W E L L S E R V I C E S

Cement Evaluation

Mickaël Allouche
Project Manager
SRPC - France
Cement Evaluation Course
1 - May 2003
 Cement Job Evaluation Course
• Acoustic log method and interpretation
 Sonic: CBL
 Ultra Sonic: USIT

Cement Evaluation Course

2 - May 2003
 Basic Sonic Principle
Basic Tool Principle

T
– A Transmitter fires an acoustic signal in
all directions
– Surrounding Media Resonates
3’
R
– Receivers record resulting sound
– Sound wave is Analyzed 5’
R

Cement Evaluation Course

3 - May 2003
 CBL Measurement Principle
Basic Tool Configuration
Tx
• 1 Transmitter – 2 Receivers
 3 ft Receiver for CBL Measurement
3 ft  5 ft Receiver for VDL Analysis

R3 • TOOL MUST BE CENTRALIZED

5 ft

CBL: CEMENT BOND LOG

R5 VDL: VARIABLE DENSITY LOG

Cement Evaluation Course

4 - May 2003
 Basic CBL Principle
Similar to a
Ringing Bell
When Fluid is behind
No
Cement Casing, pipe is free to
vibrate [ loud sound ]

When the casing is bonded

to hard cement, casing
Good vibrations are attenuated
Bond
proportionally to bonded
surface
Cement Evaluation Course
5 - May 2003
 Sonic fundamentals: 3 ft measurements
Casing
CBL Time
Window E3
T0 Minimum
Amplitude
Detection E1
Level

E2 Time

TT

NMSG CBLG

Part of the wave front, refracted straight down the casing, is

used to determine Amplitude and Transit time.
Cement Evaluation Course
6 - May 2003
 Sonic fundamentals - VDL
• Variable Density Log is the full Transmitter
firing
Casing Formation Mud

wave display of the 5-ft receiver.

• Displayed as light and dark stripes.

Amplitude (mv)
• Contrast depends on positive or time usec

negative amplitude.
• To allow easy differentiation
between casing and formation
signal.

Cement Evaluation Course

7 - May 2003
 VDL Algorythm Principle

• Recorded Waveform at one depth

• Waveform is cut for only Positive Peaks

• Peaks are compared to a Grey Scale

• Peaks are shaded and presented from

Top View

• Final Picture Vs Depth is obtained

Cement Evaluation Course

8 - May 2003
 Sonic fundamentals - VDL
• Unless casing is fully eccentered the presence of formation
arrivals is:
 A qualitative indicator of the presence of a solid material behind
the casing
 By no means a quantitative indicator of its presence

Cement Evaluation Course

9 - May 2003
 Example of a good CBL
• Check quality
 Look at TT curve
• Check CBL curve
 Relatively low amplitude
• Verify VDL
 No casing arrivals
 Formation arrivals

Example of a SALTBOND slurry 7-in. liner

Cement Evaluation Course
10 - May 2003
 Example of a “bad” CBL
• Check quality
 Look at TT curve
• Check CBL curve
 Relatively high amplitude
• Verify VDL
 Casing arrivals
 Weak formation arrivals
• Isolation???

Example from a 9 5/8-in. casing

Cement Evaluation Course
11 - May 2003
 Basic CBL Amplitude Interpretation
• Basic interpretation:
 Low measured amplitude: good cement
 High measured amplitude: no cement
• Drawback: too simplistic (examples of 100% bond amplitudes)

Casing/Cement 3 MRayl 6 MRayl

5 ½ in. 17 lb/ft 6.1 mV 1.0 mV
9 5/8 in. 47 lb/ft 12.2 mV 3.3 mV
Logging fluid: 9.0 lb/gal water base mud

Cement Evaluation Course

12 - May 2003
 Bond Index

Cement Evaluation Course

13 - May 2003
 Refined CBL Amplitude Interpretation
• Determine the bond index from:

log [ E Measured ]  log[ E Freepipe ]

BI 
log [ E 100 % ]  log[ E Freepipe ]
 The relative (not the absolute) amplitude value is the key
• Interpretation:
 In absence of any artefact the bond index is linearly related to the
percentage of the casing surface bonded with the cement
 Backed up by experiments with artificial channels filling part of the
annulus

Cement Evaluation Course

14 - May 2003
 Other definition of the bond index
• Determine the bond index from (Huawen Gai- BP - SPE 23729):

E Freepipe  E Measured
BI 
E Freepipe  E 100 %
 Based on a parallel path model (cement/free pipe) vs. a series path
model
 Assumes the energy following one path does not leak into the other which is
invalidated by experimental results
 Truth between between the two models but closer to series path model
 Gives higher values of the bond index

Cement Evaluation Course

15 - May 2003
 Refined CBL Interpretation (cont’d)
• Rule of thumb:
 BI > 80% across a minimum length means hydraulic isolation
 But a bond index of 80% means that 80% of the casing surface
is bonded with cement
 If this is due to a channel this is quite a large channel (20% or 72
degrees!)
 But in absence of ultrasonic tool response this is considered as
satisfactory!
 What if BI < 80%?

Cement Evaluation Course

16 - May 2003
 How to determine the bond index?
• What is needed?
 The free pipe and measured amplitudes: E freepipe and Emeasured
 The expected amplitude for 100% bond: E 100%
• Free pipe amplitude:
• Depends on casing size and logging fluid
 Can be measured (free pipe section) or can be derived from data
bank/model
 Wireline data
 SoniCalc software
[Link]

Cement Evaluation Course

17 - May 2003 *Mark of Schlumberger
 How to determine the bond index (cont’d)?
• 100% bond amplitude:
 Depends on casing size/thickness and logging fluid
 Depends on cement acoustic impedance* NOT compressive
strength (material behind casing has to be solid)
 Can be derived from data bank/model
 Wireline data (quite simplistic - still based on compressive strength)
 SoniCalc software

* This is not 100% correct but this is a very good approximation

Cement Evaluation Course

18 - May 2003
 Free Pipe Amplitude
5 • If no Casing-Cement bond, amplitude
is not attenuated

• This is called

3 FREE PIPE AMPLITUDE

2
CBL: Free Pipe

Cement Evaluation Course

19 - May 2003
 CBL AMPLITUDE VS. CASING SIZE

Cement Evaluation Course

20 - May 2003
 CBL-VDL Fluid Effects

Cement Evaluation Course

21 - May 2003
 FREE PIPE CHECK
CBL

100
Interpretation
100

Perfect
Chevron Patterns
Depth Match

Chevron Patterns

TT and CBL Amplitude

as espected according to Casing Size
Cement Evaluation Course
22 - May 2003
 Cement to Casing Bond
5
• If casing is well bonded,
soundwave will be attenuated

• The received CBL amplitude will be

low
3

2
CBL: Free Pipe

CBL: Good Bond

Cement Evaluation Course

23 - May 2003
 CBL chart

5.5 MRayl
2 mV
100% bond

CBL / VDL
3.5 MRayl
12 mV
100% bond

Cement Evaluation Course

24 - May 2003
 CBL attenuation and compressive strength
• Using the conventional chart with compressive strength
scale makes no sense:
 Different charts have to be used for different cement systems
e.g. foamed cement vs. neat cement
 CBL attenuation is related to acoustic impedance and not
compressive strength
 In the lab compressive strength is quite often indirectly derived
from compressional wave velocity measurement - UCA - using
empirical laws which depends on cement density!

Cement Evaluation Course

25 - May 2003
 Acoustic Impedance
Materials
8
• Acoustic tools respond
Heavy
to acoustic impedance
6
(acoustic hardness) Z Z
Setting Neat
MRayl
• Z = density x acoustic slurry Cement

velocity 4
Light
• Z is expressed in Heavy
mud

MRayl (106 kg.m-2.s-1) 2

Water Liquid

Oil

0 Gas

Cement Evaluation Course

26 - May 2003
 SoniCalc software

[Link]

Cement Evaluation Course

27 - May 2003
 Refined CBL Amplitude Interpretation
• SoniCalc software predicts:
 Cement acoustic impedance and expected attenuation
 Expected amplitudes for 100% and 80% BI
• SoniCalc software Interpretation Guide:
 If E100% < EMeasured < E80%: OK
 If EMeasured > E80%: ???
• Key in the interpretation:
 Compare the measured amplitude with a REALISTIC expected
amplitude not an absolute amplitude.
Cement Evaluation Course
28 - May 2003
 SoniCalc software for CBLF prediction
• A CBL log can display CBLF instead of CBL the
measured amplitude
• CBLF is the fluid compensated amplitude
 It reads as if the logging fluid was water
 CBLF = FCF x CBL
 FCF is derived from data bank (Wireline)
• If CBLF is displayed on the log, logging fluid entered in SoniCalc
should be water at 8.32 lb/gal

Cement Evaluation Course

29 - May 2003
 LiteCRETE in 9 5/8-in. 47# casing

CBLF FP: 55 mV
SoniCalc
100%BI: 5-6mV
CBLF min.: 3-5 mV

Cement Evaluation Course

30 - May 2003
 Deeper section

CBL = 35 mV
BI=19% (40%)

Cement Evaluation Course

31 - May 2003
 CBL interpretation problems
• In many cases CBL amplitude is much higher than
expected even though cement is known to be present
 Design and execution of cement job are known to be OK
 VDL showing formation arrivals
 Attempt to squeeze failed
• In many cases calculated BI do not make sense: e.g. 30%!

CBL amplitude is affected by many factors

other than channelling
Cement Evaluation Course
32 - May 2003
 CBL Amplitude interpretation flaws
• Measured amplitude is not due to first peak of casing signal
• Tool is not measuring what it is supposed to do: QC
• Free pipe is not constant vs. depth
• Transducer response is not consistent vs. depth
• 100% bond amplitude is different from expected
• Microannulus effect(s)

Cement Evaluation Course

33 - May 2003
 Stretching
E1 Free Pipe Signal
Good Bond Signal
T0 Threshold

TT TT’

T
In cases of Good Cement
E1 decreases and TT is detected on a non linear portion of E1

T STRETCHING is the TT increase from its value in free pipe

Cement Evaluation Course

34 - May 2003
 TT Cycle Skipping
E1 E3

T0 Threshold

TT TT’

In cases of very Good Cement E2

E1 could not reach Detection Threshold Level

T T skips to 3rd Peak [E3 ]........this is known as CYCLE SKIPPING

Cement Evaluation Course

35 - May 2003
 CBL physical limitations
• Fast formations: may give higher amplitude than expected
 Look for correlation in transit time/amplitude change
 Look for correlation amplitude/VDL/Gamma ray
 Check Sonic log
• Concentric casings: may give higher amplitude than expected
 Determine time lag between casing and previous casing arrivals
 Look for VDL: frequency change
 Look for unusual transit time skips
In both cases BI in meaningless

Cement Evaluation Course

36 - May 2003
 Fast Formation
5 Fast Formation Arrivals
In cases of good cement and

formation slowness < steel slowness

3 formation arrival arrives first

T Dolomite = 43.5 sec/ft

T Limestone = 47.5 sec/ft
2 T Anhydrate = 50.0sec/ft

The transit time and CBL amplitude

T will be affected

Cement Evaluation Course

37 - May 2003
 Fast Formation
 In the presence of FF arrivals no CBL evaluation is
5
possible, since E1 is due to Formation (Fast)
arrivals and not from Casing arrivals (usually with
3 to 5 ft receivers)

3  FF arrivals travel longer distance from Tx to RX

(casing thickness + cement thickness + formation)
than casing arrivals.
2

 Only in short spacing Tx-Rx (~ 1 ft) the casing

arrival will arrive earlier than FF arrival.
T  Tools able to measure CBL in FF are: CBT, CMT
and SSLT (shortes Tx-Rx about 1 ft
Cement Evaluation Course
38 - May 2003
 FAST FORMATION
CBL

Interpretation

Transit Time
High
Shorter than
<----------------------------------------CBL Amplitude
Casing arrivals
on areas of

fast formation

<---------------------------------------- arrivals

Cement Evaluation Course

39 - May 2003
 Fast Formations
• High(er) amplitude not
representative of cement
 TT, TTSL
 CBL, CBSL

• Correlation with GR

• Confirm with sonic VDL fingerprint

• Must have some good cement in
annulus (how much?)

Cement Evaluation Course

40 - May 2003
 Concentric casings
• Distance between peaks on
VDL (frequency)
• CBL (amplitude) measures
interferences and not
cement quality
• Must have good cement in
annulus (how much?)
• SPE 18028
 use narrow gate

Example from a 13 3/8-in. casing

Cement Evaluation Course
41 - May 2003
 CBL Quality Control
• Is the tool calibrated?
• Does the tool response repeat itself (Repeat section)?
• Are the CBL gate parameters set properly?
• Is the measured transit time as expected?
• Is the tool properly centered?
• Is the free pipe measured value as expected?
• How does the lowest measured amplitude of the log compare to the
predicted amplitude of SoniCalc software?

Cement Evaluation Course

42 - May 2003
 CBL Quality Control (cont’d)
• Is the tool calibrated?
 Usually not a problem: done on a regular basis in the shop

• Does the tool response repeat itself (Repeat section)?

 Obvious as with any other log

• Are the CBL gate parameters set properly (fixed gate)?

 CBLG: around 35 microsec (or 45).
 NMSG: Expected TT (or Free pipe TT) - 10 microsec. (or 20)
 Expected TT = 57 x 3 + Logging fluid slowness x Casing ID / 12
Gate parameter settings not valid for the DSLT

Cement Evaluation Course

43 - May 2003
 CBL Quality Control (cont’d)
• Floating gate mode (CBL->CBSL and TT->TTSL)
 The gate follows TT and is characterised by:
 SGW (sliding gate width)/AMSG: gate opens at (TT-SGW) but no before AMSG (amplitude minim
sliding gate)
 ASGL (auxiliary minimum sliding gate): gate closes at (TT-SGW+ASGL)
 Useful for checking fixed gate, and in case of change in logging fluid properties, change in casing
thickness, fast formations, etc ...

• Are the CBL gate parameters set properly (floating gate)?

 SGW/AMSG: around 80 microsec. / around 150 microsec.
 ASGL: around 100 microsec.
Gate parameter settings not valid for the DSLT

Cement Evaluation Course

44 - May 2003
 CBL Quality Control (cont’d)
• Is the measured transit time as expected?
 Measured TT < Expected TT: tool eccentering or fast formation (BI
meaningless)
 Measured TT > Expected TT: TT stretch or skip (good qualitative sign if
associated with low amplitude), concentric casing (BI meaningless)

• Is the tool properly centered?

 Measured TT shorter decreasing by more than 4 microsec. while amplitude is
decreasing shows that the tool is off centered
 Eccentered tool will overestimate BI

Cement Evaluation Course

45 - May 2003
 Tool Eccentering
Causes for Eccentralization
5 • Improper Equipment selection

[ Centralizers ] for Casing Size

• Missing or Broken Centralizer(s)

3 • Weak Centralizers in deviated wells

• Tool Damaged and/or bent

2
• Damaged Casing
Consequences
• Unbalanced sound paths
T
• Resulting waveform is meaningless

Cement Evaluation Course

46 - May 2003
 Eccentering Analysis
Short Path Waveform
Resulting Waveform Normal Waveform

T0 Threshold

TT

Delayed Waveform
If the tool is eccentered
There will be destructive interference from different sound paths

Waveform from close tool side to casing

Result is a Bad Log
Waveform from far tool side to casing not recoverable
Resulting waveform has Dramatic lower amplitude in Playback
Resembling a zone of Good Cement but with shorter Transit Time [≈ 4 s less]
Cement Evaluation Course
47 - May 2003
 CBL Quality Control (cont’d)
• Is the free pipe measured value as expected?
 First make sure that free pipe is indeed free pipe (not microannulus for example)
 What is displayed? CBL vs. CBLF (is the FCF factor OK?)
 If significantly different used measured value to determine BI unless tool is off
centered

• How does the lowest measured amplitude of the log compare to the
predicted amplitude of SoniCalc software?
 If significantly different used lowest measured value to determine BI unless tool
is off centered

Cement Evaluation Course

48 - May 2003
 Free pipe value varies vs. depth
• BI depends on free pipe amplitude
• Free pipe amplitude depends on logging fluid properties
• Logging fluid properties vary vs. depth
• Consequences:
 Limited as in most cases the impact of an error on free pipe
amplitude on BI is low.
 Can be important if attenuation is low due to low acoustic
impedance of material behind casing
• Effect accounted for with the CBT
Cement Evaluation Course
49 - May 2003
 Transducer response vs. depth
• Transducer response is affected by both temperature and
pressure
 SPE 13044 by Nayfeh et al.
• Effect accounted for only with the CBT

Cement Evaluation Course

50 - May 2003
 100% bond amplitude not as expected
• What if the material behind the casing does not have the
expected acoustic impedance?
 Cement not mixed as per design: density control, additive concentration?
 Cement system unstable: free water, sedimentation
 Poor fluid loss control: annular bridging
 Cement can get contaminated: spacer, mud, formation fluid, or even
with another cement (tail slurry sinking/swapping with lead slurry)
 Cement can develop its strength at a slower pace than expected
 Contamination and/or overestimation of temperature and/or slower strength
development at top vs. bottom of cement column

Cement Evaluation Course

51 - May 2003
 100% bond amplitude not as expected
• What if the material behind the casing does not have the
expected acoustic impedance?
 Cement acoustic impedance underestimated:
 BI can possibly be higher than 100%
 Cement acoustic impedance overestimated:
 BI lower than 100%
 But material behind casing may be uniformly distributed around the casing
and bonded to it therefore hydraulic isolation may be achieved.

Cement Evaluation Course

52 - May 2003
 Microannulus effect(s) on the CBL
• When the shear coupling casing/cement is lost:
 There is less energy lost in the surrounding medium
 Amplitude goes up
 Job looks worse than what it is
• Due to the presence of a fluid at the casing/cement interface
caused by (wet microannulus)
 Incomplete mud removal (mud/spacer film)
 Change in downhole stresses after cement job (pressure or
temperature decrease, further drilling)

Cement Evaluation Course

53 - May 2003
 Microannulus effect(s) on the CBL
• Due to the presence of a “gap” at the casing/cement interface
caused by (dry microannulus)
 Change in downhole stresses after cement job (pressure or
temperature decrease, further drilling)
 Cement can be either partially bonded or even totally debonded

Cement Evaluation Course

54 - May 2003
 Microannulus effect(s) on the CBL
• Most of the time
 Amplitude varies a lot vs. depth
 Formation arrivals can be seen on VDL
 Microannulus is though not to be continuous (cement partially
bonded)
• In some cases
 CBL nearly free pipe / formation arrivals barely visible on VDL
 Fluid film is thought to be continuous (mud/spacer film or “air” gap)
• Microannulus effect explains the meaningless BI values

Cement Evaluation Course

55 - May 2003
 Dry microannulus case 1: CBL 0 psi

Cement Evaluation Course

56 - May 2003
 Dry microannulus case 1: CBL 1000 psi

Cement Evaluation Course

57 - May 2003
 CBL and the cement/casing interface
• The CBL is strongly affected if a fluid is present at the
cement/casing interface almost regardless of it thickness because
the shear coupling is lost
 Numerical simulation show that between a few tens of microns and a few
millimeters CBL amplitude is of the order of 85% of free pipe amplitude
(laboratory experiments - SPE 25377 - give a figure of the order of 40% but
results are questionable). For larger gaps CBL amplitude is equal to free pipe.
 Field examples show that CBL amplitude can be nearly free pipe with an
average acoustic impedance lower than expected but uniformly distributed on
the USIT maps.
 For such cases CBL amplitude does not always decrease when pressure inside
the casing is increased (cannot fully close the gap?)

Cement Evaluation Course

65 - May 2003
 CBL and the cement/casing interface
• The CBL is may or may not be affected if a gas is present at
the cement/casing interface
 When the gap is very small or rather when it is not continuous along the
casing the CBL amplitude the CBL is barely affected (USI reads lower than
expected)
 As the size of the gap increases or as it becomes more continuous along the
casing the CBL becomes more and more affected (USI map show
microdebonded cement with more and more gas) until cement is fully
debonded from the casing in which case CBL reads free pipe - or even
slightly more as described in SPE 25377 (USI map becomes totally red).
 In the latter case CBL amplitude is usually sensitive to pressure inside casing

Cement Evaluation Course

66 - May 2003
 CBL Qualitative Interpretation
CONDITION TRANSIT CBL VDL
TIME AMPLITUDE
Free Pipe NORMAL HIGH Casing Arrivals
Usually No Formation Arrivals
Good Bond to Casing & Formation HIGH / NOISY LOW No Casing Arrivals
Formation Arrivals
Good Bond to Casing HIGH LOW No Casing Arrivals
Not to Formation CAN BE No Formation Arrivals
NOISY
Poor Bond to Casing NORMAL MEDIUM Strong Casing Arrivals
No Formation Arrivals
Microannulus NORMAL MEDIUM Formation Arrivals
Casing Arrivals
Channeling NORMAL MEDIUM Formation Arrivals
Casing Arrivals
Fast Formations LOW HIGH Formation Arrivals
No Casing Arrivals

Eccentered Tool LOW LOW DEPENDS

Cement Evaluation Course

67 - May 2003
 CBL global limitations
• A lot of effects tend to increase the amplitude:
 Material with acoustic impedance lower than expected
 Fluid channel
 Microannulus
• So in fact quantitative evaluation from CBL amplitude is
nearly impossible without
 Using VDL information AND analysing cement job data
 On top of this of course USIT data when available lead to an even
better evaluation
Cement Evaluation Course
68 - May 2003
 CBL global limitations
• CBL limitations amplified when dealing with low acoustic
impedance cement
 Attenuation lower that with conventional weight cement (15.8
lb/gal)
 Even more of a problem with large casing sizes which gives also a
lower attenuation

Cement Evaluation Course

69 - May 2003
 CBL pros and cons
• Pros
 Provides information on the cement/formation bond
 Works in a wide range of logging fluid density
 Can be less affected by the presence of a dry microannulus at the
cement/casing formation (gas) than the USIT.
 Detects top of cement in some cases where USIT cannot.
• Cons
 No azimuthal resolution; poor resolution in large casing sizes
 Affected by fast formations and concentric casings
 Affected by the presence of a wet microannulus at the cement/casing
interface (liquid or unset material)
Cement Evaluation Course
70 - May 2003
 CBL Common Interpretation Mistakes
• Expecting a low amplitude systematically
 Low acoustic impedance cements can give a relatively high amplitude
even when 100% bonded to the pipe
• Assuming hydraulic isolation is not achieved because
amplitude is higher than expected:
 Tool calibration?
 Fast formation or concentric casings?
 Cement weaker than expected?
 Microannulus?

Cement Evaluation Course

71 - May 2003
 CBL Common Interpretation Mistakes
• Assuming hydraulic isolation is achieved because VDL shows
formation arrivals
 A channel may still be present
• Assuming hydraulic isolation is achieved because BI is larger
than 80%
 In rare cases a channel may still be present

Too many problems with the CBL why not using it

in combination with the USIT?

Cement Evaluation Course

72 - May 2003