Pressure Transient Analysis-

Introduction and Theory

 What is Well Test Interpretation (WTI)?

• Put simply, WTI uses rate and pressure data measured at

the wellbore to estimate well and reservoir properties
• The analysis is based on the transient flow equation which
governs the well behaviour as a function of time for a
given reservoir. The science of WTI is called pressure
transient analysis or PTA
• The equations are solved assuming constant terminal
rate
• WTI is used on a well test in exploration wells, on an ad
hoc basis in development wells when problems are
suspected, and continuously with permanent downhole
gauges (PDGs).
2
 Why Test?

Very similar rocks can result in drastically-different

permeabilities.
3
 Scales

Given a Reservoir Dimension – 5km x 5km x 50m ~ 109 m3

• Core Dimensions: 30 – 100 cm3 core = 10-13

reservoir

• Log Dimensions: 150 m3 log = 10-9

reservoir

• Well Test Dimensions: 105 m3 test = 10-4

reservoir

• Surveillance Dimensions: …up to… test =1

reservoir

4
Well Testing Resolution

Resolution of
Reservoir Features:

 no resolution

 ↓

 ↓

 best resolution

5
Kjonsvik et al., 1994 SPE 28445
 q Radial Flow
p

Flow causes a pressure disturbance or dP measured at

the well which changes with time as the ‘pressure front’
moves away from the well.

The simplest ‘pressure front’ is radial flow.

6
 Darcy’s Law

In radial terms:
q k dp
V 
2 rh  dr
Where the radial distance is now the distance over which
the pressure drop occurs. At a radius of investigation re,…

q re
pe  pw  ln
2 kh rw
Where the subscripts “w” represent the “well” and the “e”
represents the effective radius in the reservoir at any
particular point in time.
7
Darcy’s Law- Radial, Linear and Spherical

kA p ( L)  p (0)
q
BL

2kh( p e  p w )
q
 re 
B ln  
 rw 

4k ( pe  p w )
q
1 1
B   
 rw re 

8
 Diffusivity Equation

To describe the movement of pressure vs. time in a

reservoir Darcy’s Law (Equation of Motion) is combined
with Conservation of Mass (Continuity equation), and
Equation of State to derive the diffusivity equation:

1   p  ct p
r  
r r  r  k t

NOTE: the terms highlighted.

9
 Diffusivity Equation

The diffusivity equation is non-linear. It is linearized by

assuming...

• Constant k, , , ct
• Low pressure gradients
• Incompressible (cf =0) and fluid saturated formation

For gas special transforms are used to linearlize the

strong pressure dependency on g, cg and Bg.

10
 Diffusivity Equation

 Behavior of o, co and Bo variables as functions of pressure- these are

assumed to be constant for the diffusivity solution solved analytically.

 This is reasonable in oilfields and rich condensate gas fields operated

with p > pbp or p > pdp

In these conditions well performance is controlled by three factors:

Darcy kh/µ, Skin and reservoir geometry/heterogeneity.

11
 Diffusivity Equation Issues

Other times the diffusivity equation has ‘issues’:

• k or porosity changes with pressure (compaction)

•  or ct changes over time (gas or water breakthrough)

• high pressure gradients (large drawdowns)

12
Diffusivity Equation and Fluid Type

13
 Radius of Investigation Rinv

The radius of investigation Rinv defines the location of the

pressure disturbance as a function of its diffusivity and time:

kt
Rinv  0.029
ct

NO dependence on rate, however rate has to be sufficient to

cause a pressure disturbance that can be measured with the
gauges.

14
 Radius of Investigation Components
RI is considered in two parts:

kt
Rinv  0.029
ct

• Higher Mobility M ( k/µ ) : ‘Darcy’s Law’

causes faster pressure transmission

• Higher Storativity F ( Øct ) : ‘Material Balance’

causes slower pressure transmission

Increasing permeability has the same effect as

decreasing any of the other parameters Ø, µ, ct
15
 Rinv moves away from well and
‘traces out’ the reservoir
q

pi Rinv Rinv Rinv Rinv

The location of the ‘pressure

front’ increases with time

well R
16
 Rinv Examples

 Radius investigated in 1 minute in a good-quality oil reservoir.

Mobility: k = 100 mD / µ = 0.8 cp
Storativity: Ø = 0.25 x ct = 10E-6

100(1 / 60)
rinv  0.029 6
 27 ft
(0.25)(0.8)(10 10 )

 Radius investigated in 1 week in a low-permeability, low-pressure

gas reservoir.
Mobility: k = 0.010 mD / µ = 0.05 cp
Storativity: Ø = 0.05, ct = 200E-6

0.010(168)
rinv  0.029 6
 55 ft
(0.05)(0.05)(200 10 )
17
 As Rinv increases the dP at the well
changes
q
Rinv(t1) Rinv(t2)

(∂p/∂r)

dP at the well includes a term that

reflects the properties at Rinv

well R
dP at the well includes a term that reflects properties at Rinv
18
 As Rinv increases the dP change at the well
becomes ever smaller
q
p
Rinv(t2) Rinv(t1)
pwf(t2) pwf(t2)

The pressure disturbance creates an expanding dP ‘sink’as the

‘pressure front’ moves away from the well with time which is
is reflected in the pressure and dP measured at the well. 19
 Volume of Investigation Vinv

 Rinv with ‘h’ results in Vinv …

‘h’

kh sh
Vinv  R hsh  0.0002637 * 4
2
t
 ct
inv

 Rinv increases proportional to the square root of time

 Vinv increases proportional to time.
 This concept can be expanded to non-radial, non-
homogeneous reservoirs. 20
 Pressure Change vs. Time at the well

5000.

pi
q*pu(t)
4950.
pressure PSI

dP The ‘u’ in pu stands for Unit Rate

4900.

Dimensions psi/stb/d (1/PI)

4850.
4800.

0. 50. 100. 150.

t
100.

+q
rates STB/D
60.
20.
-20.

0. 50. 100. 150.

Time (hours)
The dP at a well caused by a constant rate is characterised by a
Drawdown Type Curve, pu(t). This is the starting point of all PTA
21
 Drawdown Type Curve Analysis
pi  p(t ) B   kt 
1 / PI   pu (t )    f  , x 
q kh   ct L 
2

t = time
k = permeability (output)
h = thickness
L2 = area
x = model parameters (e.g. well orientation)
α,β = conversion factors
 At constant q the Drawdown Type Curve for oil radial flow (Oilfield Units):
q kh
PI  
( pi  pwf )  k 
162.6 B  log t  log  3.23  0.87 S 
  ct rw
2

 Note the additional ‘S’ term- this characterizes additional pressure change
called Skin outside of Darcy’s Law and will be discussed in detail.
22
 Superposition

 Unfortunately, in practice wells are never produced at a

constant rate as this is difficult to achieve. Also, the
pressure data is often noisy during flow.
 Therefore BUs (which are at a constant rate = 0) are
utilized with flows causing the pressure change.
 This requires the rate history q(t) to be expressed as a
piece-wise function, the result being a summation called
the Superposition function.

23
 Superposition

Single Flow :
5100.

pi
4900.
4800.

0. 50. 100. 150.

+q
50.
rates
0.
-100.

0. 50. 100. 150.

tp

24
 Superposition

5100. Single Build Up:

pi
4900.
4800.

0. 50. 100. 150.

50.
rates
0.

-q
-100.

0. 50. 100. 150.

tp

25
 Superposition

5100. Single Flow / Build up Sequence:

Flow + BU
pi
4900.
4800.

0. 50. 100. 150.

+q
50.

=Flow + BU
rates
0.

-q
-100.

0. 50. 100. 150.

tp

26
 Horner Plot
p t 
5000.

pi

p  pt   pt  0
4900.

pt  0
4800.

0. 50. 100. 150.

80.

tp t
40.
-20.

0. 50. 100. 150.

t p  t
2003/03/11-1501 : OIL

9650.
9600.
9550.
• Flow / SI the well
P PSI
• Plot the data in a Horner Plot 9500.
9450.
9400.
9350.

• Find Radial Flow Straight Line 10 2 10 3 10 4

(Tp + dT)/dT
10 5

27
 Horner Plot Straight-line Analysis

B   k  
Identify Radial Flow: pu (t )  162.6 
log( t )  log  
2 
 3.23  0.87 S 
kh   ct w 
r 
Apply Superposition: p(t )  pi  q[ pu (t p  t )  pu (t )]
 t p  t 
qB
Flow / BU: p (t )  pi  162.6 log    0.87 S
kh  t  2003/03/11-1501 : OIL
1980/12/15-1300 : OIL

9650.
Plot and you get the
4980.
famous… 9600.
Pressure during SI

9550.

HORNER PLOT
P PSI
4940.
9500.

qB
NOTE: this plot
9450.

sometimes shown Time

slope m  162.6 increases →
9400.

kh
4900.

Intercept  p *
9350.

10 2 10 3 10 4 10 5
100 101 10 2
(Tp + dT)/dT 103 104
(Tp + dT)/dT
Skin 28
 Superposition: Multi-Flow / Build Up

The Horner Plot is valid for BUs following single-rate flow

however if the flow history is varied Superposition q(t) is
made up of n constant-rate flow-periods which start at time ti
and are less than t:
n
pi  p t    qi  qi 1  pu t  ti 
i 1

This piece-wise rate history q(t) depends on:

 the rate
 the duration of rate
 the time from start time ti of each flow period.
29
 Superposition: Multi-Flow / Build Up

Multi- Rate and Build Up Sequence: 1980/12/19-1200 : OIL

4500.
Pressure, p

3500.
2500.
1500.

0. 50. 100. 150. 200. 250.

q3 q1 q0
q2
400.
Rate, q

qn
-200.

0. 50. 100. 150. 200. 250.

t t t t
Time (hours)

tn t3 t2 t1 t
30
 Superposition Plot Straight-line Analysis

For an arbitrary number of flow periods in the rate history,

with radial flow:
B n
p (t )  pi 162.6  qi  qi 1  log t  tn  ti   0.87 S
kh i 1
1980/12/15-1300 : OIL
2003/03/11-1501 : OIL

This is the semi-log

4980.
9650.
Pressure during SI

9600.
Superposition Plot 9550.
4940.
P PSI
9500.

qB
4900. 9450.

slope m  162.6
9400.

kh 0.
8.0 HR

100.
0.57 HR

200.
0.044 HR

300.

Intercept  p *
9350.

Superposition(T)
2.3 HR 0.60 HR 0.16 HR 0.041 HR 0.011 HR 0.0029 HR
15000. 20000. 25000. 30000. 35000. 40000. 45000.
Superposition(T)

Skin 31
 Enter the Derivative Plot

 The Derivative Plot became possible with high-resolution

gauges that could measure pressure change to a high
degree of accuracy, and personal computers that could
quickly calculate the derivative.
 The Derivative plot is complementary to the Superposition
plot, identifies wellbore and reservoir character that is unique
to this plot
 Most importantly it shows when wellbore storage is over.

32
 What Straight Line?

2003/03/11-1501 : OIL

9650.
9600.
9550.
P PSI
9500.
9450.
9400.
9350.

2.3 HR 0.60 HR 0.16 HR 0.041 HR 0.011 HR 0.0029 HR

15000. 20000. 25000. 30000. 35000. 40000. 45000.
Superposition(T)

The plot DOES NOT tell you which straight line is correct!
Wrong slope m = wrong interpretation!
33
 Derivative Plot Radial Flow
-1
10

Dp and Dp’ proportional to Skin

overlay at
start of
-2
DP & DERIVATIVE (PSI/STB/D)

10

WBS
derivative stabilisation
characteristic = Slope ‘m’ on superposition plot
-3

wellbore storage
10

(WBS)
hump
-4
10

-3 -2 -1 0 1 2
10 10 10 10 10 10

Delta-T (hr)

The Derivative Plot shows character in the data that is difficult to

impossible to recognize in the Horner or Superposition Plot.
34
 Derivative Plot Calculation

Derivative data is calculated as the derivative of pressure with

respect to same Superposition function used for the Horner or
Superposition plot :

For a single-rate drawdown test (no Superposition):

dp dp
(derivative)   t
d ln t  dt

For a Flow and Buildup test (Horner or Superposition):

dp
(derivative) 
d sup( q(t ), t 
35
 Derivative Plot vs. Superposition Plot
Flow & Build up
5000.
4900.
4800.

0. 50. 100. 150.

80.
40.
-20.

0. 50. 100. 150.

9980.
9960.
DP & DERIVATIVE (PSI/STB/D)
10 -1

9920. 9940.
10 -2

P PSI
9900.
10 -3

9880.
9860.
10 -4

2.8 HR 0.25 HR 0.025 HR 0.0025 HR

-3 -2 -1 0 1 2
10 10 10 10 10 10 0. 1000. 2000. 3000. 4000.
Delta-T (hr) Superposition(T)

36
 Derivative vs. ‘Semi-Log’ Plots: Radial Flow

9980.
-1

9960.
10
DP & DERIVATIVE (PSI/STB/D)

9940.
-2

STABIL

P PSI
10

9920.
9900.
dp D
-3

 0.5
10

d lnt D C D 

9880.
9860.
-4

2.8 HR 0.25 HR 0.025 HR 0.0025 HR

10

-3 -2 -1 0 1 2
10 10 10 10 10 10 0. 1000. 2000. 3000. 4000.
Delta-T (hr) Superposition(T)

Time Increases → ← Time Increases

‘Derivative Plot’ (Dp and Dp’) : ‘Horner Plot’ or ‘Superposition Plot’
Log-log scale Semi-log scale
pressure change sensitive pressure sensitive

 The log-log ‘Derivative’ and semi-log ‘Horner or Superposition’ plots

stabilize with infinite-acting radial flow.
 Skin is solved once kh ( the Darcy dP) is determined.
37
 Why does Superposition matter?

4000

3900
pressure, psi

3800

3700

3600

Last rate pre- BU2

3500

0 50 100 150 200 250 300 350 400 450 500

time, hr

• Example of over simplifying rate history

• Remember the derivative is calculated! 38
 Why does Superposition matter?

1.0E+03

1.0E+02
Dp & Dp', psi

1.0E+01

1.0E+00
1.0E-02 1.0E-01 1.0E+00 1.0E+01 1.0E+02 1.0E+03

Dt, hr

The shape of the derivative during BU depends on the accurate

measurement of all preceding rates (but in practice 3x the length
of the BU is usually sufficient)
39
 Analysis Process

A Type Curve is generally divided as follows:

• Well Behavior (aka Early Time) : Wellbore

Storage WBS, Skin related to damage, hydraulic
Time and RI Increases

fractures, partial perforation, well orientation

• Reservoir Properties (aka Middle Time) : kh
• Reservoir Geometry (aka Late Time) :
heterogeneities, fluid contacts, boundaries.

40
 ...when do we see these?

80/12/11-1400 : N/A
Early Time
The pressure response at the wellbore is
pi affected by the well behavior and reservoir
properties, and is in transience.
pressure PSI

Middle Time
Boundaries / heterogeneities start to affect
the pressure but still in transience.
Late Time
0. 50. 100. 150.

All reservoir boundaries are sensed and the

the reservoir is in pseudo-steady state.

+q
rates

Note! The actual duration

0.

the well remains in each

0. 50. 100. 150.
Time (hours) time depends on the
geometry and Rinv of the
system.
Time and RI Increases
41
 Early and Middle Time Effects
 p and PI controlled by kh (Darcy) and Skin.
 Skin change causes Dp -> Dp’ distance during radial flow to change.
 kh change causes Dp->Dp’ to move up or down.

If the only interest is whether kh or Skin has changed, short BUs

that reach radial flow are all that’s needed.
-1

-1
10

10
-2

-2
10

10
-3
-3

10
10

10 -2
10 -1
10 0
10 1
10 2 10 -2 10 -1 10 0 10 1 10 2
Delta-T (hr)

42
 Late Time Effects
 p and PI controlled by Reservoir Geometry (barriers / heterogeneity).
After radial flow derivative moves up or down.
 A long BU is useful but more useful is two BUs as often flow regimes
display the same derivative character but different pressure- the pressure
constrains the inverse problem- this is discussed in Test Design.

pi

k = 500 md, S=10

k = 500 md, S=10,

geometry

pgeo

Time (hours)
43
 What is Pressure Transient Analysis?

I S O
Input Signal Unknown Reservoir Output Signal
(Rate) (Pressure)

• I and O are known: Find S

MODEL IDENTIFICATION
INVERSE PROBLEM, NON-UNIQUE SOLUTION

• I and S are known: Find O

MODEL VERIFICATION, TEST DESIGN
DIRECT PROBLEM, UNIQUE SOLUTION

44
 Analysis Process

Step 1: MODEL IDENTIFICATION

Find a model S’ which behaves in the same way as S (i.e.
the true system); O’ has the same shape as O

Model identification is a pattern recognition problem:

• Given the measured pressure / rate data, and knowing
characteristic shapes created by well-defined flow
regimes and well geometries, identify those that could
create this type of test data.

45
 Wellbore Storage
Pressure, p
Rate, q

q surface
q sand-face

Time, t
The compressible volume of the wellbore cwVw causes a lag between the
surface and sandface rate
WBS is estimated to have finished when the sandface rate is < 1% of the
surface rate. Before that WBS masks all reservoir information in the data.
46
 Wellbore Storage

Dp and Dp’ both show the Unit Slope and WBS ‘hump’ whereas there is
nothing characteristic in the Superposition Plot.
End-WBS and calculates a wellbore volume coefficient C.

V
C  cwVw endWBS
p -1
10

A back-check of C can
DP & DERIVATIVE (PSI/STB/D)

be done with known Vw

-2
10

and cw
-3
10

dp D
1
d ln t D C D 
-4
10

-3 -2 -1 0 1 2
10 10 10 10 10 10
Delta-T (hr)

47
 Wellbore Storage

• NOTE: There is NO information in the WBS period that can

be analysed, however a change in WBS over time can
indicate a changing wellbore or fluid condition.
• Time end WBS can be estimated as follows:

 200,000  12,000 s C 
t wbs   
 kh /  

• Large WBS masks the reservoir behavior out to

the RI at time end- WBS.

48
 Changing Well Bore Storage

Pressure Change, p, and Derivative ( psi)

104 104

e
lin
PRESSURE
e
in
-1

PRESSURE

ht
-1
tl
10

ig
103
gh

10

ra
103
ai

st
DP & DERIVATIVE (PSI/STB/D)

tr

DP & DERIVATIVE (PSI/STB/D)

pe
s

slo
pe
slo

t
ni
t

U
-2

ni

-2
U
10

102

10
102

01
01

0.
04

1=
.
0.

04 =0
2=

,C
-3

=0 DERIVATIVE DERIVATIVE
,C

-3
10

C 2

IL
10
S,

10
C 1 10
IL

O
S, A
O

A G
G &
& IL
IL O
-4

-4
1 10 O
10

1 10 -3

10
-3 -2 -1 0 1 2
10 10 10 10 10 -3 -2 -1 0 1 2
10 -2 10 -1 10 10 10 2
10-3 10-2 10-1 1
Delta-T (hr)
10 102 10 10 10 Delta-T1(hr) 10 10
Elapsed time, t ( hours) Elapsed time, t ( hours)
• Wellbore below bubble point: • Wellbore below dew point:
C coefficient is gas that C coefficient is oil
transitions to oil during the (condensate) that
BU. This produces a steep transitions to oil during the
derivative increase that can BU as the condensate
be > 1 at the end of the gas revaporizes.
dominated response.
49
 Radial Flow

ENDWBS SLOPE ENDWBS

9980.
9960.
DP & DERIVATIVE (PSI/STB/D)
-1
10

9940.
P PSI
9920.
-2

9900.
10

STABIL
UNIT SLP

9880.
9860.
-3

3.2 HR 0.29 HR 0.029 HR

10

10 -3
10 -2
10 -1
10 0
10 1 0. 1000. 2000. 3000.

Delta-T (hr) Superposition (T)

Homogeneous Reservoir Homogeneous Reservoir

Static-Data and Constants slope of the line = -0.01714 PSI/cycle Static-Data and Constants
well. storage = 0.10392 BBLS/PSI Volume-Factor = 1.000 vol/vol extrapolated pressure = 10000.2 PSI Volume-Factor = 1.000 vol/vol
R(inv) at 22.46 hr = 991. METER Thickness = 3.048 METER
skin = -0.43215 Thickness = 3.048 METER Viscosity = 1.000 CP
permeability = 948.66 MD Viscosity = 1.000 CP prod. time=30.00 hr at rate=1000.000 STB/D Total Compress = .8250E-05 1/PSI
Perm-Thickness = 2891.6 MD-METER Total Compress = .8250E-05 1/PSI Rate = 1000. STB/D
P-extrap. = 10000.2 PSI Rate = 1000. STB/D skin = -0.43354 Storivity = .6287E-05 METER/PSI
R(inv) at 22.28 hr = 987. METER permeability = 948.66 MD Diffusivity = 11270. METER^2/HR
Perm-Thickness = 2891.6 MD-METER

 Derivative Plot for Radial Flow: two characteristic straight lines- Unit
Slope for WBS and STABIL (0 slope).
 Semi-Log Superposition Plot: one characteristic straight line- Slope ‘m’
50
 Using the Derivative Plot
y-axis

CDe2S

tD kh t 0.8936C x-axis
 0.000295 CD 
CD  C  ct hrw2
51
 Pressure Match PM pD/p =2/100
Time Match TM t/(tD/CD) = 0.01/.015

CDe2S
Match

1000

100

10

1 .
0.001 0.01 0.1 1 10 100
Dt hours

Matches for pD/p, t/(tD/CD) are made by overlay, and the groups are
then dimensionalized using ct, q, µ, Ø, h, rw. Note pD stabilizes at 0.5.
The equations are then solved for kh and CDe2S for Skin. 52
 Analysis Process

Step 2: MODEL VERIFICATION

Verify the consistency of the interpretation model

• I and S are known: Find O

MODEL VERIFICATION, TEST DESIGN
DIRECT PROBLEM, UNIQUE SOLUTION

Model Verification ensures that the interpretation model is

consistent with the known data.

53
 Analysis Process: Model Verification

9980.
9940.
pressure PSI
DP & DERIVATIVE (PSI/STB/D)
-1
10

9900.
9860.
0. 20. 40. 60. 80.
-2

-200. 200. 600. 1000.

10

rates STB/D
-3
10

-3 -2 -1 0 1 0. 20. 40. 60. 80.

10 10 10 10 10
Time (hours)
Delta-T (hr)

 Type Curve simulation

 Output O’ has the same shape as actual data O in the
derivative, data and semi-log plots: good solution

54
Constraining and Managing the Inverse Problem
• List the flow regimes that have characteristic behaviour:
e.g. WBS -> radial flow derivative plot stabilisation after WBS

• Identify these flow regimes in the measured data to identify

a well / reservoir system:
e.g. radial flow in a vertical well suggests homogeneous reservoir

• Predict the response given known facts about the well /

reservoir system:
e.g. a limited entry well is expected to have 2 derivative stabilisations

Compare with other knowns (geo, petrophysics, core,

etc.) and don’t forget common sense.
Finally remember you can’t interpret what you can’t see.

55