- Email: [email protected]
Borehole seismic survey using multimode optical fibers in a hybrid wireline
Accepted Manuscript Borehole seismic survey using multimode optical fibers in a hybrid wireline Gang Yu, Zhidong Cai, Yuanzhong Chen, Ximing Wang, Qinghong Zhang, Yanpeng Li, Yanhua Wang, Congwei Liu, Baoyin Zhao, Joe Greer PII: DOI: Reference:
S0263-2241(18)30339-7 https://doi.org/10.1016/j.measurement.2018.04.058 MEASUR 5459
To appear in:
Measurement
Received Date: Revised Date: Accepted Date:
19 July 2017 12 April 2018 16 April 2018
Please cite this article as: G. Yu, Z. Cai, Y. Chen, X. Wang, Q. Zhang, Y. Li, Y. Wang, C. Liu, B. Zhao, J. Greer, Borehole seismic survey using multimode optical fibers in a hybrid wireline, Measurement (2018), doi: https:// doi.org/10.1016/j.measurement.2018.04.058
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Borehole seismic survey using multimode optical fibers in a hybrid wireline Gang Yu1, Zhidong Cai1, Yuanzhong Chen1, Ximing Wang1, Qinghong Zhang1, Yanpeng Li1, Yanhua Wang1, Congwei Liu1, Baoyin Zhao2 and Joe Greer3 1 Bureau of Geophysical Prospecting (BGP) Inc., China National Petroleum Corporation (CNPC), Zhuozhou, Hebei, China, 2Jidong Oilfield, CNPC, Hebei, China, 3Silixa Ltd., Houston, TX, USA
Highlights
DAS uses optical fiber cable as distributed sensors for borehole seismic signals
DAS is a viable alternative to downhole geophone arrays for borehole seismic survey
It can dramatically reduce operating time required for a borehole seismic survey
It can achieve much higher spatial coverage than is typical of current technologies
Optical cable for borehole seismic survey will open DAS to much wider applications
Abstract Distributed Fiber Optic Sensing is increasingly recognized as a viable alternative to geophone arrays for the acquisition of borehole seismic data. The ability to deploy optical fibers into a well, either as a cable based intervention or as part of a completion string, allows for the entire wellbore to be surveyed with every source activation. This can dramatically reduce the operating time required to complete a normal survey as well as offering the opportunity to achieve much higher spatial coverage than is typical of current technologies. The ability to acquire borehole seismic data in a producing well without the need to disrupt production also offers significant benefits to the operator. Distributed acoustic sensing (DAS) is a novel technology that uses an optical fiber cable as a sensor for acoustic signals and can take almost any downhole fiber-optic installation or deployment and turn the fiber optic cable into a large downhole seismic array. This array can provide enhanced Vertical Seismic Profile (VSP) imaging and monitor fluids and pressures changes in the hydrocarbon production reservoir. Walkaway VSP data acquired over a former producing well in north eastern China provided a rich set of very high quality DAS 1
Walkaway VSP data. A standard VSP data pre-processing workflow was applied, followed by prestack Kirchhoff time migration. In the DAS pre-processing step we were faced with additional and special challenges: strong coherent noise due to cable slapping and ringing along the borehole casing. In comparison with an earlier offset VSP data set using 327-levels acquired with conventional 3C downhole geophones in the same well, the final pre-processed DAS walkaway VSP has a larger vertical aperture resulting in a wider lateral image. The single well DAS Walkaway VSP images provide a good result with higher vertical and lateral resolution than the surface seismic in the objective area. The vertical well environment without the ability to effectively “clamp” the sensor to the borehole casing wall by touching, creates a unique set of challenges. Although earth signal was recorded with almost all the shots, there was also a considerable amount of noise. Much of the noise was due to the physical placement of the wireline in the well and expressed by slapping and ringing. This paper reports on lessons learned in the handling of the wireline cable and subsequent special DAS data processing steps developed to remediate some of the practical wireline deployment issues. Optical wireline cable as a conveyance of fiber optic cables for VSP in vertical wells will open the use of the DAS system to much wider applications. Introduction Distributed Fiber Optic Sensing is increasingly recognized as a viable alternative to geophone arrays for the acquisition of borehole seismic data. The ability to deploy optical fibers into a well, either as a cable based intervention or as part of a completion string, allows for the entire wellbore to be surveyed with every source activation. This can dramatically reduce the operating time required to complete a normal survey as well as offering the opportunity to achieve much higher spatial coverage than is typical of current technologies. The ability to acquire borehole seismic data in a producing well without the need to disrupt production also offers significant benefits to the operator. Fiber optic Distributed Acoustic Sensing (DAS) is a new, fast growing technology that uses an optical fiber cable as a sensor for acoustic signals. Almost any downhole fiber-optic installation or temporary deployment can be turned into a large downhole seismic array for enhanced vertical seismic profile (VSP) imaging, monitoring of fluid and pressures changes in the hydrocarbon production reservoir. Seismic imaging and time-lapse reservoir monitoring give us critical information to guide the placement of production and water flood injection wells in our high-value reservoirs. Silixa’s intelligent distributed acoustic sensor (iDAS™) system measures the acoustic field along the fiber by sending a laser pulse into it and receiving the naturally occurring Rayleigh Backscatter from along the fiber. By analyzing this backscatter, measuring the time between the laser pulse being launched and the signal being received, the DAS can quantify the acoustic signal at all points along the fiber. An acoustic signal, coupled by friction or pressure to the optical fiber, induces dynamic strain changes along the cable. These strain changes lead to small displacements of the scattering elements and therefore to variations of the relative phases of the backscattered photons. The fiber behaves as a series of interferometers whose output is sensitive to small changes of the strain along its length caused by an acoustic signal [1]. The Distributed Acoustic Sensor system (iDAS™) records the acoustic field at every one meter along the length of the optical fiber in the well. The nature of this measurement means that the DA is able to measure and report the amplitude, frequency and phase of incident 2
energy including that associated with Seismic surveying. The phase accuracy allows for precise stacking and migrating of seismic data to enhance the signal to noise ratio, which is a key factor in the successful acquisition, and processing of high quality borehole seismic data. Stacking can be performed on signals from a single optical fiber over multiple source activations or on the response of multiple optical fibers at the same source activation. This project is for the acquisition of borehole seismic data using wireline conveyed fibers in an onshore well in China. Acquisition used shot holes drilled expressly for this VSP survey. Many distributed acoustic sensing VSP field tests have been published [2, 3, 4, 5, 6 and 7]. These case studies represent a wide range of different source types, source-receiver geometries and cable deployment methods (on tubing, behind casing). The use of fiber optic for VSP surveys has become a fairly well established application for Distributed Acoustic Sensing (DAS). Up until now, all commercial surveys reported have been on either permanently installed fibers behind the casing or carried tubing. Some research surveys have used fiber in a wireline clamped with magnets to the inside of the casing [8]. This report will describe the use of an unclamped hybrid wireline as the sensor. Recording seismic data using DAS instead of downhole geophones has advantages especially for VSP surveys. In wells with pre-existing optical cables for temperature or pressure measurements, DAS on-demand or time-lapse VSPs can be acquired without well intervention [5, 9]. In comparison to geophone recordings, the fiber optical cable covers the entire well and a full well VSP data can be recorded with a single shot. The distributed acoustic sensing VSP survey using the fiber optical cable does not need to move the tool string and repeat the shot [5]. This technique makes it economically feasible to interrogate multiple wells at the same source and full vertical coverage. Wells can be retrofitted with fiber optical cables by clamping it on tubing, pumping it inside tubing [5] or using hybrid fiber optical wireline cables [1, 3]. Most of the fiber DAS systems measure the phase change of Rayleigh backscattered light in fiber to record the acoustical wave. Assisted with the demodulation techniques including optical coherent detection, 3x3 coupler detection and phase generated carrier (PGC) detection, optical time domain reflectometry (OTDR) and optical frequency reflectometry (OFDR) based DASs are constructed. Up to now, using the single mode fiber (SMF), response frequency band from 20 Hz to 20 kHz has been realized in the 400 m long sensing section [10] and strain resolution of 80 nε has been demonstrated with the 3x3 phase demodulation scheme [11]. However, owing to the low signal to noise ratio (SNR) of Rayleigh backscattered light in SMF, further improvement at low-frequency range is hard to achieve. To enhance the SNR of the sensing signal light, FC/PC connector and ultra-weak fiber Bragg gratings (UWFBG) are used in the sensing fiber [12, 13]. Acoustical signal at 5 Hz is finally demodulated [14], however DAS sensing under 1 Hz has not been reported yet and the sensing performance at ultra-low range is not fully evaluated, of which the root lies in the random optical noise in the low frequency band. Walkaway VSP data acquired over a former producing well in north eastern China provided a rich set of very high quality DAS Walkaway VSP data. A standard VSP data pre-processing workflow was applied, followed by prestack Kirchhoff time migration. In the DAS preprocessing step we were faced with additional and special challenges: strong coherent noise due to cable slapping and ringing along the borehole casing. In comparison with an earlier offset VSP data set using 327-levels acquired with conventional 3C downhole geophones in the same well, the final pre-processed DAS walkaway VSP has a larger vertical aperture 3
resulting in a wider lateral image. The single well DAS Walkaway VSP images provide a good result with higher vertical and lateral resolution than the surface seismic in the objective area. The vertical well environment without the ability to effectively “clamp” the sensor to the borehole casing wall by touching, creates a unique set of challenges. Although earth signal was recorded with almost all the shots, there was also a considerable amount of noise. Much of the noise was due to the physical placement of the wireline in the well and expressed by slapping and ringing. According to noise analysis, we can obtain five parameters about the cable slapping and ringing noise. Those are first break time, first break amplitude, noise period, attenuation and average wavelet. The first 3 parameters can be measured directly, the fourth parameter can be obtained by curve fitting, and the last parameter can be obtained according to following steps: noise wavelet extraction, attenuation compensation, and statistical average. This paper reports on lessons learned in the handling of the wireline cable and subsequent special DAS data processing steps developed to remediate some of the practical wireline deployment issues. Optical wireline cable as a conveyance of fiber optic cables for VSP in vertical wells will open the use of the DAS system to much wider applications. DAS measurement mechanism Data acquired from distributed sensors is fundamentally different from data acquired from point sensors such as geophones, and the processing and analysis of such data benefits through being treated differently. Currently the majority of borehole seismic tools are constructed using geophones (sensors of electrical current generated by the motion of a coil in a magnetic field) that are idealized as sensitive to components of the local particle velocity for the medium at the point where the tool is clamped. The DAS interrogator measures, in a moving window, the relative strain between two sections of the fiber that are separated by a length dz, referred to here as the gauge length. The DAS response, by design, is linearly proportional to the average fiber elongation over the gauge length. The DAS optical signal processing is designed to extract, for each channel and each successive temporal sampling interval, the change in fiber strain with respect to the previous temporal sampling interval at that channel. In the DAS native data format, each digital sample is indexed by the center location of a moving window along a cable’s fiber core (the sample’s ‘channel’, z) and recording time (the sample’s ‘time’, t). Thus if u(z,t) represents the dynamic displacement of the fiber at axial location z and time t, the DAS output is a measure of:
dz dz dz dz u z + ,t + dt u z ,t + dt u z + ,t - u z - ,t 2 2 2 2
(1)
where dz and dt are the spatial gauge length and temporal sample interval respectively. As such, the DAS output can be equivalently regarded either as an estimate of the fiber strainrate:
u t z
(2)
or as an estimate of the spatial derivative of fiber particle velocity:
4
u z t
(3)
as calculated by difference operators applied in time or axial distance respectively. A measurement of strain is obtained from the DAS native strain-rate, since integration with respect to time converts strain-rate to strain (typically followed by a suitable temporal band pass filter). Moreover, for a propagating signal, integration with respect to distance is equivalent to integration with respect to time, followed by multiplication with the propagation speed.
Signal and noise When comparing signal and noise for data recorded by the DAS unit, the useable signal captured from the native output is limited by broadband noise that is inherent in the optical scattering process. Because the system response is linear and coherent with respect to the dynamic local strain, repeated stacking of DAS traces over repeated shots may be expected to result in a SNR improvement following the usual inverse square root relationship between SNR and number of stacks. However, in contrast with the case for geophone sensors, analysis of the DAS optical scattering using advanced techniques allows a significantly more optimal stacking to be performed, which substantially exceed the SNR improvement provided by simple averaging. It is accepted that when data consists of a set of measurements of common signal plus uncorrelated noise with known noise power, a weighted-mean stack gives a significantly higher SNR than a simple mean stack, given that the optimal weights are inversely proportional to the noise power. In practice, this reduces the problem of determining optimal stacking weights to the problem of estimating noise power. Exploiting detailed knowledge of the DAS scattering processes and the opto-acoustic demodulation carried out within the DAS, Silixa has developed a proprietary method to track the noise power and thereby accomplish optimal stacking. By spatially oversampling to obtain a better picture noise statistics, noise-optimal stacking within a single shot is achieved without degrading the signal characteristics. The DAS samples optical data at 25cm and then uses noise-adaptive down sampling to 2m, as part of the optical-acoustic conversion process, to produce the raw acoustic data delivered to the customer. In addition, Silixa provides de-noising of the raw acoustic data as an optional post-processing stage. The de-noising process uses the understanding that neighboring channels are exposed to a phase-shifted version of the same signal. In a perfect noise-free situation, this means that each channel can be represented as a linear combination of its neighboring channels, with the application of appropriate phase correction. De-noising works by finding the optimal complex weights to minimize the noise, and then uses these weights to compute the noiseoptimized value at each channel. This complex channel averaging takes care of the delay information and so increases SNR with, if applied correctly, minimal distortion of the signal.
5
Figure 1. The left hand panel shows a single shot of raw acoustic data; the right hand panel showing the same shot with de-noising applied. Conversion to geophone equivalent signal In some cases, it is useful to make a comparison of the DAS response to that of industry ‘standard’ geophones. Since the DAS recording is a local strain-rate or local strain (depending on whether or not the data is time-integrated), comparison to a geophone, which measures particle velocity at a fixed point, requires conversion. For propagating seismic signals, such as a harmonic plane wave, strain, displacement and particle velocity are related as follows: For εzz = extensional strain in the z direction, and uz = displacement in the z direction with velocity c, where uz = U e-iω (t-z/c) and vz = duz/dt = U (-iω) e-iω (t-z/c) is the axial particle velocity, then εzz = duz/dz = vz/c. However the relationship is more general and applies to any propagating disturbance with a stable phase function. For dynamic fiber displacement, a stably coupled propagating disturbance will be self-similar under suitable translation in space and time; that is, it will take the form u(z,t) = u(ϕ) where ϕ = (t0 + t ± z/c) is a characteristic phase function with propagation speed c. Differentiating with respect to time and distance respectively, the fibre particle velocity is:
v
u u t
(4)
and the fiber strain:
u u 1 / c (5) z
Comparing these equations, then = ± ; that is, the ratio between fiber particle velocity and fiber strain is given by the propagation speed along the fiber cable (apparent velocity) with a sign determined by the direction of propagation. Therefore, the time-integrated DAS signal is converted to equivalent geophone signal by multiplying the dimensionless strain by the local propagation speed (as determined from VSP moveout data).
6
DAS Walkaway VSP data acquisition During a survey in a formerly producing CNPC well in north-eastern China, near Tangshan, Hebei [15], walkaway vertical seismic profile (VSP) data were acquired using Silixa’s iDASTM system. The primary objective was to evaluate the velocity and anisotropy characteristics for use in a surface seismic survey. A secondary objective was to investigate the overall performance and operational feasibility of using DAS for VSP surveys. In January 2015, BGP and Silixa shot a Walkaway VSP survey as part of a larger commercial 3D surface seismic campaign in a 4,004 m vertical well in north eastern China. This 6 year old oil producing well was plugged back above its perforations. A total of 386 dynamite shot points were arranged into one in-line Walkaway and one cross-line Walkaway. Figure 4 shows the overall shot line layout, each shot position and its corresponding shot number (top) and receiving cable deployment (bottom). The surface conditions are characterized by a single narrow causeway surrounded by shallow fishponds. A denser shot Walkaway line was placed on the causeway due to ease of placing charges. Two multimode 62.5 μm fibers which make up part of a commercially available 5,000 m hybrid optical-electric wireline cable were spliced together at the cable head to form a continuous 10 km down then up-going distributed sensor. This geometry gave two side-byside stackable fiber channels. During conventional use of this wireline cable, these multimode fibers are for communications with downhole logging tools. The DAS interrogator, triggered by a Real Time System source controller, was used to record 6 kg dynamite shots. Recording was setup to measure 0.25m intervals at 1 ms sample rate along 8 km of the fiber optic cable. The first and last portions stayed on the cable drum. Data acquisition took place over the course of 5 days, working daylight hours, with 386 successful shot points acquired [16].
Figure 2. The two panels show data from a single shot, before (left panel) and after (right panel) the application of the velocity transform. The hybrid wireline cable was rigged up as illustrated in Figure 4. The significant point is that this is a standard wireline rig up with only a weight at the bottom and normal depth 7
control procedures. In future surveys, there would be nothing to prevent the use of standard logging tools before or after the seismic survey. The electrical conductors and two of the fibers were still available.
Figure 3. DAS Walkaway VSP shot line layout (left) and receiving hybrid wireline cable deployment (right).
8
Figure 4. Distributed Acoustic Sensing VSP survey receiving hybrid electric-optic wireline configuration Diagram. DAS Walkaway VSP data analysis and processing Field data was saved in the contractor’s proprietary format. Afterward, the data was converted to SEG-Y format (Society of Exploration Geophysicists Y Format) using acoustic conversion algorithms that involve noise-adaptive resampling to reduce the spatial sampling from 0.25m to 2m. The process also includes the straight channel alignment and stacking of the VSP data recorded through the two legs of fiber in the well. As a result, the signal to noise ratio is greatly improved by removing unwanted system generated optical noise. The noise suppression process provides us with a 2,000-level Walkaway VSP data set covering the full well depth and does not distort the seismic signal measured by DAS. The physical property measured by DAS is the rate of change of strain (strain rate) within the fiber. The units within DAS measurement files and the delivered SEG-Ys are proportional to this quantity. During the preparation of the final data set, the section of the data covering channels within the well was extracted. This is channels between the Blowout Preventer (BOP), the bottom of the well and back to the BOP again. The measured depth relative to 9
ground level of each channel was calculated using the field recorded depth QC measurements (tap test) and comparison with the well completion diagram. The receivers corresponding to the BOP were assigned a depth of 1.4 m above ground level; the bottom receiver was assigned a depth of 4,004 m below ground level. A linear interpolation was applied between these positions to assign a depth to each receiver. In the pre-processing step, we were faced with additional and special challenges: strong coherent noise due to cable slapping and ringing along the borehole casing. A special patent pending algorithm and procedure were developed to remove the strong coherent cable slapping and ringing noise (Figure 5).
Figure 5. DAS Walkaway VSP raw data (left) and cable slapping and ringing noise removing results (right) comparison. Following is the detailed description of the data analysis and the coherent noise suppression method: 1. Wavefield analysis In a single -VSP shot record, we can recognize several waves, such as downgoing and upgoing P-wave, but we also find that there are strong coherent cable slapping and ringing noises in the record (Figure 6a). The analysis shows that: 1. The downgoing P-wave is clearly visible, but its first break is not clear; 2. The upgoing P-wave is clearly visible, but it is disturbed by strong coherent noise; 3. The downgoing and upgoing P-S wave is visible; 4. There are few multiple waves; 5. Cable slapping and ringing noise is the main disturbance.
10
(a)
(b)
(c) Figure 6. Raw VSP data (a) recorded by the DAS acquisition system and the hybrid wireline, synthetic cable slapping and ringing noise wavefield (b), and cable slapping and ringing noise removed DAS Walkaway VSP wavefield (c). 2. Noise period analysis We analyze the noise period of regular cable slapping and ringing noise by statistic method as follows: For each seismic trace, the maximum time ( Tmax ) of recognizable period is read firstly, and then maximum period numbers ( nmax ) are counted. Finally, we calculate a single noise period ( T ), it is expressed as follow:
T Tmax / nmax
(6) Significantly, there are more than one noise period in some traces. 3. Noise wavelet analysis
11
In each seismic trace of raw wavefield, wavelet extraction includes the following steps: finding the noise wavelet ( Wi ) in each period, calculating these wavelets to obtain '
normalized wavelets ( Wi ), then stacking all wavelets to form an average wavelet ( Wave ), it is expressed as follow:
Wave 1 / nmax i max Wi ' 1 n
(7)
Because some traces have more than one noise periods, then these traces have different noise wavelets. 4. Noise attenuation characteristics analysis Cable slapping and ringing noise travels up and down in optic fiber, and its attenuation characteristics are different than that of seismic wave attenuation. The attenuation characteristics are dependent on optic fiber material and its length. According to many tests and detailed analysis, we proved that the best fitting method of the attenuation curve was not the exponential fitting, but the base fitting. When the noise travels a time ( t ), the raw amplitude ( A0 ) at the position of its first break is decreased to new amplitude ( At ), it is expressed as follow:
At At xt
(8)
t
Where x is the attenuation factor, and x is the base number of the factor. It can be calculated by attenuation curve statistic method (Figure 7).
Figure 7: Attenuation curve and the average curve obtained through attenuation curve statistic method. For the coherent cable slapping and ringing noise suppression method, following steps were adopted:
12
Figure 8. Illustration of noise fitting process. Using the formula (4) to each seismic trace, the cable slapping and ringing noise wavefield is synthesized (Figure 6b). 1. Noise fitting According to the above analysis, the cable slapping and ringing noise ( N ) can be fitted by first break time ( T0 ), first break amplitude ( A0 ), noise period ( T ), average wavelet (Wave ) t
and attenuation parameter ( x ) (Figure 8).
N i max A0 x 0 1 n
T i 1T
*Wave
(9)
2. Noise reduction The raw wavefield ( S ) is subtracted by the synthetic cable slapping and ringing noise wavefield ( N ), and then leaves noise removed DAS Walkaway VSP wavefield ( S res ) (Figure 4c). Sres S N (10) Comparing the wavefield in Figure 6a and Figure 6c, obviously, cable slapping and ringing noise is effectively suppressed, and the upgoing wave is clearer in the noise removed DAS Walkaway VSP wavefield ( S res ). 13
Figure 9. Autocorrelation data of raw wavefield (top) and noise removed wavefield (bottom). After DAS fibre channel alignment and stacking of VSP data recorded through the two legs of fibre in the well, a typical VSP processing sequence was applied: trace editing, noise attenuation, spherical divergence correction, up/down separation and up-down deconvolution. Figure 10 shows the comparison of common source gather (1,000 m offset) of the DAS Walkaway VSP data before and after wavefield separation processing. 3. Quality Control It is well known that autocorrelation is an effective checking method for the regular coherent noise suppression. Figure 9 shows two autocorrelation datasets of raw wavefield and noiseremoved wavefield. By comparison, the side lobe of noise removed wavefield autocorrelation is much less than that of the raw wavefield, the suppression method is proved to be effective and feasible. DAS Walkaway VSP data imaging After up/down wave-field separation, we first perform the VSP data coordinate transformation through wave-field continuation. Then based on the theory of bending rays and preserved amplitude weight function formula, a stable ray tracing was used to establish, numerically, the relationship between time-distance curve of the reflected wave and offset angle of incidence mapping. Finally, high precision common imaging point (CIP) was used in gathering VSP-CDP (Common Depth Point) imaging and pre-stack Kirchhoff time migration (PSTM) to image the up going reflection VSP data.
14
Figure 10. Comparison of DAS Walkaway VSP raw data (left) and wavefield separation results (right). Method 1. VSP-CDP imaging VSP-CDP imaging is the basic method of VSP imaging, among the improve methods, the ray tracing imaging is the most representative method [17, 18], its steps as follow:
Velocity model building. According to the zero-offset VSP data and the 2-D or 3-D seismic velocity, the velocity model is built. Ray tracing. According to the velocity model, the location and travel time of reflect point is calculated. Coordinate calculating. The coordinate of reflect point is calculated by using interpolation method. VSP data assignment. For each trace of VSP data, according to the above steps, each sample is assigned the calculated coordinate. Imaging gridded. Given a value of group interval and gridded the imaging area, and then put the VSP data into the correspondent locations. VSP-CDP imaging. Time-depth transformed is the last step. Figure 11 shows the final VSP imaging.
Method 2. Pre-stack Kirchhoff time migration The pre-stack Kirchhoff time migration of VSP is based on Huygens's principle [19].
I r, t 0
1 cos U V dA x , t U x , t 0 0 0 R 2π RV t R t0 t
(11)
V
Among the function, A0 is the locations of sources and geophones in earth's surface, is the angle of emergence of the ray form imaging point to a receiver, R is the distance of seismic wave transmission from reflection point to receiver, cos z / R is the declination factor 15
( z is the depth of imaging point), it shows the amplitude changes with the emergence angle when the scattering point is as a secondary source, t R / V is the delay time, t0 is the reflection time from source to receiver, and V is the transmission velocity of seismic wave.
Figure 11. Surface seismic profile (left) and DAS Walkaway VSP image (VSP-CDP mapping) insert into surface seismic profile (right) along the Walkaway VSP survey line with denser VSP shots. The edge of VSP image is muted.
Figure 12. Surface seismic profile (left) and DAS Walkaway VSP image (VSP-CDP mapping) insert into surface seismic profile (right) along the Walkaway VSP survey line with coarser VSP shots. Figures 11 shows the surface seismic section along the North East Walkaway VSP line that has higher shot density and the VSP image inserted into surface seismic profile. The surface seismic image along the North West Walkaway VSP line with coarser shot density and its VSP image are inserted into surface seismic section as shown in Figure 12. Both Walkaway
16
VSP images not only match well with surface seismic image they also show higher vertical resolution and more detailed structures. Conclusions A Walkaway VSP survey acquired with fiber optics over a former producing well in northeastern China provided very large aperture Walkaway VSP data. Strong coherent noise due to cable slapping and ringing along the borehole casing created special challenges in the preprocessing step that was overcoming by designing a new and effective noise removal algorithm. Afterward, standard VSP data pre-processing workflow could be applied, followed by pre-stack Kirchhoff time migration. The final pre-processed DAS data has a larger vertical aperture than previous VSP with geophones resulting in a wider lateral image from a single well. The single well DAS Walkaway VSP images provide a good result with higher vertical and lateral resolution as well as more detailed structures than the surface seismic in the objective area. The vertical well environment without the ability to effectively clamp the sensor to the borehole casing wall creates a unique set of challenges. Although earth signal was recorded with almost all the shots, there was also a considerable amount of noise. Much of the noise was due to the way the wireline was physically placed in the well and expressed by slapping and ringing. Among the lessons learned on handling the hybrid wireline cable during acquisition were to fill the top part of the well with water, stabilize the hybrid wireline between wellhead and recording truck, run some slack, use a T-Bar at the wellhead to hold the downhole hybrid wireline cable and use smaller dynamite charges on near offset shots. Special DAS data processing steps were developed to remediate some of the practical wireline deployment issues. Optical wireline cable as a conveyance of fiber optic cables for VSP in vertical wells will open the use of the DAS system to much wider applications. Acknowledgements We would like to thank Jidong Oilfiled and CNPC for allowing the publication of this article. We would also like to thank BGP and Silixa management for their support and guidance during the data acquisition, processing and interpretation phases of this project. This study was funded by the China 13th Five Year Plan DAS Borehole Seismic Project (2016ZX05018004-003) and China National Natural Science Fund Grants 41527805. References [1] Frignet, B. and Hartog, A., 2014, Optical vertical seismic profile on wireline cable. SPWLA 55th Annual Logging Symposium, Abu Dhabi, May 18-22. [2] Daley, T., Freifeld, B., Ajo-Franklin, J., Dou, S., Pevzner, R., Shulakova, V., Kashikar, S., Miller, D., Götz, J., Henninges, J. and Lüth, S., 2013, Field testing of fiber-optic distributed acoustic sensing (DAS) for subsurface seismic monitoring. The Leading Edge, 32(6), 699706. [3] Hartog, A., Frignet, B., Mackie, D. and Clark, M., 2014, Vertical seismic optical profiling on wireline logging cable. Geophysical Prospecting, 62(4), 693-701. [4] Madsen, K., Thompson, M., Parker, T. and Finfer, D., 2013, A VSP field trial using distributed acoustic sensing in a producing well in the North Sea. First break, 31, 51-56. 17
[5] Mateeva, A., Lopez, J., Potters, H., Mestayer, J., Cox, B., Kiyashchenko, D., Wills, P., Grandi, S., Hornman, K., Kuvshinov, B., Berlang, W., Yang, Z. and Detomo, R., 2014, Distributed acoustic sensing for reservoir monitoring with vertical seismic profiling. Geophysical Prospecting, 62, 679- 692. [6] Parker, T., Shatalin, S. and Farhadiroushan, M., 2014, Distributed Acoustic Sensing – a new tool for seismic applications，First Break, 32, 61-69. [7] Lesnikov, V., Allanic, C., 2014, DAS VSP acquisition - perspectives and challenges: 76th EAGE Conference and Exhibition - Workshops, [8] La Follett, J., Wyker, B., Hemink, G., Hornman, K., Lumens, P., Franzen, A., 2014, Evaluation of fiber-optic cables for use in Distributed Acoustic Sensing: commercially available cables and novel cable designs: 84th Annual International Meeting, SEG, Expanded Abstracts. [9] Mateeva, A., J. Mestayer, B. Cox, D. Kiyashchenko, P. Wills, J. Lopez, S. Grandi, K. Hornman, P. Lumens, A. Franzen, D. Hill, and J. Roy, 2012, Advances in distributed acoustic sensing (DAS) for VSP: 82nd Annual International Meeting, SEG, Expanded Abstracts. [10] He, X. G., Xie, S. G, Liu, F., Cao, S., Gu, L. J., Zheng, X. P. and Zhang, M., 2017, Multi-event waveform-retrieved distributed optical fiber acoustic sensor using dual-pulse heterodyne phase-sensitive OTDR. Optics Letters, 42, 442- 445. [11] Masoudi, A., Belal, M. and Newson, T.P., 2013, A distributed optical fibre dynamic strain sensor based on phase-OTDR, Measurement Science and Technology, 24, 085204. [12] Steinberg, I., Shiloh, L., Gabai, H. and Eyal, A., 2015, Over 100km long ultra-sensitive dynamic sensing via Gated-OFDR, International Conference on Optical Fibre Sensors, Curitiba, Brazil, 96341B. [13] Wang, C., Shang, Y., Liu, X. H., Wang, C., Yu, H. H, Jiang, D. S. and Peng, G. D., 2015, Distributed OTDR-interferometric sensing network with identical ultra-weak fiber Bragg gratings. Optics Express, 23, 22, 29038- 29046. [14] Zhu, F., Zhang, Y. X., Xia, L., Wu, X. L. and Zhang, X. P., 2015, Improved Φ-OTDR Sensing System for High-Precision Dynamic Strain Measurement Based on Ultra-Weak Fiber Bragg Grating Array. Journal of Lightwave Technology, 33, 23, 4775- 4780. [15] Yu, G., Li, Y. P., Wang, X. M., Greer J., Clarke A., 2015, Deep vertical seismic profiling using multimode optical fibers in a hybrid wireline - lessons learned: SPE Fiber Optic Sensing Workshop, Napa, CA, USA. [16] Yu, G., Cai, Z. D., Chen, Y. Z., Wang, X. M., Zhang, Q. H., Li, Y. P., Wang, Y. H., Liu, C. W., Zhao, B. Y. and Greer, J., 2016, Walkaway VSP using multimode optical fibers in a hybrid wireline, The Leading Edge, July 2016, 936-940. [17] Chen G., Canales L. and Peron J., 1999, VSP mapping error for dipping horizons using a horizontally layered model. 69th SEG Annual Meeting. 18
[18] Chen G., Peron J. and Canales L., 2000, Rapid VSP-CDP mapping of 3-D VSP data. Geophysics, 65(5), 1631-1640. [19] Schneider, W. 1978, Integral formulation for migration in two and three dimensions. Geophysics, 43(1), 49~76
19
S0263-2241(18)30339-7 https://doi.org/10.1016/j.measurement.2018.04.058 MEASUR 5459
To appear in:
Measurement
Received Date: Revised Date: Accepted Date:
19 July 2017 12 April 2018 16 April 2018
Please cite this article as: G. Yu, Z. Cai, Y. Chen, X. Wang, Q. Zhang, Y. Li, Y. Wang, C. Liu, B. Zhao, J. Greer, Borehole seismic survey using multimode optical fibers in a hybrid wireline, Measurement (2018), doi: https:// doi.org/10.1016/j.measurement.2018.04.058
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Borehole seismic survey using multimode optical fibers in a hybrid wireline Gang Yu1, Zhidong Cai1, Yuanzhong Chen1, Ximing Wang1, Qinghong Zhang1, Yanpeng Li1, Yanhua Wang1, Congwei Liu1, Baoyin Zhao2 and Joe Greer3 1 Bureau of Geophysical Prospecting (BGP) Inc., China National Petroleum Corporation (CNPC), Zhuozhou, Hebei, China, 2Jidong Oilfield, CNPC, Hebei, China, 3Silixa Ltd., Houston, TX, USA
Highlights
DAS uses optical fiber cable as distributed sensors for borehole seismic signals
DAS is a viable alternative to downhole geophone arrays for borehole seismic survey
It can dramatically reduce operating time required for a borehole seismic survey
It can achieve much higher spatial coverage than is typical of current technologies
Optical cable for borehole seismic survey will open DAS to much wider applications
Abstract Distributed Fiber Optic Sensing is increasingly recognized as a viable alternative to geophone arrays for the acquisition of borehole seismic data. The ability to deploy optical fibers into a well, either as a cable based intervention or as part of a completion string, allows for the entire wellbore to be surveyed with every source activation. This can dramatically reduce the operating time required to complete a normal survey as well as offering the opportunity to achieve much higher spatial coverage than is typical of current technologies. The ability to acquire borehole seismic data in a producing well without the need to disrupt production also offers significant benefits to the operator. Distributed acoustic sensing (DAS) is a novel technology that uses an optical fiber cable as a sensor for acoustic signals and can take almost any downhole fiber-optic installation or deployment and turn the fiber optic cable into a large downhole seismic array. This array can provide enhanced Vertical Seismic Profile (VSP) imaging and monitor fluids and pressures changes in the hydrocarbon production reservoir. Walkaway VSP data acquired over a former producing well in north eastern China provided a rich set of very high quality DAS 1
Walkaway VSP data. A standard VSP data pre-processing workflow was applied, followed by prestack Kirchhoff time migration. In the DAS pre-processing step we were faced with additional and special challenges: strong coherent noise due to cable slapping and ringing along the borehole casing. In comparison with an earlier offset VSP data set using 327-levels acquired with conventional 3C downhole geophones in the same well, the final pre-processed DAS walkaway VSP has a larger vertical aperture resulting in a wider lateral image. The single well DAS Walkaway VSP images provide a good result with higher vertical and lateral resolution than the surface seismic in the objective area. The vertical well environment without the ability to effectively “clamp” the sensor to the borehole casing wall by touching, creates a unique set of challenges. Although earth signal was recorded with almost all the shots, there was also a considerable amount of noise. Much of the noise was due to the physical placement of the wireline in the well and expressed by slapping and ringing. This paper reports on lessons learned in the handling of the wireline cable and subsequent special DAS data processing steps developed to remediate some of the practical wireline deployment issues. Optical wireline cable as a conveyance of fiber optic cables for VSP in vertical wells will open the use of the DAS system to much wider applications. Introduction Distributed Fiber Optic Sensing is increasingly recognized as a viable alternative to geophone arrays for the acquisition of borehole seismic data. The ability to deploy optical fibers into a well, either as a cable based intervention or as part of a completion string, allows for the entire wellbore to be surveyed with every source activation. This can dramatically reduce the operating time required to complete a normal survey as well as offering the opportunity to achieve much higher spatial coverage than is typical of current technologies. The ability to acquire borehole seismic data in a producing well without the need to disrupt production also offers significant benefits to the operator. Fiber optic Distributed Acoustic Sensing (DAS) is a new, fast growing technology that uses an optical fiber cable as a sensor for acoustic signals. Almost any downhole fiber-optic installation or temporary deployment can be turned into a large downhole seismic array for enhanced vertical seismic profile (VSP) imaging, monitoring of fluid and pressures changes in the hydrocarbon production reservoir. Seismic imaging and time-lapse reservoir monitoring give us critical information to guide the placement of production and water flood injection wells in our high-value reservoirs. Silixa’s intelligent distributed acoustic sensor (iDAS™) system measures the acoustic field along the fiber by sending a laser pulse into it and receiving the naturally occurring Rayleigh Backscatter from along the fiber. By analyzing this backscatter, measuring the time between the laser pulse being launched and the signal being received, the DAS can quantify the acoustic signal at all points along the fiber. An acoustic signal, coupled by friction or pressure to the optical fiber, induces dynamic strain changes along the cable. These strain changes lead to small displacements of the scattering elements and therefore to variations of the relative phases of the backscattered photons. The fiber behaves as a series of interferometers whose output is sensitive to small changes of the strain along its length caused by an acoustic signal [1]. The Distributed Acoustic Sensor system (iDAS™) records the acoustic field at every one meter along the length of the optical fiber in the well. The nature of this measurement means that the DA is able to measure and report the amplitude, frequency and phase of incident 2
energy including that associated with Seismic surveying. The phase accuracy allows for precise stacking and migrating of seismic data to enhance the signal to noise ratio, which is a key factor in the successful acquisition, and processing of high quality borehole seismic data. Stacking can be performed on signals from a single optical fiber over multiple source activations or on the response of multiple optical fibers at the same source activation. This project is for the acquisition of borehole seismic data using wireline conveyed fibers in an onshore well in China. Acquisition used shot holes drilled expressly for this VSP survey. Many distributed acoustic sensing VSP field tests have been published [2, 3, 4, 5, 6 and 7]. These case studies represent a wide range of different source types, source-receiver geometries and cable deployment methods (on tubing, behind casing). The use of fiber optic for VSP surveys has become a fairly well established application for Distributed Acoustic Sensing (DAS). Up until now, all commercial surveys reported have been on either permanently installed fibers behind the casing or carried tubing. Some research surveys have used fiber in a wireline clamped with magnets to the inside of the casing [8]. This report will describe the use of an unclamped hybrid wireline as the sensor. Recording seismic data using DAS instead of downhole geophones has advantages especially for VSP surveys. In wells with pre-existing optical cables for temperature or pressure measurements, DAS on-demand or time-lapse VSPs can be acquired without well intervention [5, 9]. In comparison to geophone recordings, the fiber optical cable covers the entire well and a full well VSP data can be recorded with a single shot. The distributed acoustic sensing VSP survey using the fiber optical cable does not need to move the tool string and repeat the shot [5]. This technique makes it economically feasible to interrogate multiple wells at the same source and full vertical coverage. Wells can be retrofitted with fiber optical cables by clamping it on tubing, pumping it inside tubing [5] or using hybrid fiber optical wireline cables [1, 3]. Most of the fiber DAS systems measure the phase change of Rayleigh backscattered light in fiber to record the acoustical wave. Assisted with the demodulation techniques including optical coherent detection, 3x3 coupler detection and phase generated carrier (PGC) detection, optical time domain reflectometry (OTDR) and optical frequency reflectometry (OFDR) based DASs are constructed. Up to now, using the single mode fiber (SMF), response frequency band from 20 Hz to 20 kHz has been realized in the 400 m long sensing section [10] and strain resolution of 80 nε has been demonstrated with the 3x3 phase demodulation scheme [11]. However, owing to the low signal to noise ratio (SNR) of Rayleigh backscattered light in SMF, further improvement at low-frequency range is hard to achieve. To enhance the SNR of the sensing signal light, FC/PC connector and ultra-weak fiber Bragg gratings (UWFBG) are used in the sensing fiber [12, 13]. Acoustical signal at 5 Hz is finally demodulated [14], however DAS sensing under 1 Hz has not been reported yet and the sensing performance at ultra-low range is not fully evaluated, of which the root lies in the random optical noise in the low frequency band. Walkaway VSP data acquired over a former producing well in north eastern China provided a rich set of very high quality DAS Walkaway VSP data. A standard VSP data pre-processing workflow was applied, followed by prestack Kirchhoff time migration. In the DAS preprocessing step we were faced with additional and special challenges: strong coherent noise due to cable slapping and ringing along the borehole casing. In comparison with an earlier offset VSP data set using 327-levels acquired with conventional 3C downhole geophones in the same well, the final pre-processed DAS walkaway VSP has a larger vertical aperture 3
resulting in a wider lateral image. The single well DAS Walkaway VSP images provide a good result with higher vertical and lateral resolution than the surface seismic in the objective area. The vertical well environment without the ability to effectively “clamp” the sensor to the borehole casing wall by touching, creates a unique set of challenges. Although earth signal was recorded with almost all the shots, there was also a considerable amount of noise. Much of the noise was due to the physical placement of the wireline in the well and expressed by slapping and ringing. According to noise analysis, we can obtain five parameters about the cable slapping and ringing noise. Those are first break time, first break amplitude, noise period, attenuation and average wavelet. The first 3 parameters can be measured directly, the fourth parameter can be obtained by curve fitting, and the last parameter can be obtained according to following steps: noise wavelet extraction, attenuation compensation, and statistical average. This paper reports on lessons learned in the handling of the wireline cable and subsequent special DAS data processing steps developed to remediate some of the practical wireline deployment issues. Optical wireline cable as a conveyance of fiber optic cables for VSP in vertical wells will open the use of the DAS system to much wider applications. DAS measurement mechanism Data acquired from distributed sensors is fundamentally different from data acquired from point sensors such as geophones, and the processing and analysis of such data benefits through being treated differently. Currently the majority of borehole seismic tools are constructed using geophones (sensors of electrical current generated by the motion of a coil in a magnetic field) that are idealized as sensitive to components of the local particle velocity for the medium at the point where the tool is clamped. The DAS interrogator measures, in a moving window, the relative strain between two sections of the fiber that are separated by a length dz, referred to here as the gauge length. The DAS response, by design, is linearly proportional to the average fiber elongation over the gauge length. The DAS optical signal processing is designed to extract, for each channel and each successive temporal sampling interval, the change in fiber strain with respect to the previous temporal sampling interval at that channel. In the DAS native data format, each digital sample is indexed by the center location of a moving window along a cable’s fiber core (the sample’s ‘channel’, z) and recording time (the sample’s ‘time’, t). Thus if u(z,t) represents the dynamic displacement of the fiber at axial location z and time t, the DAS output is a measure of:
dz dz dz dz u z + ,t + dt u z ,t + dt u z + ,t - u z - ,t 2 2 2 2
(1)
where dz and dt are the spatial gauge length and temporal sample interval respectively. As such, the DAS output can be equivalently regarded either as an estimate of the fiber strainrate:
u t z
(2)
or as an estimate of the spatial derivative of fiber particle velocity:
4
u z t
(3)
as calculated by difference operators applied in time or axial distance respectively. A measurement of strain is obtained from the DAS native strain-rate, since integration with respect to time converts strain-rate to strain (typically followed by a suitable temporal band pass filter). Moreover, for a propagating signal, integration with respect to distance is equivalent to integration with respect to time, followed by multiplication with the propagation speed.
Signal and noise When comparing signal and noise for data recorded by the DAS unit, the useable signal captured from the native output is limited by broadband noise that is inherent in the optical scattering process. Because the system response is linear and coherent with respect to the dynamic local strain, repeated stacking of DAS traces over repeated shots may be expected to result in a SNR improvement following the usual inverse square root relationship between SNR and number of stacks. However, in contrast with the case for geophone sensors, analysis of the DAS optical scattering using advanced techniques allows a significantly more optimal stacking to be performed, which substantially exceed the SNR improvement provided by simple averaging. It is accepted that when data consists of a set of measurements of common signal plus uncorrelated noise with known noise power, a weighted-mean stack gives a significantly higher SNR than a simple mean stack, given that the optimal weights are inversely proportional to the noise power. In practice, this reduces the problem of determining optimal stacking weights to the problem of estimating noise power. Exploiting detailed knowledge of the DAS scattering processes and the opto-acoustic demodulation carried out within the DAS, Silixa has developed a proprietary method to track the noise power and thereby accomplish optimal stacking. By spatially oversampling to obtain a better picture noise statistics, noise-optimal stacking within a single shot is achieved without degrading the signal characteristics. The DAS samples optical data at 25cm and then uses noise-adaptive down sampling to 2m, as part of the optical-acoustic conversion process, to produce the raw acoustic data delivered to the customer. In addition, Silixa provides de-noising of the raw acoustic data as an optional post-processing stage. The de-noising process uses the understanding that neighboring channels are exposed to a phase-shifted version of the same signal. In a perfect noise-free situation, this means that each channel can be represented as a linear combination of its neighboring channels, with the application of appropriate phase correction. De-noising works by finding the optimal complex weights to minimize the noise, and then uses these weights to compute the noiseoptimized value at each channel. This complex channel averaging takes care of the delay information and so increases SNR with, if applied correctly, minimal distortion of the signal.
5
Figure 1. The left hand panel shows a single shot of raw acoustic data; the right hand panel showing the same shot with de-noising applied. Conversion to geophone equivalent signal In some cases, it is useful to make a comparison of the DAS response to that of industry ‘standard’ geophones. Since the DAS recording is a local strain-rate or local strain (depending on whether or not the data is time-integrated), comparison to a geophone, which measures particle velocity at a fixed point, requires conversion. For propagating seismic signals, such as a harmonic plane wave, strain, displacement and particle velocity are related as follows: For εzz = extensional strain in the z direction, and uz = displacement in the z direction with velocity c, where uz = U e-iω (t-z/c) and vz = duz/dt = U (-iω) e-iω (t-z/c) is the axial particle velocity, then εzz = duz/dz = vz/c. However the relationship is more general and applies to any propagating disturbance with a stable phase function. For dynamic fiber displacement, a stably coupled propagating disturbance will be self-similar under suitable translation in space and time; that is, it will take the form u(z,t) = u(ϕ) where ϕ = (t0 + t ± z/c) is a characteristic phase function with propagation speed c. Differentiating with respect to time and distance respectively, the fibre particle velocity is:
v
u u t
(4)
and the fiber strain:
u u 1 / c (5) z
Comparing these equations, then = ± ; that is, the ratio between fiber particle velocity and fiber strain is given by the propagation speed along the fiber cable (apparent velocity) with a sign determined by the direction of propagation. Therefore, the time-integrated DAS signal is converted to equivalent geophone signal by multiplying the dimensionless strain by the local propagation speed (as determined from VSP moveout data).
6
DAS Walkaway VSP data acquisition During a survey in a formerly producing CNPC well in north-eastern China, near Tangshan, Hebei [15], walkaway vertical seismic profile (VSP) data were acquired using Silixa’s iDASTM system. The primary objective was to evaluate the velocity and anisotropy characteristics for use in a surface seismic survey. A secondary objective was to investigate the overall performance and operational feasibility of using DAS for VSP surveys. In January 2015, BGP and Silixa shot a Walkaway VSP survey as part of a larger commercial 3D surface seismic campaign in a 4,004 m vertical well in north eastern China. This 6 year old oil producing well was plugged back above its perforations. A total of 386 dynamite shot points were arranged into one in-line Walkaway and one cross-line Walkaway. Figure 4 shows the overall shot line layout, each shot position and its corresponding shot number (top) and receiving cable deployment (bottom). The surface conditions are characterized by a single narrow causeway surrounded by shallow fishponds. A denser shot Walkaway line was placed on the causeway due to ease of placing charges. Two multimode 62.5 μm fibers which make up part of a commercially available 5,000 m hybrid optical-electric wireline cable were spliced together at the cable head to form a continuous 10 km down then up-going distributed sensor. This geometry gave two side-byside stackable fiber channels. During conventional use of this wireline cable, these multimode fibers are for communications with downhole logging tools. The DAS interrogator, triggered by a Real Time System source controller, was used to record 6 kg dynamite shots. Recording was setup to measure 0.25m intervals at 1 ms sample rate along 8 km of the fiber optic cable. The first and last portions stayed on the cable drum. Data acquisition took place over the course of 5 days, working daylight hours, with 386 successful shot points acquired [16].
Figure 2. The two panels show data from a single shot, before (left panel) and after (right panel) the application of the velocity transform. The hybrid wireline cable was rigged up as illustrated in Figure 4. The significant point is that this is a standard wireline rig up with only a weight at the bottom and normal depth 7
control procedures. In future surveys, there would be nothing to prevent the use of standard logging tools before or after the seismic survey. The electrical conductors and two of the fibers were still available.
Figure 3. DAS Walkaway VSP shot line layout (left) and receiving hybrid wireline cable deployment (right).
8
Figure 4. Distributed Acoustic Sensing VSP survey receiving hybrid electric-optic wireline configuration Diagram. DAS Walkaway VSP data analysis and processing Field data was saved in the contractor’s proprietary format. Afterward, the data was converted to SEG-Y format (Society of Exploration Geophysicists Y Format) using acoustic conversion algorithms that involve noise-adaptive resampling to reduce the spatial sampling from 0.25m to 2m. The process also includes the straight channel alignment and stacking of the VSP data recorded through the two legs of fiber in the well. As a result, the signal to noise ratio is greatly improved by removing unwanted system generated optical noise. The noise suppression process provides us with a 2,000-level Walkaway VSP data set covering the full well depth and does not distort the seismic signal measured by DAS. The physical property measured by DAS is the rate of change of strain (strain rate) within the fiber. The units within DAS measurement files and the delivered SEG-Ys are proportional to this quantity. During the preparation of the final data set, the section of the data covering channels within the well was extracted. This is channels between the Blowout Preventer (BOP), the bottom of the well and back to the BOP again. The measured depth relative to 9
ground level of each channel was calculated using the field recorded depth QC measurements (tap test) and comparison with the well completion diagram. The receivers corresponding to the BOP were assigned a depth of 1.4 m above ground level; the bottom receiver was assigned a depth of 4,004 m below ground level. A linear interpolation was applied between these positions to assign a depth to each receiver. In the pre-processing step, we were faced with additional and special challenges: strong coherent noise due to cable slapping and ringing along the borehole casing. A special patent pending algorithm and procedure were developed to remove the strong coherent cable slapping and ringing noise (Figure 5).
Figure 5. DAS Walkaway VSP raw data (left) and cable slapping and ringing noise removing results (right) comparison. Following is the detailed description of the data analysis and the coherent noise suppression method: 1. Wavefield analysis In a single -VSP shot record, we can recognize several waves, such as downgoing and upgoing P-wave, but we also find that there are strong coherent cable slapping and ringing noises in the record (Figure 6a). The analysis shows that: 1. The downgoing P-wave is clearly visible, but its first break is not clear; 2. The upgoing P-wave is clearly visible, but it is disturbed by strong coherent noise; 3. The downgoing and upgoing P-S wave is visible; 4. There are few multiple waves; 5. Cable slapping and ringing noise is the main disturbance.
10
(a)
(b)
(c) Figure 6. Raw VSP data (a) recorded by the DAS acquisition system and the hybrid wireline, synthetic cable slapping and ringing noise wavefield (b), and cable slapping and ringing noise removed DAS Walkaway VSP wavefield (c). 2. Noise period analysis We analyze the noise period of regular cable slapping and ringing noise by statistic method as follows: For each seismic trace, the maximum time ( Tmax ) of recognizable period is read firstly, and then maximum period numbers ( nmax ) are counted. Finally, we calculate a single noise period ( T ), it is expressed as follow:
T Tmax / nmax
(6) Significantly, there are more than one noise period in some traces. 3. Noise wavelet analysis
11
In each seismic trace of raw wavefield, wavelet extraction includes the following steps: finding the noise wavelet ( Wi ) in each period, calculating these wavelets to obtain '
normalized wavelets ( Wi ), then stacking all wavelets to form an average wavelet ( Wave ), it is expressed as follow:
Wave 1 / nmax i max Wi ' 1 n
(7)
Because some traces have more than one noise periods, then these traces have different noise wavelets. 4. Noise attenuation characteristics analysis Cable slapping and ringing noise travels up and down in optic fiber, and its attenuation characteristics are different than that of seismic wave attenuation. The attenuation characteristics are dependent on optic fiber material and its length. According to many tests and detailed analysis, we proved that the best fitting method of the attenuation curve was not the exponential fitting, but the base fitting. When the noise travels a time ( t ), the raw amplitude ( A0 ) at the position of its first break is decreased to new amplitude ( At ), it is expressed as follow:
At At xt
(8)
t
Where x is the attenuation factor, and x is the base number of the factor. It can be calculated by attenuation curve statistic method (Figure 7).
Figure 7: Attenuation curve and the average curve obtained through attenuation curve statistic method. For the coherent cable slapping and ringing noise suppression method, following steps were adopted:
12
Figure 8. Illustration of noise fitting process. Using the formula (4) to each seismic trace, the cable slapping and ringing noise wavefield is synthesized (Figure 6b). 1. Noise fitting According to the above analysis, the cable slapping and ringing noise ( N ) can be fitted by first break time ( T0 ), first break amplitude ( A0 ), noise period ( T ), average wavelet (Wave ) t
and attenuation parameter ( x ) (Figure 8).
N i max A0 x 0 1 n
T i 1T
*Wave
(9)
2. Noise reduction The raw wavefield ( S ) is subtracted by the synthetic cable slapping and ringing noise wavefield ( N ), and then leaves noise removed DAS Walkaway VSP wavefield ( S res ) (Figure 4c). Sres S N (10) Comparing the wavefield in Figure 6a and Figure 6c, obviously, cable slapping and ringing noise is effectively suppressed, and the upgoing wave is clearer in the noise removed DAS Walkaway VSP wavefield ( S res ). 13
Figure 9. Autocorrelation data of raw wavefield (top) and noise removed wavefield (bottom). After DAS fibre channel alignment and stacking of VSP data recorded through the two legs of fibre in the well, a typical VSP processing sequence was applied: trace editing, noise attenuation, spherical divergence correction, up/down separation and up-down deconvolution. Figure 10 shows the comparison of common source gather (1,000 m offset) of the DAS Walkaway VSP data before and after wavefield separation processing. 3. Quality Control It is well known that autocorrelation is an effective checking method for the regular coherent noise suppression. Figure 9 shows two autocorrelation datasets of raw wavefield and noiseremoved wavefield. By comparison, the side lobe of noise removed wavefield autocorrelation is much less than that of the raw wavefield, the suppression method is proved to be effective and feasible. DAS Walkaway VSP data imaging After up/down wave-field separation, we first perform the VSP data coordinate transformation through wave-field continuation. Then based on the theory of bending rays and preserved amplitude weight function formula, a stable ray tracing was used to establish, numerically, the relationship between time-distance curve of the reflected wave and offset angle of incidence mapping. Finally, high precision common imaging point (CIP) was used in gathering VSP-CDP (Common Depth Point) imaging and pre-stack Kirchhoff time migration (PSTM) to image the up going reflection VSP data.
14
Figure 10. Comparison of DAS Walkaway VSP raw data (left) and wavefield separation results (right). Method 1. VSP-CDP imaging VSP-CDP imaging is the basic method of VSP imaging, among the improve methods, the ray tracing imaging is the most representative method [17, 18], its steps as follow:
Velocity model building. According to the zero-offset VSP data and the 2-D or 3-D seismic velocity, the velocity model is built. Ray tracing. According to the velocity model, the location and travel time of reflect point is calculated. Coordinate calculating. The coordinate of reflect point is calculated by using interpolation method. VSP data assignment. For each trace of VSP data, according to the above steps, each sample is assigned the calculated coordinate. Imaging gridded. Given a value of group interval and gridded the imaging area, and then put the VSP data into the correspondent locations. VSP-CDP imaging. Time-depth transformed is the last step. Figure 11 shows the final VSP imaging.
Method 2. Pre-stack Kirchhoff time migration The pre-stack Kirchhoff time migration of VSP is based on Huygens's principle [19].
I r, t 0
1 cos U V dA x , t U x , t 0 0 0 R 2π RV t R t0 t
(11)
V
Among the function, A0 is the locations of sources and geophones in earth's surface, is the angle of emergence of the ray form imaging point to a receiver, R is the distance of seismic wave transmission from reflection point to receiver, cos z / R is the declination factor 15
( z is the depth of imaging point), it shows the amplitude changes with the emergence angle when the scattering point is as a secondary source, t R / V is the delay time, t0 is the reflection time from source to receiver, and V is the transmission velocity of seismic wave.
Figure 11. Surface seismic profile (left) and DAS Walkaway VSP image (VSP-CDP mapping) insert into surface seismic profile (right) along the Walkaway VSP survey line with denser VSP shots. The edge of VSP image is muted.
Figure 12. Surface seismic profile (left) and DAS Walkaway VSP image (VSP-CDP mapping) insert into surface seismic profile (right) along the Walkaway VSP survey line with coarser VSP shots. Figures 11 shows the surface seismic section along the North East Walkaway VSP line that has higher shot density and the VSP image inserted into surface seismic profile. The surface seismic image along the North West Walkaway VSP line with coarser shot density and its VSP image are inserted into surface seismic section as shown in Figure 12. Both Walkaway
16
VSP images not only match well with surface seismic image they also show higher vertical resolution and more detailed structures. Conclusions A Walkaway VSP survey acquired with fiber optics over a former producing well in northeastern China provided very large aperture Walkaway VSP data. Strong coherent noise due to cable slapping and ringing along the borehole casing created special challenges in the preprocessing step that was overcoming by designing a new and effective noise removal algorithm. Afterward, standard VSP data pre-processing workflow could be applied, followed by pre-stack Kirchhoff time migration. The final pre-processed DAS data has a larger vertical aperture than previous VSP with geophones resulting in a wider lateral image from a single well. The single well DAS Walkaway VSP images provide a good result with higher vertical and lateral resolution as well as more detailed structures than the surface seismic in the objective area. The vertical well environment without the ability to effectively clamp the sensor to the borehole casing wall creates a unique set of challenges. Although earth signal was recorded with almost all the shots, there was also a considerable amount of noise. Much of the noise was due to the way the wireline was physically placed in the well and expressed by slapping and ringing. Among the lessons learned on handling the hybrid wireline cable during acquisition were to fill the top part of the well with water, stabilize the hybrid wireline between wellhead and recording truck, run some slack, use a T-Bar at the wellhead to hold the downhole hybrid wireline cable and use smaller dynamite charges on near offset shots. Special DAS data processing steps were developed to remediate some of the practical wireline deployment issues. Optical wireline cable as a conveyance of fiber optic cables for VSP in vertical wells will open the use of the DAS system to much wider applications. Acknowledgements We would like to thank Jidong Oilfiled and CNPC for allowing the publication of this article. We would also like to thank BGP and Silixa management for their support and guidance during the data acquisition, processing and interpretation phases of this project. This study was funded by the China 13th Five Year Plan DAS Borehole Seismic Project (2016ZX05018004-003) and China National Natural Science Fund Grants 41527805. References [1] Frignet, B. and Hartog, A., 2014, Optical vertical seismic profile on wireline cable. SPWLA 55th Annual Logging Symposium, Abu Dhabi, May 18-22. [2] Daley, T., Freifeld, B., Ajo-Franklin, J., Dou, S., Pevzner, R., Shulakova, V., Kashikar, S., Miller, D., Götz, J., Henninges, J. and Lüth, S., 2013, Field testing of fiber-optic distributed acoustic sensing (DAS) for subsurface seismic monitoring. The Leading Edge, 32(6), 699706. [3] Hartog, A., Frignet, B., Mackie, D. and Clark, M., 2014, Vertical seismic optical profiling on wireline logging cable. Geophysical Prospecting, 62(4), 693-701. [4] Madsen, K., Thompson, M., Parker, T. and Finfer, D., 2013, A VSP field trial using distributed acoustic sensing in a producing well in the North Sea. First break, 31, 51-56. 17
[5] Mateeva, A., Lopez, J., Potters, H., Mestayer, J., Cox, B., Kiyashchenko, D., Wills, P., Grandi, S., Hornman, K., Kuvshinov, B., Berlang, W., Yang, Z. and Detomo, R., 2014, Distributed acoustic sensing for reservoir monitoring with vertical seismic profiling. Geophysical Prospecting, 62, 679- 692. [6] Parker, T., Shatalin, S. and Farhadiroushan, M., 2014, Distributed Acoustic Sensing – a new tool for seismic applications，First Break, 32, 61-69. [7] Lesnikov, V., Allanic, C., 2014, DAS VSP acquisition - perspectives and challenges: 76th EAGE Conference and Exhibition - Workshops, [8] La Follett, J., Wyker, B., Hemink, G., Hornman, K., Lumens, P., Franzen, A., 2014, Evaluation of fiber-optic cables for use in Distributed Acoustic Sensing: commercially available cables and novel cable designs: 84th Annual International Meeting, SEG, Expanded Abstracts. [9] Mateeva, A., J. Mestayer, B. Cox, D. Kiyashchenko, P. Wills, J. Lopez, S. Grandi, K. Hornman, P. Lumens, A. Franzen, D. Hill, and J. Roy, 2012, Advances in distributed acoustic sensing (DAS) for VSP: 82nd Annual International Meeting, SEG, Expanded Abstracts. [10] He, X. G., Xie, S. G, Liu, F., Cao, S., Gu, L. J., Zheng, X. P. and Zhang, M., 2017, Multi-event waveform-retrieved distributed optical fiber acoustic sensor using dual-pulse heterodyne phase-sensitive OTDR. Optics Letters, 42, 442- 445. [11] Masoudi, A., Belal, M. and Newson, T.P., 2013, A distributed optical fibre dynamic strain sensor based on phase-OTDR, Measurement Science and Technology, 24, 085204. [12] Steinberg, I., Shiloh, L., Gabai, H. and Eyal, A., 2015, Over 100km long ultra-sensitive dynamic sensing via Gated-OFDR, International Conference on Optical Fibre Sensors, Curitiba, Brazil, 96341B. [13] Wang, C., Shang, Y., Liu, X. H., Wang, C., Yu, H. H, Jiang, D. S. and Peng, G. D., 2015, Distributed OTDR-interferometric sensing network with identical ultra-weak fiber Bragg gratings. Optics Express, 23, 22, 29038- 29046. [14] Zhu, F., Zhang, Y. X., Xia, L., Wu, X. L. and Zhang, X. P., 2015, Improved Φ-OTDR Sensing System for High-Precision Dynamic Strain Measurement Based on Ultra-Weak Fiber Bragg Grating Array. Journal of Lightwave Technology, 33, 23, 4775- 4780. [15] Yu, G., Li, Y. P., Wang, X. M., Greer J., Clarke A., 2015, Deep vertical seismic profiling using multimode optical fibers in a hybrid wireline - lessons learned: SPE Fiber Optic Sensing Workshop, Napa, CA, USA. [16] Yu, G., Cai, Z. D., Chen, Y. Z., Wang, X. M., Zhang, Q. H., Li, Y. P., Wang, Y. H., Liu, C. W., Zhao, B. Y. and Greer, J., 2016, Walkaway VSP using multimode optical fibers in a hybrid wireline, The Leading Edge, July 2016, 936-940. [17] Chen G., Canales L. and Peron J., 1999, VSP mapping error for dipping horizons using a horizontally layered model. 69th SEG Annual Meeting. 18
[18] Chen G., Peron J. and Canales L., 2000, Rapid VSP-CDP mapping of 3-D VSP data. Geophysics, 65(5), 1631-1640. [19] Schneider, W. 1978, Integral formulation for migration in two and three dimensions. Geophysics, 43(1), 49~76
19