SPE 132220

A Case Study: Profiling Gas Production in the Tubing/Casing Annulus, Using Noise/Temperature Logging Techniques M. M. Molenaar, SPE, Shell Canada Ltd.; B. M. Cowan, SPE, Weatherford Canada Partnership; E. Fidan, SPE, Shell International Exploration & Production, Inc.; D. A. Cuthill, SPE, Weatherford Canada Partnership

Copyright 2010, Society of Petroleum Engineers This paper was prepared for presentation at the CPS/SPE International Oil & Gas Conference and Exhibition in China held in Beijing, China, 8–10 June 2010. This paper was selected for presentation by a CPS/SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract To reduce liquid loading on multizone unconventional gas wells, tubing can be run to a depth below the lowest perforation interval. Gas flows down the tubing/casing annulus and flows up the tubing, eliminating liquid buildup. For production surveillance, this wellbore configuration is not conducive to obtaining conventional production logs.

Conventional production profiling techniques involve repositioning the tubing string or removing it altogether. If the tubing remains in place during logging, the costs associated with pulling the tubing are eliminated; production is not suspended; and the risks associated with well control are reduced. Also by not modifying the wellbore configuration, fluid velocities are not affected and the log results more closely represent the actual production profile. Ideally the well should be logged without manipulating the tubing to provide a representative production profile.

Noise/temperature logging has been used for many years to assist in locating sources of fluid flow behind casing. The use of this technique to obtain a pseudo or qualitative production-flow profile in the tubing/casing annulus was explored to enable the tubing to remain in the wellbore and obtain a measurement of the flow behind pipe.

In the 1970s, tests where conducted to quantify wellbore inflow using noise logs. The research was recently used to evaluate numerous wells in western Canada with favorable results. This paper discusses the logging method and presents comparisons to profiling results from conventional production-logging techniques with emphasis on the cost savings to the operator.

Introduction In general, tight-gas developments require a dense spacing of wells, ranging from a few to several wells per section. On these wells, regulators often require that a production-logging survey be acquired periodically. In Canada these requirements vary from province to province. Often, in the absence of regulatory requirements, operators must determine if the relative contribution of each completion zone continues to deliver gas as originally identified in logs. These tight-gas reservoirs are usually completed over several zones in stacked-pay, and each zone has varying subsurface qualifiers. These qualifiers can include different system permeability, stress distribution, and natural fractures—all tend to increase the delivery anisotropy between zones, especially when all zones are commingled. This anisotropy in producibility can create early water loading of some zones or even complete loss of production from zones within the first 3 months of a well being placed on production.

It is critical to run a periodic flow profile that assists in predicting earlier-than-expected termination of the completed net pay. If remediation is required, corrective action could include refracturing, or if the well is liquid loading, then running a plunger or foam injection could be recommended.

The water flowback during production usually becomes a hindrance to effective gas lift, forcing operators to run production tubing deep immediately from the onset of the production. Deep-run tubing, which is hung past the deepest set of perforations, can help lift the liquids and delay liquid loading in the wellbore. Gas inflow is directed downward within the tubing/casing annulus, thereby lifting the liquid buildup from the wellbore as the gas is produced up the tubing.

Normally monitoring is performed using downhole production-logging tools (PLT) to provide a production profile across the perforated intervals. These tools must be exposed to the wellbore and be run within the flowstream to measure fluid flow rates, pressure, temperature, and density. If wells have deeply run tubing, traditional production logging can only be accomplished by

2 SPE 132220

pulling the tubing to a depth that is at least above the top perforated interval. Repositioning the tubing presents numerous challenges, including the costs associated with moving the tubing string; possible loss of well control as the tubing is manipulated; and at the time of logging the well not being produced under the normal producing conditions. Also to reposition the tubing, the well must be shut in, and as a result, stabilized flow may not be obtained at the time of logging the well. A preferred method would enable production logging without manipulation of the tubing.

Noise- and temperature-logging tools have been used for many years to identify flow behind casing within cement channels and for surface vent-source location. Noise logging uses a sensitive microphone to record the sounds/noises attributable to pressure drops as fluids move through spaces behind the casing. Because the normal application of noise/temperature logging is to identify flow behind pipe, it was thought that the technique could be applicable for production surveillance for deep-run tubing wellbores. Noise/temperature logging was performed on a number of tight-gas wells with deep-run tubing to investigate the applicability of the technique for production logging. Case studies are presented that demonstrate this new technique, and when the noise/temperature and PLT where run, favorable comparisons are observed. Noise Logging A noise log is a record of the sounds measured at different positions in the borehole. The sounds of moving fluids or the hiss of escaping gas are caused by disturbances in a liquid/gas interface or by turbulence in the fluid stream. It is the sounds of this turbulence that can be used to detect flow. In a wellbore environment, the noise log is very effective for gas detection as the gas flows up through liquid, but it is also effective for the detection of various kinds of gas, water, or oil single-phase flow, including channeling behind pipe (assuming that there is adequate turbulence in any given situation to produce enough noise).

Because fluid turbulence generates sound, high-noise amplitudes indicate locations of greater turbulence, such as leaks, channels, and perforations. Noise logging is used primarily for detecting channels in the cement sheath for gas influx. It has also been used to measure flow rates, identify open perforations, detect sand production, and locate gas-liquid interfaces (Hill, 1990).

The noise log is a presentation of a series of stationary readings plotted against depth, as shown in Fig. 1. Stationary measurements are preferred to avoid noise generated because of tool movement against the casing and also resulting from the wireline motion through the surface-hanging equipment. The noise tool was first introduced in 1955 (Enright, 1955). However only after many laboratory studies where completed in the early 1970s did the tool begin to have wider usage. In 1973, McKinley et al (1973) outlined the utility of the noise log for both qualitative and quantitative applications. Logging Approach The main purpose of using a noise/temperature-logging technique was the capability to log the well without substantial modification of the wellbore from the normal production configuration. However because the tools were to be run within the tubing, it was necessary to divert production up the tubing-casing annulus to eliminate the noise effect of production fluid movements across the tool housing. Gas flow is diverted into the annulus, up and out to surface facilities while logging. The noise created by gas and liquid flow is recorded along all the perforating intervals outside of production tubing while the log is acquired from inside the tubing.

Well Preparation. The gas flow is diverted through a valve assembly connected to the annulus. Flow is directed into a flowline or a surface-testing unit to permit monitoring of the surface-flowing conditions during logging. In some instances a plug was set in the bottom of the tubing to allow liquid to be placed in the tubing to assess the effects of better sonic and thermal coupling between the logging tools and the noise source. When liquid was added, the level was maintained above the top-most perforated interval, as indicated in Fig. 2. The well production was allowed to stabilize while running the tools into the well.

Logging Tool String. The noise-logging tool is simply a sensitive microphone and amplifier that is used to detect the sounds produced by fluids either entering the wellbore or flowing behind pipe. Typically a piezoelectric-crystal, sound detector is used. The detector generates a voltage that is transmitted up the wireline conductor to a surface panel. The noise amplitude spectrum is filtered through many high-pass filters that present the noise amplitudes above 200, 600, 1,000, 2,000, 4,000, and 6,000 Hz. These different frequency ranges can be tied to different noise sources or could be an indicator of the fluid-flow regimes. Frequently the noise tool is paired with a temperature tool to provide auxiliary and complimentary data and to assist with the interpretation. A through-tubing collar locator and gamma ray tool is also run to enable depth correlation, as shown in Fig. 3.

Logging procedure. With the well flowing up the annulus, the logging tool string is run downward to the bottom of the tubing, at the same time recording a flowing temperature survey. The noise tool is pulled upwards taking stationary readings every 2 to 2.5 m across the perforations and at 5 m station stops between zones. The well is subsequently shut-in, and additional noise/temperature surveys are run at approximately 1.5 and 2.5 hr after shut-in.

SPE 132220 3

Analysis method The analysis method uses equations originally developed by McKinley et al (1973). The relationships were derived for calculating a flow rate within a channel that essentially describes the wellbore configuration for deep-run tubing with flow diverted up the annulus. The interpretive technique for a deep-run tubing application relies on the following assumptions: • The fluid phase behind the casing is constant. • The pressure drop for all the entries is constant. • Noise-tool, correction factors are assumed to be constant (correction factors for line type and length are not used). • The fluid in the tubing is constant and homogeneous.

The analysis consists of first identifying cooling areas on the producing temperature log. These areas would indicate cooling behind the tubing as the gas exits the perforations and expands into the tubing/casing annulus. The noise log is then visually analyzed for indications of fluid entries across perforated intervals that had not been identified as productive with the temperature log. Once all potential productive perforations have been identified, the amplitude of the 1,000-Hz noise curve is averaged across each interval. Based on research conducted by McKinley et al (1973), the single-phase gas rate at the entry points can be determined:

( )P

NQΔ

−⋅=

65 *1000 (1)

Where: ΔP = pressure drop, psi Q = flowrate, ft3/D (downhole conditions) N*

1000 = normalized peak-peak noise amplitude (> 1000 Hz), mV

If the pressure drop is assumed constant, then the pressure drop can be ignored as inflow is calculated on a percentage basis.

( )65 *1000 −⋅= NQ (2)

To use the methods developed by McKinley et al (1973), the noise-tool readings must be calibrated to a standard EPRC sonde

by applying various corrections to the noise amplitude:

GMTL FFFFNN ⋅⋅⋅⋅= 1000*1000 (3)

Where: N*

1000 = normalized noise amplitude, mV N1000 = recorded noise level, mV FL = line correction factor FT = tool gain factor FM = meter factor FG = geometry factor

Several of the corrections noted in Eq. 3 are based on outdated analog measurements, as a result several assumptions where

made to account for technological tool improvements: • Tool gain factor, FT = 10 (tool gain was 100) • Line factor, FL = 1 (line variations not a factor) • Meter factor, FM = 1 (digital meter was used) • Geometry factor, FG = 30 (gas in both tubing and annulus), FG = 10 (fluid-filled tubing)

The percentage inflow per interval is then calculated by dividing the interval 1000-Hz averaged amplitude by the cumulative

effective interval total:

4 SPE 132220

∑=

=

= zi

ii

ii

Q

QQ

1

(%) (4)

Where: Qi(%) = percentage inflow for interval i Qi = normalized “flowrate” for interval i z = total number of intervals

Case Studies The application of this interpretive technique for noise and temperature measurements was investigated as a cost-effective surveillance method for tight-gas wells to measure the well inflow profiles without moving the tubing. Several field trials were conducted to prove the technology and to establish the application range as a potential pseudo-PLT. Initially the trials were run in a fluid-filled tubing to achieve maximum coupling. The last field trial was an experiment with the tubing gas-filled. Well Configuration. In multi-stacked, vertical wells in the Western Canadian Basin, tubing is run to mitigate liquid loading. In general the vertical wells are completed with a cemented production casing and multiple intervals are perforated in each well, normally with 0.5 to 1 m interval lengths. In some wells there may be 20 individual perforated intervals, as shown in Fig. 3. The tight-gas wells are then fracture-stimulated, most commonly with water-based fracturing fluids, either by large volumes of “slickwater fracturing technology” or smaller volumes of crosslinked fracturing fluids to place higher concentrations of proppant. In some hybrid applications, slickwater and crosslinked gel stages are combined for optimization of fracture conductivity.

The hydraulic fracture flowback stage is usually performed quickly (from immediate flowback to flowing back several zones in 2 to 3 days) to ensure that the completions fluid is recovered efficiently. Because of the generally tight nature of these reservoirs, the completions fluid cleanup takes long periods of time, even after a well is put on production. In addition, formation water production is common, although usually at very low rates. To improve the lift of fluids and mitigate liquid loading a tubing or velocity string is installed, usually past the deepest set of perforations.

Well 1. The well was completed with seven perforated zones and six hydraulic fracturing stages. After the hydraulic fracture flowback of the well, a PLT survey (pressure/temperature/fluid-density/spinner-flowmeter) was run inside the 114.3 mm casing before the 60.3 mm velocity string was installed. A flow profile was developed from the PLT data and used as the baseline for future production surveillance activities, as indicated in Fig. 4.

After the well was on production for approximately 1 month, the well was conditioned to a flow rate similar to what had been maintained during the PLT survey, and a noise/temperature log was run. The well was flowed up the casing with tubing filled with liquid and shut in throughout the survey. Stationary noise readings were obtained at specified intervals under both flowing and shut-in conditions. Temperature surveys were also run down under flowing and shut-in conditions.

The noise-log response suggested that single-phase flow (gas) was present. Based on the values of the noise generated from perforation inflow into the casing and temperature-log responses, a flow profile was developed using the noise amplitude of the 1,000-Hz curve (as described by the analysis method). The profile for the relative percentage inflow for each perforated interval is presented in Fig. 5.

Results. The earlier run PLT survey permitted a direct comparison to the noise/temperature-derived production profile. The

noise profile was found to be comparable to the PLT results, as indicated in Fig. 6. Both the noise and PLT survey profiles identified the same zones and intervals to be the major producers. The noise log appears to show minor gas entry locations more clearly than the PLT-generated profile. The noise profile assigns more flow to Zone C (3800 to 3950 m) and to Zone A (3500 to 3700 m) and also measures a higher contribution from Zones E–G (4000 to 4200 m), as shown in Table 1.

Upon closer examination, the PLT spinner-flowmeter survey indicates a strong slug-flow response in the area of Zone A, as indicated in Fig. 7a. As a result, the interpretation of the PLT data is more subjective, and this could account for the difference in the interpreted inflow profiles. Fig. 7b also shows that the PLT spinner over Zone B suffers from the slug-flow response. The noise survey shows that the reduced annular flow area, resulting from the installation of the velocity string, results in better lifting of the well fluids (and also potentially less slugging), which opens up the bottom zones to more flow.

SPE 132220 5

Well 2. The well was completed with five perforated zones and five hydraulic-fracturing stages. This well was used to determine the operational requirements for noise logging in a “dry” (gas-filled) tubing versus a “wet” (liquid-filled) tubing.

For wells that have been on production for some time, the energy (reservoir pressure) required to clean up produced liquids from the wellbore lessens. As a result, it is not advantageous to introduce additional liquid to the wellbore. Also there could be instances where there is a reason not to expose the formation and the hydraulic fracture to new fluids that could negatively impact the deliverability of the well. However, a gas-filled tubing reduces the coupling between the noise/temperature tool and the tubing-casing annulus.

To demonstrate the operating range of the tool and to understand the extent of the loss in sensitivity, a field trial was conducted. The noise/temperature tools were logged in a gas-filled tubing, and directly thereafter, in a liquid-filled tubing after a plug had been installed in the tubing bottom. The well was representative of a well that has been on production for some time—the flow rate of this well was considerably less than Well 1 and was very close to liquid-loading conditions. As a result this well provided a challenge for the noise survey even under optimum “wet” logging conditions.

Results. The results of the noise/temperature logging and the inflow profile interpretations are shown in Fig. 8. The noise log that was run under dry conditions was comparable to the log run under wet conditions—similar zones are noted to be effective regardless of well conditions (wet or dry) under which the measurement is acquired (Table 2).

In the application of the noise-logging tool for surveillance, it is important to understand that the main objective is to capture the changes in the inflow distribution over the producing length of the well over time. The results of this trial demonstrate that wells can be logged dry, and producing zones can be identified. However the results do not appear to be as definitive as with a wet-tubing string. Conclusions The noise/temperature method of profiling production was shown to be a cost-effective option for production logging in tight-gas wells with deep-run tubing. The cost effectiveness is realized through minimal downtime for the logging operation (no deferred production); an effective gas lift that is continuously supported while logging because the tubing remains in place; the capability to correlate to previously obtained through-casing PLT to provide surveillance of lift efficiency; and no costly hoist and/or service rig is required. It has been demonstrated that qualitative inflow profiles can be obtained directly after installation of the tubing, whether the tubing is liquid- or gas-filled. Benefits. • No need to pull tubing to log which eliminates the need for snubbing. • A representative flow profile can be obtained as compared to PLT. • The noise/temperature string has no moving parts to fail. • The less complex tool string results in lower logging costs. • Pressure-control requirements are minimal in situations where the tubing is plugged before logging. Future considerations. • Can a noise profile be obtained while flowing the well up the tubing? • Does the profile change if the flow is directed up the tubing? • Is the profile more or less accurate than that obtained using a spinner in a slug flow regime? • Can liquid carbon dioxide, nitrogen, or methanol be substituted for water as the tubing-fluid?

Acknowledgements The field trials were initiated by the Shell Unconventional Gas Centre of Excellence (UG CoE) in Canada, as part of the ongoing technology development and deployment, to support the onshore North America gas ventures. The authors extend their thanks to the members of the Deep Basin venture team, study team, completion team, and surveillance team for their contributions to the project. Special thanks to B. Oz and D. Miller for supporting this study.

The authors wish to thank J. Woodford and the members of the Weatherford crew for their efforts in acquiring the data used for this study. Thank you also to P. Huber for his advice, guidance, and contributions in developing and refining the methodology used in this study.

The authors would also like to thank Shell Canada Ltd. for graciously allowing publication of this paper.

6 SPE 132220

References Enright, R.J. 1955. Sleuth for Down-Hole Leaks. Oil & Gas J. (28 February 1955) 78–79. Hill, A.D. 1990. Production Logging – Theoretical and Interpretive Elements. Monograph Series, SPE, Richardson, Texas 10: 112-121. Lohuls, G., Flexhaug, L., Huber, P., and Baker, C. 1991. Coiled Tubing/Production Logging in Highly Deviated and Horizontal Wellbores. Oral

presentation given at the CIM/AOSTRA 1991, Banff, Canada, 21–24 April. McKinley, R.M., Bower, F.M., and Rumble, R.C. 1973. The Structure and Interpretation of Noise from Flow Behind Cemented Casing. JPT.

(March 1973) 329–338. Smolen, J. 1996. Fluid Movement: Noise Logging. In Cased Hole and Production Log Evaluation, Chap. 14, 255-268. Tulsa, Oklahoma:

PennWell Publishing Company.

SPE 132220 7

Fig. 1: Example of a noise/temperature survey for channel detection.

8 SPE 132220

Fig. 2: Schematic of common wellbore configuration.

Fig. 3: Noise/temperature tool string configuration.

SPE 132220 9

Fig. 4: PLT survey acquired before installation of the velocity string on Well 1.

10 SPE 132220

Fig. 5: Production profile generated from noise/temperature data for Well 1.

SPE 132220 11

Fig. 6: Overlay of production profiles generated from the PLT and noise/temperature. Flowing temperature surveys show good agreement indicating well was producing under similar conditions.

12 SPE 132220

Fig. 7a: Illustration of the slug-flow response and poor repeatability on the spinner-flowmeter survey from the PLT log for Well 1, Zone A.

Fig. 7b: Illustration of the slug-flow response and poor repeatability on the spinner-flowmeter survey from the PLT log for Well 1, Zone B.

SPE 132220 13

Fig. 8: Production profiles generated for Well 2 show relatively good agreement between the profiles developed from "wet" and "dry" logging conditions.

14 SPE 132220

TABLE 1–SUMMARY OF NOISE AND PLT PROFILES FOR EXAMPLE WELL 1

Formation Perforation Interval (m) Qg (%) - Noise Qg (%) - PLT Comments

Zone A 3482.0 - 3482.3 0.0 0.0 6% total flow difference between surveys 3497.0 - 3497.3 2.7 0.0 3504.0 - 3504.3 0.0 8.7 3509.0 - 3509.3 0.0 0.0 3512.0 - 3512.3 0.0 0.0

Zone B 3624.0 - 3624.3 2.9 13.4 Noise (16.8%) versus PLT (34.9%) 3629.0 - 3629.3 4.5 4.7 3671.0 - 3671.3 3.7 5.4 3679.0 - 3679.6 5.7 11.4

Zone C 3855.0 - 3855.3 0.0 0.0 Noise (64.3 %) versus PLT (51.3 %) total production 3864.0 - 3864.3 31.6 31.1 3868.5 - 3868.8 0.0 0.0 3893.0 - 3893.3 32.7 20.2 3897.0 - 3897.3 0.0 0.0

Zone D 4010.0 - 4011.0 0.0 0.0 Ineffective perforations 4029.5 - 4030.5 6.2 0.0 Noise (9.3 %) versus PLT (1.5 %) 4034.0 - 4034.3 0.0 1.0 4039.5 - 4040.5 0.0 0.0 4049.0 - 4049.6 3.1 0.5 4055.0 - 4055.3 0.0 0.0

Zone E 4068.5 - 4069.5 0.0 0.0 No indications of inflow from the Zone E 4072.0 - 4072.3 0.0 0.0 4075.0 - 4076.0 0.0 0.0 4077.0 - 4077.3 0.0 0.4 4082.0 - 4083.0 0.0 0.0 4084.0 - 4085.0 0.0 0.0 4085.0 - 4086.6 0.0 0.0 4086.0 - 4087.0 0.0 0.0 4092.0 - 4092.3 0.0 0.0 4094.0 - 4096.0 0.0 0.0

Zone F 4113.0 - 4115.0 2.7 0.0 Minor inflow noted on noise log Zone G 4176.0 - 4176.3 1.2 1.4 Noise (4.2%) versus PLT (2.3%) from this interval

4191.0 - 4191.3 0.9 0.5 4195.0 - 4195.3 1.0 0.5 4240.0 - 4240.3

1.1

0.0 4240.0 -4240.3 0.0 4267.0 - 4267.3 0.0 4276.0 - 4276.3 0.0

SPE 132220 15

TABLE 2–SUMMARY OF WET AND DRY PROFILES FOR EXAMPLE WELL 2

Formation Perforation Interval (m) Qg (%) - Wet Qg (%) - Dry

Zone A 3300.0 – 3300.6 0.0 0.0 3388.0 – 3388.6 0.0 0.0 3415.0 – 3415.6 0.0 0.0 3422.0 – 2422.9 0.0 0.0

Zone B 3560.0 – 3560.6 3.8 6.7 3572.0 – 3572.6 4.5 12.0 3605.0 – 3605.6 4.1 0.0 3616.0 – 3616.3 0.0 0.0 3631.0 – 3631.6 0.0 0.0

Zone C 3657.5 – 3658.4 5.2 0.0 3676.0 – 3676.3 0.0 0.0 3680.0 – 3680.3 0.0 0.0 3710.5 – 3711.1 0.0 0.0 3723.0 – 3723.6 6.3 0.0

Zone D 3802.0 – 3803.2 8.9 0.0 3822.5 – 3823.4 8.5 5.8 3832.0 – 3832.6 8.0 3.2

Zone E 4075.5 – 4076.1 16.8 27.2 4086.0 – 4086.9 18.6 21.3 4097.0 – 4079.6 17.3 24.0 4103.0 – 4103.6 N/A N/A