. As on February

2012 their MTBF was 239 days, maximum current MTBF reached 436 days.

The gas separator disassembling showed no wearing of the working parts, liner, base and head.

Based on trial results, Borets vortex gas separators GSVB5-250 were recognized as successful and recom-

mended for the operation in the wells of the Rosneft Company.

5, Moldavskaya st., Moscow, 121467, Russia

Tel./Fax: + 7 495 660 - 21 - 90
e-mail: office@[Link]