Analysis of the acoustic response in water and sand of different fiber optic sensing cables Joachim Hofmann, Massimo Facchini, Mark Lowell Affiliation: Brugg Cables – Fiber Optic Systems

ABSTRACT Distributed Acoustic Sensing (DAS) is a highly promising technology to efficiently monitor assets for energy production and transportation, both off- and on-shore, such as boreholes, pipelines and risers. The aim of the hereby-presented measurements is to evaluate the sensitivity of the different optical fiber cables to acoustic signals in sand and water, independently from the DAS read-out unit type and manufacturer. Acoustic sensing cables specifically designed by BRUGG Cables are characterized and compared to standard telecommunication cables. The spectral response of each cable was quantified using an all-fiber Mach-Zehnder interferometer. The response was also measured with calibrated microphones in order to convert the measurements into absolute physical units (Pascal). The measurement campaign is part of an investigation program for a reliable DAS system, which comprises the sensing cable (including installation procedure), the interrogator unit and suitable software. Keywords: Distributed Acoustic Sensing (DAS), fiber optics sensing, acoustic sensing cable, harsh environment TEF, Mach-Zehnder interferometer

1. INTRODUCTION 1.1 Choosing the right acoustic sensing cable For fiber optic sensing applications, the sensing elements, optic fibers, are packaged in a fiber optic cable. The fiber optic cable becomes in essence the fully packaged sensor. Can a fiber optic cable be designed to a better sensor, specifically for distributed acoustical sensing (DAS)? This study addresses this question. Until now there has been a lack of knowledge about how to chose the construction parameter of an acoustic sensing cable in terms of geometry (stranded, tube-encapsulated), materials (metal, polymer, Kevlar, glass roving) and type of sub-element (tight- or lose-tube, with and without gel) to build a sensor optimally designed for the specific application and environment. Not only should the sensor provide suitable spectral acoustic sensitivity, but also be compatible with the application environment and installation procedures. Particularly critical are the reliability requirements in downhole applications that require Tube Encased Fiber (TEF) cables. This study includes one TEF configuration among other fiber optic cables designed for sensing and telecommunication means. The sensor structure may need to be matched to the specific the environment where the cable is used. The spectral response of a monitoring system is strongly influenced by the nature of the propagation medium and by how and where the sensor is installed: immerged in a liquid milieu (petroleum, water), embedded in soil, attached to a structure (borehole casing, pipeline, or other industrial facility) or, more rarely, suspended in air. In the present study, attention is focused on the most recurrent situations, where the sensor is immerged in water or buried in sand. Depending on the application type and environment, acoustic monitoring systems must be able to efficiently measure in specific frequency ranges, which are summarized in Table 1 for some application fields [1].

Application Frequency range

Downhole seismic 1 - 500 Hz

Land seismic 1 - 500 Hz

Streamer 1 - 500 Hz

Subsea cables 1 - 500 Hz

Downhole acoustic 5 - 5000 Hz

Pipeline leakage > 20 kHz

Table 1: Typical frequency ranges for geophysical applications

There are further Distributed Acoustic Sensing (DAS) applications where acoustic signals propagate at higher frequencies (above 20 kHz), such as monitoring of liquid natural gas pipelines, or in other industry fields, e.g. partial discharge detection in high voltage power lines, or early crack detection of stainless steel structural ropes in cable stayed bridges. No regulations or standard procedures currently exist for the qualification of acoustic sensing cables to be used by installers, operators and end-users that provides guidance for selecting the most appropriate acoustic sensing cable. The present work proposes a methodology to achieve a comparative evaluation of different sensors, in specific environments and over defined spectral ranges, independently from any commercial DAS read-out unit type and brand. 1.2 Scope of the present work To evaluate and compare the spectral response functions of different sensing elements of various cable design, two real application scenarios are stimulated by placing 10 meter samples in either a water tank, for the frequency range from 500 Hz to 50 kHz, or buried in a sand box, for the frequency range from 20 Hz to 500 Hz. In account of the intrinsic spatial distribution nature of sensing cables as used in DAS, their sensitivity should be normalized to the unit length.

2. DAS APPLICATION FIELDS AND MEASUREMENT TECHNIQUE 2.1 Potential application fields of DAS Distributed Acoustic Sensing (DAS) is applied in the Oil and Gas industry for a wide variety of monitoring system for smart/intelligent wells, along pipelines, subsea structures such as umbilicals or risers, and in energy production and transport facilities, both on-shore and off-shore. The aim of DAS monitoring is to optimize production or monitor system integrity or health through data-informed decision-making. The following potential applications for DAS are amongst those already implemented [2]: •

down-hole applications: • • • • • • • • • • • • • •

distributed flow measurement in oil wells sand detection in wells - provides distributed condition monitoring of multi-stage sand control completions to facilitate remedial steps to isolate problem zones gas breakthrough - identification and isolation gas lift optimization - monitor operational status of individual gas lift valves, adjustment overall gas lift program electric submersible pumps (ESP) monitoring - electrical motor condition monitoring using noise pattern recognition techniques intelligent completions performance - condition monitoring of downhole zonal flow control devices e.g. for erosion effects leak detection/well integrity and formation subsidence monitoring - allowing remedial activities to be planned seismic - e.g. vertical seismic profiling velocity measurements and tube wave detection hydraulic fracture stimulation diagnostics (cross-well, in-well) wireline distributed acoustic sensing (various apps) real-time casing imager (RCTI) for real-time completion monitoring (RTCM) monitoring of permanent casing installations, often performed in tandem with distributed temperature sensing (DTS) carbon dioxide sequestration monitoring (in-well)

•

Other application fields in the oil and gas industry, e.g. on pipelines and, in general, transport and production facilities: • • • • •

leakage detection structural integrity identify flow regime – e.g. slugging perimeter security seismic reservoir monitoring

Even more applications suitable for DAS will be identified as the technology further matures with improved performance of read-out units (higher sensitivity, extended spatial range, enhanced spatial resolution), specific sensors design (increased reliability in harsh environment and under heavy duty), more reliable data processing algorithms (less false alarms, better capability of discriminating and identifying different phenomena), and commonly agreed regulations and industrial standards.

2.2 Brief review of DAS measurement techniques To become functionally a sensor, the fiber optic cable sensor must be connected to an instrument that interrogates the fibers and provides the sensor data. These instruments containing the optical and electrical hardware along with the special software are called interrogators. As is practiced today, interrogators utilize time-resolved DAS monitoring adopted from typically two distinct types of optical measurement techniques: DASI and DASP. The first and earliest practiced form is intensity-based (DASI). DASI relies on the Coherent Optical Time Domain Reflectometry (COTDR), where the intensity signal resulting from self-interfering backscattered optical pulses is measured. High-sensitivity photodetectors are required for DASI based interrogators. This simple approach yields a nonlinear and compressed measure of the acoustic signal. Despite signal infirmities, DASI interrogators are still quite effective when used for applications such as leak and intrusion detection. However, it isn’t particularly effective when signal-processing techniques require coherent array processing methods. The second form of DAS, based on more recent developments and typically requiring more expensive optoelectronic technology, is phase-based (DASP). DASP interrogation is not prejudiced by DASI-type errors and adopts an interferometry approach, where optical phase demodulation is used to obtain a precise measure of the acoustic signal. This approach is intended to provide for stable and linear representations of the distributed acoustic signals and is highly appropriate for coherent array processing, commonly used for conventional geophones, accelerometer and hydrophone sensor arrays. The measurement results presented in the present study provide an evaluation of the acoustic sensitivity of fiber optic sensors that apply mostly for use with DASP interrogators.

3. SELECTION OF FIBER OPTIC ACOUSTIC SENSING CABLES FOR TESTING Fiber optic cables are of two primary designs: tight-buffered and loose-tube. In tight-buffered fibers an additional polymeric layer is applied bonded to the optical fiber. With a tight-buffered fiber, the strain or pressure waves are efficiently coupled from within the cable surroundings to the sensing fiber. Problems that can be encountered with tight-buffered fibers in cables are that the fibers are prone to microbending, optical attenuation increase, or even reliability issues. Loose-tube designs contain the fibers loosely in tube, usually with the optical fibers suspended in a gel, typically introduced for water blocking capacity. These tubes are developed to protect the fiber from damage and performance deterioration due to handling and environmental stresses. The gel provides viscous drag on the fiber and may be accountable for good sensitivity to small strains down to very low frequencies. Common practice suggests that loosetube cables without gel filling are generally to be avoided, providing reduced sensitivity compared to gel-filled cables. Ideally, acoustic sensing fiber optic cables should be built with sub-elements and materials with a mass density and typical sound velocity similar to that of the neighboring elements in the cable and of the environment surrounding the cable. This would offer the smallest acoustic impedance with the soil and therefore the highest acoustic sensitivity. The cable geometry, i.e. diameter, also has an impact on the acoustic coupling efficiency. Although these ideals for designing

sensing cables with high sensitivities are well understood, often the choice of a sensor must be balanced against cost issues, security and operational specifications, rough installation and use procedures, compatibility with extreme environment conditions, such as high temperature and hydrostatic pressure, presence of aggressive substances, extreme mechanical stresses. In this complex paradigm the choice of sensing cables for DAS falls over a broad spectrum of very distinctive construction typologies for different applications. The fiber optic cables tested in the present work (presented in Table 2) represent some of the DAS sensor categories used today in the field. They range from simple optical fiber buffer tubes, standard telecommunication cables, specifically designed acoustic sensing cables, to downhole-dedicated Tube Encapsulated Fiber (TEF) [9]. Acoustic sensing cables, like BRUsens AC1 and BRUsens AC2, have already proven their efficiency in field applications of DAS, both in soilburied configuration (e.g. intrusion detection) or tightly attached to the structure under monitoring (e.g. leak detection in pipelines). Specially designed cables have been developed to increase the spatial sensitivity of a fiber optical cable. In the BRUsens High-Resolution sensing cable, the optical fibers are wound with a lay length of some millimeters, in order to pack 13 meters of optical fibers in one meter of cable length. This construction expedient increases the spatial resolution of the sensor by a factor 13 and amplifies its sensitivity to acoustic stimuli. The TEF downhole cable type is represented by Brugg Cables exclusive design BANTAM TEF. This is a harsh environment cable purposely developed as a sensing cable with small outer diameter to cope with reduced space requirements compared to usual control lines and TECs usually installed in oil wells. It is designed to withstand the harsh conditions found in downhole environments: elevated temperature, extreme hydrostatic pressure, and high concentration of aggressive substances. Its reliability is assured by the specialty hydrogen-resistant optical fibers with controlled Fiber Excess Length (FEL), additionally protected by an inner metal tube, filled with high temperature scavenging gel, and a rugged thick wall outer tube made of Incoloy A825.

Name

Schematic cross-section

Short construction and functional description

Tightbuffer

Tight-buffered multi layer fiber.

Plastic loose tube

Plastic loose-tube, gel-filled with optical fibers

BRUclean BC150W

Fiber optic cable with central gel-filled plastic loose-tube, with glass armoring and strain relief, corrugated steel armoring for direct burial and good rodent protection, outer HDPE sheath.

BRUclean BC 300

Fiber optic cable with stranded gel-filled plastic loose-tubes, glass outer armoring and strain relief, outer HDPE sheath.

BRUsens HighResoluion

Special fiber optic cable with optical fibers in the core, including some electrical conductors. High-resolution sensing fibers in the outer layer allowing for about 13 m of special optical fiber in one meter of cable.

BRUsens Acoustic AC1

Small fiber optic sensing cable for acoustic sensing applications. Acoustic coupled fiber in metal tube providing hermetic seal with rugged outer PA sheath.

BRUsens Acoustic AC4

Fiber optic cable for acoustic sensing applications with special tightbuffered fibers, aramid strain-relief and PUR outer sheath.

BANTAM TEF©

Fiber optic cable for harsh environment such as downhole applications. Gel-filled inner metal tube with optical fibers protected in rugged thick wall outer tube Incoloy A825.

Table 2: Selection of fiber optic cables under test. The schematic cross-sections are not to scale.

4. OPTICAL MEASUREMENT SETUP AND ACOUSTIC TEST BENCH 4.1 Fiber optics phase-based measurement technique (DASP) fundamentals For technical reasons discussed in previous sections, such as the compatibility with the DASP measurement methodology, predominantly used in field applications, it has been decided to develop within the present study a measurement setup based on the phase analysis of the optical signal. This measurement configuration provides also advantages in terms of cost-effectiveness and ease of realization with commercially available optoelectronic devices. Acoustic emissions are pressure variations in an elastic medium. The sensing principle of an intrinsic fiber optic acoustic sensor is based on the change in the optical path length induced by the acoustic pressure waves [3, 4]. The phase ! of the signal propagating in an optical fiber of length L is given by: ! ! !!! !

!!!!!"" !

!,

(1)

where " is the propagation constant, #eff the effective refractive index of the fiber glass at the wavelength $ of the optical signal. Fiber optic acoustic phase measurement systems are usually calibrated as a function of the effect of pressure units, with the normalized sensitivity defined as the relative phase shift induced by a given acoustic pressure at a specific frequency. The pressure sensitivity of the optical phase in a fiber is defined by the term: !" !!!"

!"!

(2)!

This expression is known as the normalized acoustic phase responsivity (NR) and is expressed in Pa-1. Where ! is defined by equation (1) mainly by the fiber length and %! is the shift in the optical phase due to a pressure change %P. In an interferometric system the optical phase variation due to dynamic strain influences on the measurement arm is transformed in amplitude changes in the intensity of the interference signal and is duly measured by a photodetector. 4.2 Optoelectronic measurement setup In order to compare different fiber optic cable designs, an optical measurement setup (composed by an actively stabilized fiber optic interferometer) was combined with an acoustic emission test system (namely two ad-hoc designed acoustic test benches in water and in soil). The optical measurement system is an all fiber homodyne Mach-Zehnder interferometer [5, 6, 7] schematically shown in Fig. 1. Taking into account cost and usability, the optical source is a Thorlabs LS5000 DWDM laser module. This laser source delivers 20mW optical power and is suited for optical communication in the C and L telecom bands (1530 – 1611 nm), where most commercial DAS read-out units work. The linewidth of this laser module is below 1 MHz, providing a coherence length in the order of 10 to 20 meters. This characteristic is taken into account in the design of the interferometer and choice of the length of the cable samples under test. The optical signal from the laser source is split into two fibers with a 3dB 1x2 coupler. The fiber optic sensor cable under test is placed in the measurement arm of the interferometer. A piezo-electric cylindrical transducer (PZT) with an optical fiber tightly spooled on is used in the reference arm as phase modulator for allowing homodyne detection. Since the optical fibers present in the reference arm of the interferometer are polarization maintaining fibers and the optical fibers present in the cables are non-polarizating fibers, a manual fiber polarization controller is introduced in the sensor arm to adjust the polarization in the system for optimal interference modulation and to compensate for spurious low-frequency drifts. The interference signal is then split through a 3dB 2x2 coupler into two light beams in counterphase and detected by a balanced amplified photodetector (PD), model Thorlabs PDB440C. The PD differential output, coupled into an integrated high-pass filter, is processed by a specifically developed spectral analysis software. The two separate monitor outputs of the two photodiodes are subtracted to one another in a difference amplifier to improve the signal-to-noise ratio and are used as input in the feedback loop.

A phase difference of !/4 between the sensing and the reference arms is needed to keep the system in quadrature condition. This is achieved by a feedback system, which uses the PZT actuator as a phase modulator. Since the system tends to be subject to low-frequency drifts due to temperature changes and environmental noise, an automatic feedback control circuit has also been implemented to improve detection stability.

Actuatordriver

Integrator

Lowpass

Diff.amplifier

Feedback system Photodetector PD1

Reference Fiber

PD2 Differential

Piezo-cylinder

Coupler Laser

Coupler Acquisition and control software

Polarizer to PC

Water tank AE waves Ref. Mic.

Sensor Fiber

Fig. 1: Schematic configuration of the all fiber homodyne Mach-Zehnder interferometer and acoustic test bench.

4.3 Test benches for the analysis of acoustic sensing cables Two test benches have been conceived with the purpose of simulating real application conditions, where the fiber optic acoustic sensing cables are immersed in a liquid environment or buried in the soil. • Test Bench 1: Water Tank The water tank has dimensions of 0.8 m x 0.55 m x 0.5 m. and is placed in a larger box filled with sand, to guarantee better acoustic isolation from environmental noise. • Test Bench 2: Sand Box The sand box has dimensions 1.0 m x 0.6 m x 0.5 m and is completely filled with compacted sand. To remove structure-borne sound disturbance, both boxes are placed on vibration absorbers. In either test bench, the fiber optic cables under investigation are placed into the test bench along with an acoustic emission source and a reference hydrophone, respectively geophone. Real application scenarios are simulated by acoustically stimulating the cables in the frequency range from 500 Hz to 50 kHz in Test Bench 1 (water tank) and between 20 Hz and 500 Hz for the cables in Test Bench 2 (buried in sand). 4.4 The water tank acoustic response analysis of immersed cables The test configuration is conceived to compare the acoustic sensing performance of different fiber optic cables while submersed in water and simultaneously normalize it to the response of a calibrated reference hydrophone, model Brüel & Kjaër 8103. The selected hydrophone is a small, high-sensitivity transducer providing absolute sound measurements over the frequency range 0.1 Hz to 180 kHz with a sensitivity of "211 dB re 1V/µPa. An omnidirectional transducer hydrophone, model Neptune Sonar D140, is used as acoustic emission source, with nominal bandwidth of 150 kHz. Particular attention is paid to the installation geometry of the different elements. When an acoustic wave of e.g. 50 kHz (which falls in the spectrum of interest) is propagating in water, where the sound velocity is 1481 m/s at 20°C, the acoustic wavelength is in the order of a few centimeters. Since the water tank has a dimension in the order of 1 meter in

each direction, larger than the stimulating wavelength, a uneven phase difference of the acoustic field would be present along the sensor occupying an arbitrary trajectory and position relative to the Acoustic Emission (AE) source in the water tank, eventually resulting in a reduction of the average of the dynamic pressure detected by the spatially integrated measurement, as performed on the cable samples. This effect randomly influences the evaluation of the integral acoustic sensitivity of the sensor. To circumvent this uncertainty, the omnidirectional acoustic source was placed in the center of a number of circular coils of cable of predefined radius (in the order of 10 cm). See the schematic arrangement in Figure 2. To avoid edge effect influences, the cable coil is located in the middle of the measurement box, as shown in Figure 3. In this way the fiber optic cable perceives nearly simultaneously the acoustical signal, inphase at each radial position.

Figure 2: Relative positioning of the acoustic source, sensor under test and calibration hydrophone (respectively geophone).

The total length of the sensing fiber is 10 meters in each optical sensor. This corresponds to 10 meters of the tested cable itself, for all designs in which the optical fibers are placed longitudinally in the sensor. The same situation applies, with some approximation, to designs in which the sensing fibers are stranded with a sufficient long lay length. The characteristics of our optoelectronic measurement setup are tuned to cope with the acoustic sensitivity of such sensor lengths. A divergent case is represented by the “BRUsens High-Resolution” sensing cable, in which the sensing optical fibers are wound with a lay length of some millimeters, so that 1 meter of cable contains 13 meters of fiber. For practical reasons (limited space in the test box) 10 meters of optical fiber was the standard test length, corresponding to (10 m / 13) # 0.8 m of high-resolution sensing cable. Furthermore the in-phase condition of the acoustic stimulus on the sensing optical fiber is not systematically assured in this case.

Fig. 3: Measurement setup in water. Several coils of cables under test and the omnidirectional transducer hydrophone as acoustic emitter are visible.

Fig. 4: Measurement setup in sand. The sound shaker driver is visible in the lower part of the picture. The cable samples and the reference geophone are buried in the soil.

4.5 The sand box for acoustic response analysis of buried cables The setup is configured to compare the acoustic sensitivity of different fiber optic cables embedded in granular solids. The acoustic sensitivity of the optical sensor is referenced to the responsivity of a calibrated geophone, model Aaronia Geo10. The selected geophone works in the frequency range 4 Hz - 1 kHz (typ.), with a sensitivity of 28.8 V/m/s. The spectral range of interest for in-soil acoustics in the present investigation extends between 20 Hz and 500 Hz. The speed of sound in sand is influenced by the granularity and compaction of the soil and is strongly nonlinearly dependent on the acoustic signal amplitude. A typical value of the speed of sound for low-amplitude vibrations in soil is found in

the literature [8] to be 110 m/s. In the present investigation the acoustic wavelength is typically larger than 1 meter. The acoustic source, a Sinustec St-Bs 250 Bass-Shaker, is located 0.5 m in front of the cable coil (approximated by a cylinder of 10 cm radius and less then 10 cm height) so that the acoustic signal reaches each cable section in phase. This situation provides for optimal integral acoustic response. An eventual drop in the acoustic cable sensitivity may appear in the higher spectral region. The cable under test is located 0.2 m from the wall of the stimulation box as shown in Figure 4.

5. MEASUREMENT RESULTS AND INTERPRETATION 5.1 Acoustic measurements in water The frequency response of the fiber optic cable samples selection presented in Section 3, Table 2 is analyzed in the spectral range between 500 Hz and 50 kHz of interest for DAS applications in water. As a first step the acoustic transfer function of a compact tight-buffered optical fiber is compared with that of the calibrated reference hydrophone, model Brüel & Kjaër 8103, both immersed in water. Afterwards the samples of the more complex cable designs are measured and their acoustic sensing performance is normalized to that of the tight-buffered fiber. 5.2 Fiber optic sensor reference and hydrophone calibration The diagram in Figure 5 shows the measured acoustic spectral response of 10 meters of tight-buffered single-mode optical fiber. Figure 6 shows the measured frequency response of the reference hydrophone to the same acoustic emission. The signal amplitude is expressed in logarithmic scale (dB) as Spectral Response (Hydrophone) = 20 x Log10 (Output Voltage / 1V). The black dots represent the signal amplitude as measured by the optical setup (homodyne interferometer and optoelectronic acquisition system) at each stimulation acoustic frequency. The squares represent the measurement noise level, mainly generated by external disturbance sources and fluctuation in the measurement system (temperature-induced polarization and phase drifts, amplitude drift in the laser source). The red triangles represent the level of the total harmonic distortion (THD) of the measurement setup expressed in %, at each frequency, as sensed the tight-buffered optical fiber (Fig. 5) or by the reference hydrophone (Fig. 6). The THD is measured with a resolution of 0.05%. The THD, defined as the ratio of the sum of the powers of all harmonic components to the power of the fundamental frequency, quantifies the effect of spectral distortions imputable to nonlinearities in the acoustic emitter, propagation medium, resonance box, acoustic sensor, and acquisition system.

!"#$%&'()&#*"+,*#)+-)095&+"0+,#)7,)8'%#&)

!"#$%&'()&#*"+,*#)+-)./01234#)+".$'()62#&)7,)8'%#&)

#!$

#"%&!,$ #"%&!+$

#$

#"%&!*$ #"%&!)$

!"#$

#"%&!($ #"%&!'$ !$

#!!!!$

-!!!!$

,!!!!$

+!!!!$

!"!#$ *!!!!$

D90B;