Analytical and Interpretive Problems of Production Logging

LYNN D. JONES* CARDINAL SURVEYS, I N C .

MEMBER AlME MIDLAND, TEX.

Abstract thorough analysis will minimize unnecessary remedial

work to both injection wells and producing wells which
Analysis o f down-hole fluid movement surveys cannot often results from cook book methods of interpretation.
be cataloged. Although there are a f e w so-called "cook
book" methods designed to provide an easy, rapid analysis
for some surveys, these should be used sparingly. W h e n Flowmeter Surveys
standardized interpretations are applied to fluid movement This discussion on flowmeters is limited to the packer-
surveys with total disregard t o the seemingly infinite n u m - type tool since analysis of the continuous flowmeter or
ber o f variables affecting analysis, misleading results can spinner survey is generally recognized as being highly in-
and d o occur. C o o k book interpretations can cause erron- terpretive. The packer flowmeter is a very accurate tool
eous r e ~ u l twith
~ the flowmeter or spinner surveys, easiest for down-hole fluid measurement. Its major limitation is
of all the down-hole fluid movement surveys t o interpret. that it can measure fluid in only the pipe or borehole.
T h e v a r i o u ~ types o f tracer and temperature surveys The flowmeter can provide neither information about the
should never be analyzed with any sort o f standardized distribution of the fluid after it leaves the borehole in in-
interpretation. Some o f the more c o m m o n problems re- jection wells nor information concerning the location of
sulting from standardized interpretive procedures are pre- the flow channels prior to entering the borehole in pro-
sented in this article. Possible results o f misinterpretation ducing wells. A unique problem involving profiling open-
are presented t o provide a background for more effective hole injection wells with a flowmeter was discussed by
analysis of down-hole fluid movement surveys. Jones.' In this case, natural vertical fractures in the for-
mation precluded an accurate profile by the flowmeter.
Introduction Fluid flow through the spinner section of the tool created
enough friction pressure to cause the fluid to bypass the
Production logging, the analysis of down-hole fluid flowmeter through the vertical fractures. The reduction in
movement, generally requires more interpretation than measured rate through the flowmeter indicated that a zone
primary logging before satisfactory results are reached. above it was taking fluid, when actually the water was re-
Some standard interpretation procedures have evolved in ' entering the borehole below the flowmeter and going on
an effort to simplify the analysis of fluid movement logs. down the hole.
As long as these cook book procedures are recognized as Although the problems of obtaining accurate production
being simply a basis for initiating the complete interpre- profiles are less numerous and severe, one problem war-
tation, then valid, usable results will be obtained. If a rants consideration. For optimum results, flowmeter pro-
standardized form is used for the complete interpretation, duction profiles require much more down-hole information
the analysis of that particular production log is only on intermittent gas-lift wells than on wells utilizing con-
partially complete, and the results may not indicate what tinuous gas-lift systems. The producing rate during each
is actually happening. cycle of intermittent systems is considerably greater than
Many things affect analysis of down-hole fluid move- the PI of the well. Consequently, an instantaneous down-
ment. A few down-hole variables include hole size, lami- hole rate obtained by a flowmeter would be totally inac-
nar flow, counter or crossflow, tool efficiencies, logging curate since the well would be pumping off.
speed and even, in some cases, improper running of the Trace A on Fig. 1 is a 15-second flowmeter recording at
survey itself. Although all of these factors do not always a point above all producing intervals, indicating a producing
affect the analysis of all of the various types of production rate of 480 BJD. Since this is an intermittent gas-lift well
logs, their effect should not be indiscriminately neglected. and does not produce at a steady rate, a 15-second recording
This article is not intended to imply that a more thor- represents little more than an instantaneous rate. Trace B
ough analysis of the down-hole fluid movement survey presents the flowmeter recording during the entire 6%-
will give the complete picture every time. However, more minute lifting period. The average producing rate for Trace
Original manuscript received i n Society of Petroleum Engineers office B is 288 B/D, but the picture is not complete until the entire
March 13, 1967. Revised manuscript received July 3, 1967. Paper ( S P E gas-lift cycle is monitored by the flowmeter. Trace C pre-
1781) was resented a t S P E 6th Permian Basin Oil Recovery Confer-
ence
.--. in Midland. Tex.. Mav 8-9. 1967. O C o a y r i ~ h t1967 American
. he!d --- sents the entire 10-minute cycle. During the last 3% min-
Institute of Mining, ' ~ e t a l l u r g i ~ a sad
l, Petraeu& Engineers, Inc.
,%Presentlywith Texas Pacific Oil Go., Midland, Tex. 'References given a t end of paper.

AUGUST, 1967 Reprinted f r o m A u ~ u s t , 1967, Issue of JOURNAL OF PETROLEUM TECHNOLOGY 993

Utes of the cycle, flow is downward- the result of the above 0.075 ft/sec.' However, velocity measurements alone
fluid in the tubing seeking the lifting fluid level. Over the do not result in a complete survey.
entire cycle, fluid was rising a t an average rate of 288 Logging or tracing runs should always s~ipplementthe
B / D for 6% minutes and falling at an average rate of 263 velocity measurements. These measurements generally pro-
B/D for 3% minutes for a net surface producing rate of vide the quantitative portion of the survey while the log-
109 B/D. The average daily producing rate from this well ging runs provide only the qualitative portion of the sur-
prior to the survey was 100 B/D. vey. Attempts have been made to adopt procedures where-
This example used a procedure which could be classi- by quantitative results could be obtained by the logging
fied as standardized. The actual mathematics involved in runs. Two of these methods have been used rather ex-
calculating the producing rate is standardized, but the way tensively, one utilizing an insoluble radioactive isotope
the information was obtained is not. The entire cycle (matched density beads) and the other utilizing a soluble
shown in Fig. 1 was only one of many obtained on the isotope (Self method). A standardized procedure is used
well. Sufficient cycles should be monitored a t different to interpret both methods, and under ideal conditions both
depth levels to insure a stabilized producing rate. will give fairly reliable results. Since ideal conditions rarely
exist, the limitations of these methods should be recog-
Tracer Surveys nized.
The matched density beads method is based on the prin-
Several methods and tools can be categorized under ciple that the beads will plate out on the formation face
tracer surveys, all of which are to provide ultimately an in the same proportion that the interval is taking fluid.
injection or production profile. Johnson and Morris' pre- The resultant radioactive buildup as seen by the gamma
sented various methods of obtaining velocity surveys (con- ray trace represents the profile. Quantitative values are
trolled interval log, controlled time log and tracer-ejector assigned to the various intervals after planimetering the
velocities). Of these three, only the tracer-ejector lends increase in radioactive intensity above base log. Connolly'
itself to a standard analytical procedure with minimum showed how fractures in the formation greatly affect in-
possible interpretive error. A high degree of accuracy terpretation by this procedure.
is maintained when making velocity measurements with a The Self method is a proposed procedure for obtaining
centralized tracer-ejector tool as long as the velocity is quantitative results in shot holes. The intensity of gamma
ray responses to a slug of soluble radioactive isotope mov-
ing down the hole is compared through a series of logging
or tracing runs. The sum, in inches, of the height and
base of the gamma ray response on the logging paper for
each tracing run is the basis of comparison. The sum of
the height and base of the 100 percent rate (established
well above the first possible zone of entry) is established
as the basis for calculating the profile. As fluid is lost to
the formation, the radiation intensity of the succeeding
gamma ray trace will be less. Consequently, the height-base
sum will be less than the 100 percent rate by some propor-
tionate amount. However, highly irregular boreholes are
a major problem in trying to obtain a quantitative profile
by tracer surveys. It is ironic that the Self method is pro-
claimed to be the answer to the irregular borehole when
this one condition is the greatest cause of interpretive error.
A brief review of the basic properties of gamma rays
provides the answer to why this happens. Gamma rays are
high-energy electromagnetic waves with energies expressed
in units of million electron volts." When passing through
matter they are gradually absorbed, each additional inch
of material reducing their intensity by a given percentage.
The absorbing power is generally expressed as the half-
value thickness (HVT) which is the thickness of material
required to reduce a gamma ray to half of its intensity.
The H V T of water for 1-131, the most common radio-
active isotope used for tracer surveys, is approximately
2% in. Obviously the number of HVT's can vary con-
siderably with large variations in borehole diameter.
Abrupt hole size variations from 12 to 25 in. are very com-
mon in shot holes. Test data indicate that soluble 1-131
spreads readily in the fluid being traced, and the assump-
tion that radioactivity spreads evenly through the borehole
horizontally is valid. It could be deduced from the above
that a reduced radioactive response would be seen' with
relatively large increases in borehole sizes without any loss
of fluid to the formation.
Fig. 2 presents a theoretical case to simplify the ex-
planation of this phenomenon. In the upper portion of the
Fig, 1-Flowmeter recordings comparing 15-second (A), figure, the induced radiation is confined to 1 H V T and, for
6%-minute (B) and 10-minute (C) cycles. simplicity, a gamma count of 400 counts/second is assumed.
The lower portion of Fig. 2 shows this same amount of movement of the tracing tool through the slug as it moves
radiation spread out over a borehole radius of 4 HVT's. down-hole.
Each H V T now contains an equal amount of isotope equi- Fig. 3 is test proof that the above theoretical example is
valent to one-fourth of the radiation intensity of the bore- valid. The same amount of radioactivity was recorded by
hole radius of 1 HVT. The gamma detector now records a scintillation gamma ray detector in 2 % - , 12- and 24-in.
100 counts/second for HVT A, 50 counts/second for HVT pipe. The reduction in radiation counting rate is a direct
B, etc., for a total count of 187.5 counts/second, less than result of the increased number of HVT's between 2 % -
half of the original count. Consequently, the height of and 24-in. boreholes. The height of deflection in the 2 % -
deflection on the recording chart would be reduced by over in. pipe is 7 in. and the height of deflection in the 12-in.
one half. The height-base sum of the radioactive response pipe is 3 in. If no fluid were being lost to the formation,
would indicate a loss in fluid if the base of the response did the base of the radioactive response would have to increase
not increase in width proportionately. Vertical spreading of by 4 in. to indicate no loss by the Self method. On stand-
radioactive slugs occurs due to both laminar flow and ard logging paper with API chart divisions, 4 in. would
represent 80 ft on a 5-in. logging scale. In other words,
if no fluid were being lost to the formation and the radia-
tion response were reduced by 4 in. due to hole enlarge-
ment, the radioactive slug that was being traced would
have to spread vertically in the hole at least 80 ft. Although
not impossible, this is highly improbable.
GAMMA RECORD l N G OF Since there is no way of knowing if fluid loss to the
DETECTOR formation is actually occurring, the above method cannot
possibly produce quantitative results over shot sections.
In cased holes and fairly gauge open holes where the bore-
hole sizes change very little, this method gives fairly good
results. It can also provide limited qualitative interpretive
data in highly irregular boreholes, but should be limited
I I I to qualitative results only.
100 1 100 1 100 / 100
GAMMA CPS 1 CPS 1 CPS 1 CPS Absolute Temperature Surveys
I I I
DETECTOR
ACTUAL 100
COUNTING R A T E
50
I
' 25
I
125-1875CPS
Interpreting temperature surveys is perhaps the most
controversial subject in the entire realm of production log-
ging today. Many physical factors, in addition to human
error, affect interpretation of both absolute and differential
Fig. 2-Example case showing the theoretical effect of temperature surveys. The basic theory and interpretation of
HVT on gamma radiation. temperature logs have been presented by N o w a k , 3 i r d 6
and others, so only problems affecting interpretation will
be discussed.
Probably the greatest physical problem in obtaining sat-
isfactory static temperature surveys is that of down-hole
counterflow. This condition will exist in all injection wells
in which the zone being flooded is a stratified formation
with large variances in permeability. Most of the carbon-
ate reservoirs in the Permian Basin fall in this category.
In some wells, counterflow may exist only a short period
of time; in others, it is known to exist for over 48 hours.
The extremity of counterflow is dependent on cumulative
injected volumes, injection rates and pressures, degree of
reservoir fill-up and reservoir stratification. There is cur-
rently no known method of predicting the degree of coun-
terflow other than an actual measurement with either a
flowmeter or tracer survey.
The series of tracing runs in Fig. 4 shows the severity of
counterflow in an injection well under supposedly static
conditions. A series of radioactive slugs was ejected be-
tween the various zones taking fluid and monitored over
a period of 1 % hours. With the exception of the top slug
which was moving upward due to backflow, the rest of
the fluid was moving downward to the lowest zone. For-
tunately, counterflow presents interpretative problems for
temperature surveys only in those wells with large gross in-
tervals. If only short zones are being flooded (10 to 20 f t ) ,
temperature surveys still give good qualitative results even
if counterflow is occurring within this interval. The log
will definitely indicate if the fluid is being confined to the
pay section. This problem is a strong argument for also
running an injecting temperature survey.
Fig. 3-Test case proving the effect of HVT on gamma One major problem of temperature log interpretation is
radiation. positively discerning between channeling in the borehole

A U G U S T , 1967
and channeling out in the formation. Utilizing data de- in Fig. 7. This condition is diflicult to interpret by tem-
termined by Ramey' and evaluating over 300 after-fractur- perature logs alone. A tracer survey, because of its depth-
ing temperature logs, both theoretically and empirically, of-investigation limitation, would tell if this represented
Agnew8 concluded that sharply defined temperature anom- a remedial conditiori. If an upward movement were traced
alies actually denote fractured rock. If the anomaly ex- with scintillation detectors, a borehole channel would
tends either above or below the zone of interest and ad- exist. The volume of fluid leaving the pipe at this point
verse conditions result therefrom (i.e., excessive water pro- would govern the need for remedial operations. If no up-
duction) then very little can be done to remedy the situa- ward movement of fluid wcre traced, the fluid move-
tion. If anomalies are broken by a return to the approxi- ment would be out in the formation, thus minimizing the
mate gradient by the temperature traverse, a borehole possibility of corrective measures. This is a prime example
channeling condition is indicated. where more than one survey is required to pinpoint the
Figs. 5 and 6 show temperature logs depicting these con- problem.
ditions. Fig. 5 shows a sharply defined anomaly extending Identifying channeling conditions by temperature sur-
well below the productive zone. This indicates that the veys in water injection wells is more difficult than identi-
fracture is in the rock; if the lower extremity of the frac- fying channeling conditions by after-fracturing tempera-
ture has penetrated a prolific water-bearing zone, cement ture logs. Primarily, this is due to the large difference in
squeezing is unlikely to be successful. volumes of fluid that have been pumped into the wells.
Fig. 6 shows the condition which can be repaired. The Even with "River-fracs", the volume of treating fluid will
temperature curve returns almost to the gradient above seldom exceed 10,000 to 12,000 bbl, whereas injection wells
the productive zone before indicating another zone of will have cumulative injection volumes of several hundred
fracture. This is indicative of a borehole channel which thousand barrels. Assuming that the injection presswes
is more easily remedied than the vertical fracture out in are maintained below that required to induce formation
the rock. fracturing, the primary heat transfer mechanism when
There is one condition denoted by after-fracturing tem- injection ceases will be unsteady-state radial conduction.
perature logs which does not fit the previous interpreta- There will also be a certain amount of heat transfer due
tions. When sharply defined anomalies are not seen and to unsteady-state linear conduction, but it will not be the
the temperature traverse gradually returns to gradient result of fracturing. Vertical flow of heat between the zone
either above or below the zone of interest, a definite con- of intake and adjacent zones will be linear. In addition,
clusion cannot be reached. It is impossible to tell if the there could be some fluid crossflow between zones. The
anomaly is the result of a fracture well out in the for- net result is generally a temperature traverse similar to that
mation or a borehole channel. shown in Fig. 7.
Note the gradual return to gradient above the anomaly Many water input wells were hydraulically fractured

Fig. 4--Measuring counterflow in injection well under apparent static conditions.

JOURNAL O F PETROLEUM TECHNOLOGY

 prior to conversion to injection service. However, if
enough fluid has been injected into the well to extend the
effective temperature borehole radius to the horizontal ex-
tent of the fracture, then the temperature traverse will
still be primarily affected by unsteady-state radial conduc-
tion. There is no way of knowing when any of the above
conditions exist to the extent that they control the tem-
perature anomaly. Consequently, when analyzing injection
well temperature logs, the additional factors involved cre-
ate many more interpretive problems.

Differential Temperature Suweys

Differential temperature surveys can provide good sup-
plemental information to absolute temperature logs, but
only in a qualitative manner. There are essentially two
types of differential tools: those utilizing two sensing ele-
ments and those utilizing a single sensing element. The A
Priori differential tool uses the single-element principle and
is an improvement over the two-element principle because
the temperature of undisturbed fluid is being measured
and recorded at all times. A current fallacy is the quanti-
tative interpretation being applied to it.
The suggested theory supporting quantitative analysis
by the A Priori differential tool is totally contrary to all
previous temperature logging theories contributed by the
industry. Previous work supports the theory that the bore-
hole temperature of water moving by zones of intake will
remain constant. The proponents of the A Priori differen-
tial temperature theory "suggest" that the old theory is
wrong and "suggest" that the borehole temperature of the
Fig. 5-Temperature curve indicating fracturing in
water will decrease across zones of intake (provided the
the rock. fluid flow is laminar) and the temperature traverse will

Fig. 7-Example of temperature curve which cannot be

conclusively interpreted; indicates need for other
Fig. &Temperature curve indicating a borehole channel. surveys to isolate problem.

A U G U S T , 1967
have a negative slope across these zones." Since the A Pri- 2. Johnson, W. B, and Morris, B. P.: "Review of Tracer Surveys",
ori tool has been designed to successfully measure negative Paper 906-9-E, A P I Southwest Dist., Div. of Prod., Midland,
Tex. (March 18. 1964).
slope, it is "suggested" that only the differential deflection
to the left of some indiscriminately selected vertical line 3. Connolly, E. T.: "Reiume and Current Status of the Use of
represents fluid intake, with quantitative volumes being Logs in Production", paper presented a t SPR'LA Logging Sym-
posium, Dallas (May 4-7, 1965).
determined by the area under the differential deflection.
All of this new theory has been advanced after making the 4. Introduction to Schlurnberger Yell Logging, Schlmnberger Well
assumption that thermal conductivity occurs only within Surveying Corp. Document 8 (1958) 83.
the zone of intake and that none occurs between the zone 5. Nowak, T. J.: "The Estimation of Water Injection Profiles f ~ o m
of intake and adjacent zones. This assumption eliminates Temperature Surveys", Trans., AIME (1953) 198, 203-212.
from consideration the one variable which creates the prin- 6. Bird, J. M.: "Interpretation of Temperature Logs", Drill. and
ciple interpretive problem of analyzing injection well tem- Prod. Prac., A P I (1954) 187.
perature logs. Consequently, a standardized interpretation
is derived, apparently through the additional assumption 7. Ramey, H. J., Jr.: "Wellbore Heat Transmission", .I.
Pet. Tech.
(April, 1962) 427-435.
that an ignored problem does not affect the final analysis.
The A Priori tool is a step in the right direction in that 8. Agnew, R. G.: "Evaluation of Fracture Treatments with Tem-
perature Surveys", J. Pet. Tech. (July. 1966) 892-898.
a highly sensitive, single-element, differential temperature
tool has been needed. But the quantitative claims tend to 9. Johns, Earl: "Tracing Fluid Movements With a New Tempera-
create doubt in the validity of all the data which can be ture Technique", paper S P E 1750 presented at S P E Sympociurn
obtained with this tool. Fort Worth (March 5-7, 1967). ***
on Mechanical Engineering Aspects of Drilling and Production,

Conclusions
There is no single production logging tool or survey on L Y N ND. JONESis a petroleum en-
the market today that will consistently give a complete gineer with Texas Pacific Oil Co. re-
analysis of down-hole fluid movement. Standardized inter- sponsible for evaluating performance o j
pretive procedures many times tend to hinder rather than secondary recovery operations. H e was
aid proper interpretation. Knowledge of the limitations of previously a sales engineer with Cardi-
the various tools available will provide the basis of good. nal Surveys Co. in Odessa, Tex., active
sound analysis, resulting in more efficient and economical in sales and evaluation of production
operations in both primary and secondary phases of pro- logs. Jones was a petroleunz enginecr
duction. with Pan American Petroleum Corp.
from 1957 to 1965 working in all phases
References o f production engineering, primary and sec&ary' reser-
voir engineering and unitization work. A registered profes-
1. Jones, L. D.: "Proper Tool Selection for Refined Production sional engineer in Texas, he received his BS degree in
and Injection Well Profiling". paper presented at the Southvest-
ern Petroleum Short Course, Lubbock, Tex. (April 21-22, 1966). petroleum engineering from The U . o f Texas in 1957.