4.1. VSP of SMF
Single-mode researchers such as the authors of [
23] have introduced DAS for borehole seismic applications. DAS has the potential to expand beyond its initial use and encompass other surveillance domains, including production and completion reservoir monitoring. This extension is made feasible by leveraging existing fiber optic networks for enhanced data acquisition and monitoring capabilities. However, applying a suitable DAS data acquisition and processing workflow is also crucial in obtaining the best images; hence, many papers have utilized diverse DAS outputs that are applicable to exploration activity. VSP data are used to complement seismic data through the process of separating the signals from multiple sources that overlap in time and were acquired by deploying seismic sensors in a wellbore and recording the response of the subsurface to a controlled seismic source [
50].
Figure 5 shows a schematic deployment for land VSP acquisition. DAS offers the distinct advantage of remarkable spatial sampling, spanning the complete length of a well.
Furthermore, as the underlying technology, SMF optic cable presents enhanced robustness compared to conventional geophones. This robustness mitigates the potential hazards of deploying mechanical and electrical components within a well [
51]. This makes SMF an excellent tool for replacing geophones, providing subsurface properties, in addition to having been used widely in various subsurface applications.
In 2015, the authors of [
52] located an indication of a reservoir of the Aquistore CO
2 geological storage project through elementary imaging using DAS in Paleozoic carbonates and Mesozoic sandstones and shales. SMF optical fiber cables were permanently installed inside steel tubing, strapped behind casing, and cemented in placed in observation wells during completion, even though it was only the fiber optic cable deployed in the observation well that eventually operated well. SMF within the observation well was employed to conduct comparisons with geophone VSP measurements. A processing framework was executed for each shot obtained through both DAS and conventional geophones. The general processing framework encompassed trace balancing; correction for spherical divergence, notches, and low-pass filtering; and a frequency-wavenumber filtering algorithm of the down-going wavefield. To compare it with geophones, DAS trace gathering was superimposed with the geophone traces within a specific depth range. SMF DAS and vertical-component geophone traces were also imaged for analysis. The results indicated that geophone data distinctly showcase a higher SNR than DAS data. However, it is worth noting that the DAS for SMF SNR has significant room for enhancement through the application of noise attenuation techniques and the averaging of data over cable lengths, as it is used in geophone results. In contrast, numerous equivalent reflections are obscured by noise within the DAS data for the same depth range. However, DAS still has the advantage of potentially replacing many geophones within one SMF. This may potentially generate subsurface images by notably enhancing spatial resolution and coverage when compared to traditional geophones. Moreover, this improved performance can be achieved at a relatively lower cost.
DAS VSP has also proven to be useful in acquiring VSP data in carbon capture and geological storage (CCS) projects [
53]. By applying multiwell 3D DAS VSP data acquired at the CO
2CRC Otway, DAS data were processed and analyzed for their applicability for continuous lime-lapse monitoring of CO
2 injection. The acquisition process was performed by deploying four SMF cables cemented along the well, where two SMF cables were connected with constellation fiber (CF) and each SMF cable was connected to an iDAS v3 interrogator unit. The DAS data were recorded under a strain rate with a 5 m depth interval and correlated with vibroseis source sweep data, enabling first-breaks picking. Further DAS VSP data processing included vertical stacking until 3D migration and plotting with 3D seismic cross sections. This helped in clearly imaging the key stratigraphy horizon and CO
2 injection target within the SMF cable. This resulted in a good correspondence in the seismic reflectors in the datasets while comparing a 3D DAS VSP image with surface 3D seismic imaging. The 3D DAS VSP shows higher image resolution and provides enhanced precision in determining the depth of reflectors within the well area. Therefore, the implementation of multiwell DAS VSP emerges as a viable approach for continuous monitoring systems. The results of multiwell DAS VSP could have been improved by taking more action to tackle missing seismic volumes from different typical wells.
In 2016, the authors of [
54] conducted research to validate DAS VSP data integrity with a conventional geophone at the Aquistore site for subsurface CO
2 storage. This data acquisition used an SMF cable with a total fiber length of 5.4 km, and the configuration involved the embedding of the initial portion of the cable near the surface, while the subsequent section was permanently cemented behind casing and vertically positioned along the cable deployment. The acquisition also involved active sources, such as dynamite and vibroseis, to cover entire boreholes. Each near offset and far offset shot were applied to dynamite shot gather to indicate direct arrival, reflection events, and ground roll in shallow sections. These three DAS events were compared with geophones in terms of noise characterization and identified the noise as optical system noise, optical fading, common mode noise, or checkerboard noise. Initially, the raw DAS data on strain rate were also converted to geophone-equivalent units to achieve a good comparison of the two raw data acquisitions for further assessment. Both up- and down-going events managed to match the polarity between geophone and DAS data. The up- and down-going wave separation might have been enhanced by a suitable denoising filter between the two input trace gathers. The result indicated that converting DAS VSP data to equivalent geophone data can help with the up- and down-going separation step without using deep filters. With the mitigation of depth calibration in seismic DAS, the data processing workflow can be improved.
The authors of [
55] carried out a DAS VSP field survey to evaluate the effectiveness of fiber optic DAS by comparing four cable deployment approaches, namely with cables behind the well casing, behind an inflatable liner, clamped to production tubing, and deployed with wireline logging. The SMF cable was lowered along the well with a different approach and connected to a DAS interrogator. To generate seismic waves, active sources were employed using P-wave and an Envirovibe S-Wave Vibrator. During the final imaging of VSP for Common Depth Point (CDP) transform, DAS data attenuation accrued for all cable deployment methods. Noise attenuation reduced the SNR, influencing the quality of the seismic image. The results indicated that the deployment method with the fiber cable cemented behind casing was considered the most effective method for cable coupling with the subsurface layer, with a high-quality generated image in the pre-defined velocity model. Such an improvement can be generated by a depth-calibrated SMF cable against a reference to account for source statics resulting from variations in elevation. This correction was essential in achieving optimal resolution in the subsurface imaging.
Further DAS measurement research was carried out by the authors of [
56] to conduct real-time DAS VSP acquisition and processing using an SMF cable with seismic-source synchronization and real-time DAS data processing to provide contemporary work. The SMF cables were deployed outside the casing in vertical and lateral sections along the wellbore. The configuration of the DAS system used a homodyne DAS interrogator, which was systematically connected to fiber stretchers. This configuration was used for its capacity to integrate all supplementary seismic source signals that directly connected into the optical data stream while the obtained data were in velocity. This integration was accomplished with a sequence of piezoelectric fiber stretchers positioned in line with the SMF cable. The DAS system was able to continuously monitor the stretching fiber. The DAS maximum data sampling frequency for SMF was able to reduce the noise power spectral density of the DAS data stream to achieve real-time processing. The data processing workflow was conducted by converting raw DAS data to strain rate with weighted stacking for certain shot points and correlated with vibroseis sweep wavelets and common-mode noise denoising in SEG-Y format. The signal strength for SMF cable from DAS VSP shot records was computing SNR in the first break, and the noise level from RMS energy indicated a lower SNR along the channel. The findings demonstrated that utilizing an SMF cable in a DAS system facilitates an efficient data acquisition process and enables real-time generation of seismic and navigation data sampled at specific intervals. These DAS data adhere to geophysics seismic data formats, reducing the time delay between data acquisition and the transfer of field data outputs.
Fiber optic applications have achieved remarkable outcomes when deployed in high-spatial-resolution seismic sensors using cost-effective telecommunications fiber optics installed along the drilled well to monitor hydrocarbon reservoirs [
57,
58,
59]. Permanently installing a seismic array cable within a borehole can streamline equipment deployment, decrease operational complexities, and offer notable advantages such as enhanced repeatability and the potential for real-time data acquisition [
60]. SMF has been widely used to monitor reservoirs in various subsurface targets.
Research was conducted to evaluate DAS image quality of subsea carbon storage reservoirs and assess monitoring design concepts [
61]. Subsea DAS was deployed within a single ultra-low-loss signal transmission SMF cable through subsea infrastructure that connects the DAS interrogator unit to a borehole SMF deployed within a dual-transmission SMF with a remote circulator. This concept was employed to counteract diminished optical signals by using an enhanced backscatter of SMF. The synthetic shot records were augmented with raw DAS records in terms of optical noise to replicate the varying SNR in SMF scenarios, namely a base case involving a dry tree with common SMF and a subsea scenario entailing a single transmission fiber and common SMF. The subsea environment with a single transmitting fiber and standard SMF had a negative value of net DAS noise gain and lower DAS pulse repetition rate loss. The alteration in the DAS noise floor within the context of the two scenarios involving SMF was assessed by analyzing the overall gains or losses resulting from various physical factors influencing system performance. The result highlighted the repercussions of low SNR in terms of subsea DAS image resolution. The authors concluded that employing a single transmitting fiber with standard SMF was not preferable for the subsea environment, especially for characterizing reservoir formations. The single signal transmitted in fiber optics with standard SMF was filled with noise, in contrast with the other situation for subsea transmission fiber. These findings can inform the selection of reservoir monitoring and data acquisition designs using SMF during offshore reservoir research. A possible further improvement of subsea DAS is enhancement by selecting an appropriate vendor of DAS SMF interrogator units to handle noise. This would involve the choice of a relevant interrogator vendor before conducting the subsea DAS survey.
Subsequently, a DAS SMF with active sources generated by rotary seismic sources was applied to facilitate autonomous permanent reservoir monitoring in the CO2CRC Otway Field Site, Australia [
62], by utilizing rotary seismic data as an active source that accurately controls routinely operated signal systems (ACROSS) [
63,
64]. The DAS data were continuously collected and recorded in five wells with SMF cable cemented behind the casing and underwent automated processing at the close of each day. The data collected through the permanent reservoir monitoring system using DAS SMF and rotary seismic receivers yielded a notably higher SNR ranging from roughly 50 dB to over 100 dB for near-offset data shot. The DAS using SMF cable successfully showcased favorable data quality attributes, encompassing SNR and data repeatability. This was evident in time-lapse operations for the monitoring carried out using the DAS system. This research can be improved by enhancing the spectral-frequency content of seismic data by using dual-motor rotary seismic sources, subsequently combining high-resolution images for time-lapse acquisitions and the achievement of a high SNR for autonomous reservoir monitoring. Research applying SMF cables to VSP input DAS data is summarized in
Table 2 below.
4.3. ML Application in SMF Processing
Efficient automated processing schemes for raw DAS data have achieved significant importance. SMF cables have the capacity to consistently monitor acoustic signals and vibrations across extensive distances, offering high sensitivity and an elevated update rate [
68,
69]. DAS SMF measurement conventionally captures numerous complex backscatter profiles in a short time and performs real-time data processing. Given the substantial volume of generated data, it is essential to devise automatic, efficient, and precise tools for signal processing. As optical data are complex, have a high bitrate, and must be processed in real time, state-of-the-art machine learning techniques are ideal for extracting relevant information [
70]. Furthermore, ML can potentially decrease the amount of storage space needed for data and minimize the time required for data processing to facilitate comprehensive analysis [
71].
Figure 6 shows general ML processing applied to DAS data and its output.
Recent research reported in [
72] applied classical ML and deep learning algorithms to SMF DAS event detection. ML was compared with the deep learning result, and the benefits and limitations of both methods were analyzed for each applied algorithm. The classical ML workflow involved DAS data collection, preprocessing, feature extraction and classification, model creation, tracking, and evaluation of event detection. DAS data were acquired using a simple sensor unit with standard SMF and connected to an interferometer. SMF cable was deployed in a 17 km cable length, with a gauge length of 10 m and a phase-sensitive OTDR interrogator unit. The DAS data were collected in a suburban area near Vienna, Austria. The used ML algorithms were random forest, decision tree, and support vector machine (SVM), while the deep learning algorithms were deep neural networks (DNNs) and convolutional neural networks (CNNs). Deep learning techniques were employed for SMF event detection, utilizing the ability to automatically extract relevant attributes from the raw DAS data. This approach obviates the necessity for extensive domain knowledge. The implementation of this method offers an achievable approach for optimum sensor fusion in SMF sensors in DAS systems. A comparative analysis of outcomes between ML and deep learning methodologies demonstrated that DAS could detect various continuous events along the pipeline, such as human digging and tapping. Moreover, it can identify significant events in other SMF systems.
Despite the numerous benefits, DAS data can still be contaminated by various types of noise from the ambient environment and the optical instruments. Supervised ML techniques were applied to target specific noise types to improve the SNR and denoising (noise reduction) of DAS data [
73]. Recent research implemented supervised ML method DAS-N2N (Noise2Noise) to denoise DAS data without clean data [
74]. The DAS data were acquired in SMF cable laid out in a 1 km cable length, with sampling at 1000 Hz, channel spacing of 1 m, and a 10 m gauge length. SMF cables placed within the cable jacket were spliced at one end to build a looping cable. The results obtained from DAS-N2N were contrasted with the outcomes of conventional bandpass filtering. In this comparison, DAS-N2N demonstrated its ability to attenuate noise within a frequency range identical to that of the target signal, a task that frequency-based filtering cannot achieve. Furthermore, DAS-N2N exhibited a processing speed approximately ten times faster than self-supervised learning methods. The result indicated that DAS data are remarkably lightweight and efficient and more capable of data processing compared to the acquisition time when fine-tuned with a single GPU. DAS-N2N outperforms conventional frequency bandwidth filtering routines enhanced using compiled low-level computation regarding processing speed. Further improvement in SMF data imaging could be achieved by optimized supervised ML models compiled and compressed for DAS processing on suitable high-performance devices and edge networks, which will be appropriate for offshore or remote monitoring scenarios.
Table 3 highlights a summary of all research on SMF for microseismic monitoring and the ML applications discussed above.
4.4. VSP of MMF
The subjectivity of VSP usually depends on the interpretation and calibration of the downhole seismic arrays that can be addressed by applying fiber optic installations. Field research in a hybrid wireline using MMF cable was conducted to assess the results of a borehole seismic survey [
75]. Two MMF cables with 62.5 μm core diameters were deployed within a 5 km hybrid optical–electric wireline cable and connected via splicing at the cable head. This fusion created a length of 10 km up-going and down-going DAS. A real-time active system source controller with a 6 kg dynamite shot triggered the DAS interrogator. The recording configuration was established to capture measurements at intervals of 0.25 m, with a sampling rate of 1 ms and encompassing 8 km along the MMF cable. The MMF cable was set to acquire the strain rate (rate change) within the fiber as the physical property. Raw DAS VSP data in the preprocessing step showed that there was a high coherence in MMF cable slapping and ringing noise; this noise was further processed in the following VSP sequence: seismic trace editing, denoising filter, spherical divergence correction, up–down separation, and deconvolution. DAS walkaway VSP imaging data were processed to build a velocity model. This resulted in a wider lateral image from a single well with a larger vertical aperture than VSP with a geophone. The processing of DAS data yielded improved outcomes, addressing certain practical challenges associated with the hybrid wireline in borehole deployment. The result indicated that DAS with MMF cable walkaway VSP provided an excellent result with vertical and lateral imaging resolution and managed to detail the structure in the objective area. The application of MMF in the exploration field can help minimize the operation costs of VSP in vertical wells, owing to its utilization of a DAS system and its expanded applicability and reliability.
The authors of [
56] used fiber optic installations with MMF cable for sensing according to the principle of Raman DTS applied to real-time DAS VSP data acquisition and processing. MMF cables were synchronized with seismic sources by directly encoding auxiliary seismic signals onto the optical data stream. This was applied through the incorporation of piezoelectric fiber optic stretchers aligned with the sensing fiber cables. Inside the schematic interrogator unit, MMF cables were connected to SMF cables using an SM–MM converter. MMF cables were deployed through the circulator and sensing fiber along a far offset of the measurement. MMF was deployed as a scrambler to transfer energy from the fundamental mode to all the (low-loss) modes. This is essential to increase the amount of light power emitted. The DAS VSP shot record for MMF was recorded at a 16 kHz pulse rate and a gauge length of 20 m. The MMF result, in terms of system performance, indicated that MMF yielded a 2 dB reduction in SNR for DAS VSP, while the optical settings of DAS interrogators were placed at 5 dB. This implies that MMF measurement managed to have sufficient capacity for the DAS interrogator and acquired DAS VSP data in the MMF cable with 8 dB loss. However, a limitation of the researchers findings is that the MMF installation has shallow image due to the increase in attenuation with significant hydrogen darkening. For future work, there should be a sufficient SM–MM converter for DAS sensing of MMF deployed in legacy DTS installations to improve SNR and signal loss.
Improving reservoir monitoring relies on many factors like real-time data analysis and detailed subsurface imaging, leading to a better utilization of fiber optic technology. Accurate imaging of these fiber optic measurement data is essential in lateral andvertical spatial distributions in reservoir monitoring. In their advanced research on 4D reservoir monitoring, the authors of [
76] applied DAS 4D VSP to MMF cable in a deep-water environment to reveal injected water-sweep dynamics. MMF was deployed in active wells to perform detailed monitoring of the water flood injection in two deep hydrocarbon reservoirs in a field in the Gulf of Mexico. In the technology trials, the research compared DAS VSP and Ocean Bottom Node (OBN) to obtain the 4D imaging performance of the two data sources. The DAS data were continuously measured over the years to understand the influence of small active sources and active well noises in 3D and 4D DAS VSP on MMF imaging resolution. The progression began with a DAS VSP recording in 2015 within an active well while water flooding injection operations were ongoing. Subsequently, DAS fibers captured in a P-cable trial featuring 300 events from three sources in early 2017 served as benchmarks for additional iterations. Furthermore, in late 2017, a standalone DAS VSP survey was conducted, distinct from the simultaneous OBN deployment, encompassing three active injector wells. The measurements encountered additional challenging conditions for DAS, including notable well deviation that resulted in diminished sensitivity to desired reflections, as well as the presence of extended MMF lengths that contributed to heightened levels of optical noise. However, this measurement endeavor yielded crucial insights into acquisition and processing methodologies. It was discovered that DAS 4D images exhibited qualitative similarity to OBN 4D outcomes. Furthermore, the repeatability of DAS images proved to be exceptional, as demonstrated by a well-controlled Normalized Root Mean Square (NRMS). The reservoirs are sandstone reservoirs with turbidite deposits with high-quality sands composed of different lobes with varying thicknesses across the field located on deep-water acreage. The results of DAS MMF and OBN were compared in two reservoirs within the same vertical conformity of the time-lapse signals in both datasets. These results show that frequent 4D DAS VSP can assist in characterizing all phases of a water flooding injection; the data were also used for sweep monitoring of the evolution of the injector. On the other hand, the survey result demonstrated that the associated costs remained reasonable, while the resultant seismic images proved to be well-suited for monitoring the immediate surroundings of wells. These images facilitate the assessment of injector performance, the evolution of sweep processes, the progression of water towards producers, and the containment dynamics within the reservoir. However, the study highlighted several limitations of DAS in MMF VSP regarding flow- and injection-related well noise. Injection noise predominantly presents itself through the propagation of robust tube waves along the wellbore, characterized by comparatively lower velocities in DAS data. A data processing framework should be developed to incorporate insights from 4D DAS VSP data modeling, adding well log data to extract physical properties of reservoirs and consider the understanding of reservoir dynamics behavior. Nevertheless, an improved 4D DAS VSP may offer a promising starting point for real-world DAS of MMF for reservoir monitoring in a thick sand layer. Future research should delve into comparative analyses with other active seismic sources, explore the deployment MMF cable with evolving technologies, and assess DAS acquisition parameters involving deployment in reservoirs.
Table 4 below summarizes the works reviewed in this section.
4.5. Microseismic of MMF
Fiber optic cables have the capability to be permanently installed by cementing them behind a protective casing or temporarily deployed using a wireline method [
77,
78]. Once a fiber optic cable has been deployed in a downhole environment, it can serve various functions, including the monitoring of microseismic activity. Despite the utilization of MMF, DAS can effectively produce estimations and source mechanisms that provide sufficient coverage for obtaining a singular-moment tensor solution in microseismic measurements. Characterization of source parameter estimation using DAS with MMFrecorded microseismic data was conducted in [
79]. An MMF cable was deployed within a treatment well located in the Meramec Shale Formation in Oklahoma. This cable spanned the whole vertical distance of the well, extending from the surface down to the desired depth.
Figure 7 shows an illustration of microseismic event measurement in hydraulic fracturing by DAS fiber optic cables. As a result, almost 1000 channels were recorded through this installation. During the treatment of two wells, comprehensive measurements were conducted to assess DTS, strain, and microseismic activity along the full depth of the wells. DAS on MMF achieved excellent waveform imaging for two microseismic events with S-wave polarity reversals. This is because DAS microseismic events with consistent amplitude patterns were observed using event classification considering polarity reversals and horizontal and vertical MMF distance along the fiber. The resulting variation in nodal planes for DAS microseismic events showed a shear-wave amplitude pattern and even managed to generate a nodal plane angle picked manually by considering the horizontal and vertical axis distances of the microseismic event from the treatment well. The result of using DAS for seismic monitoring indicated that the MMF was more accurate in characterizing source features, specifically in estimating moment tensors. Future endeavors might to improve signal processing techniques and consider anisotropy in subsurface properties and its effects on the propagation of seismic waves detected by DAS systems.
Machine learning methodologies have been extensively employed to enhance the efficiency and precision of data processing [
80,
81,
82]. Several ML strategies have been tested in academic literature to increase the rate of processing of DAS microseismic data; another such strategy is the utilization of deep learning [
83].
The authors of [
84] proposed a new and efficient data processing workflow for MMF DAS microseismic data using an ML-assisted approach. A convolutional neural network (CNN) algorithm allowed for microseismic event detection, and U-Net modeling was used for both P- and S-wave arrival-time picking. MMF cables were previously deployed with two multiwell DAS datasets acquired during hydraulic fracturing well completions in western Canada. MMF was deployed using a 4 m gauge length, 4 m channel spacing, and a 2000 Hz sampling rate, and data were converted to strain. The developed ML-assisted workflow was highlighted in the event detection and 1D arrival time picking after raw DAS data denoising. The results of applying the CNN algorithm were compared to the conventional short-term average/long-term average (STA/LTA) approach and geophone for microseismic event detection. The CNN detection method demonstrated a decreased false-positive ratio, contributing to a reduction in the time required for quality control. This makes it suitable for effectively handling the substantial volumes of data generated by DAS MMF. The findings indicated that the CNN method exhibited a reduced false-trigger rate and led to an expansion in the size of the microseismic event catalogue. A CNN model can automatically detect microseismic events (hypocenter location) recorded by low-SNR DAS with MMF cable data with excellent model accuracy and superior efficiency. Subsequently, the result of microseismic comparison with a 3C geophone revealed that DAS images have a lower sensitivity in weak events and low frequencies compared to geophones. The innovative aspect of this work lies in the implementation of ML algorithms, specifically the CNN model workflow. This workflow can be retrained as additional data become accessible, enabling the implementation of transfer learning. In other words, it becomes feasible to train a network using DAS data from one well and subsequently apply it to DAS data obtained from other wells. This approach presents an effective way to automatically address the DAS processing workflow through the use of an imaging algorithm. The author indicated that the limitation of this work was that the number event detection results of the 3C geophone was more detectable than that of DAS-CNN. However, the outcomes indicated reasonable counts of microseismic event occurrences observed during each step, indicating the reliability of the CNN algorithm. Future work could focus on developing more automated tools to make use of intricate phase data, including reflections, guided waves, and coda waves. A summary of the relevant literature on microseismic detection and ML can be found in
Table 5 below.
4.6. Subsurface Imaging of Combined SMF and MMF
The development of subsurface imaging by utilizing DAS technology rely on many elements, like fiber optic cables, which enable compatibility with an interrogator and data acquisition. The selection of optical SMF and MMF for DAS measurement is contingent upon the specific application and desired aim. Combined SMF and MMF have not been widely employed for the purpose of DAS measurement as a prevalent practice. DAS commonly depends on SMF cables, owing to its capacity to preserve signal integrity at extended distances and deliver measurements with high resolution [
79]. However, it is conceivable that advancements have taken place in the scope of fiber optics and distributed sensing, considering the potential evolution of technology and practices in subsurface imaging. The combination of the two types of fiber optics is usually contained inside a single cable equipped with optical SMF and MMF, where MMF is commonly used for DTS measurements and SMF for DAS measurements.
The authors of [
28] carried out simultaneous acquisition of DAS VSP with SFM and MMF cables at the Aquistore CO
2 storage facility. This field acquisition also utilized a (3C) wireline geophone array for quantitative comparison with DAS measurement results. The fiber optic cables comprised one MMF cable loop utilized for DTS measurements, alongside one SMF cable employed for telemetry to downhole gauges and to conduct DAS measurement. During the first well completion process, the cables were installed outside of the well casing using cement. The simultaneous DAS VSP acquisitions of SMF were used as data for the full 3D surface explosive survey as an active source, and MMF was used for a significant fraction of the survey. All active sources shot data were measured with a geophone array, and a 3C VSP geophone and DAS fiber optics were used for recording. The obtained raw measurement data were recorded as strain rate, with further data processing converting them to particle velocity for equivalence with geophones for comparison. The SMF and MMF measurement results were compared within the same source shot map. The comparison result indicated that that there was no quantitatively significant distinction in the quality of the two recordings. However, it was observed that the SMF data exhibited a slightly higher level of noise. This discrepancy is likely attributed to a little variation in the used interrogator units rather than being influenced by the type of fiber deployed. Other findings include that MMF was quite efficient because of the numerous wellbores in which MMF was deployed for DTS measurement. The comparison of DAS VSP data with geophone response was consistent, showing similar trace gather within each shot point. Further DAS VSP data processing in weighted stacking revealed in improvement in SNR and 3D migrated DAS VSP imaging. Data acquired from SMF and MMF with identical coupling in the fiber optic demonstrated comparable SNRs and sensitivity. The comparable SNR and sensitivity of MMF installed in a borehole for DTS can be considered for VSP acquisition as well. The result of DAS VSP migrated image volume in this field acquisition at the Aquistore site for continuous monitoring can be expanded to 3D imaging in future work.
In the study reported in [
85], DAS measurements of seismic properties were conducted on the Store Glacier ice sheets, recording englacial and subglacial seismic properties. DAS measurements were recorded along seismic shots as an active survey to obtain P-waves and S-waves at a depth of 1043 m in the formation, and VSP and data were sampled at 10 m with vertical resolution. The fiber optic cables were installed enclosed in a gel-filled, stainless-steel capillary tube in a borehole on Store Glacier. This cable consisted of two SMFs for DAS measurements and four MMFs for DTS measurements. The interrogator unit was set to a 4000 Hz sampling frequency, with a sample length every 1 m along the cable and a 10 m gauge length for approximately half the expected seismic wavelength. A seismic source was generated using a sledgehammer along the shot 1 m away from the well surface and 500 m offset from the wellbore. The fiber optic cables were recorded continuously without any source. The DAS VSP measurements were able to image 0 m offset and 300 m offset VSP and detect direct SV-waves and P-waves. The utilization of the DAS recording technique facilitated the quantification of changes in the properties of P- and SV-waves. This enabled the identification of transitions associated with ice-crystal fabric and environmental temperature, which influenced the seismic signal. Additionally, the presence of a subglacial layer composed of consolidated silt was successfully identified. The measurement result indicated that DAS offers evident advantages in the characterization of seismic properties of ice masses with a high image resolution. It is also worth noting that DAS cables possess the capability to be monitored for DTS, enabling the combined analysis of seismic and thermal data. Nevertheless, a limitation of DAS is the disadvantage of cables in terms of applicability as a surfaced receiver in P-wave seismic surveys; however, the ability of DAS to image distinct seismic responses at roughly 10 m intervals throughout the ice column can compensate for these drawbacks. Subsequently, the deployment of DAS interrogators incurs significant expenses, but these costs can be mitigated by the resulting improvement in survey effectiveness. Once permanently deployed, the DAS cable consists of SMF and MMF, enabling continuous field monitoring to focus solely on source deployment. In comparison, traditional seismic surveys necessitate ongoing efforts for the relocation and operation of both source and receiver systems. In future work, MMF cables for DTS measurement should be observed in the same vertical borehole to verify the depth in DAS VSP.
The research reported in [
86] applied DAS technology for seismic exploration to image subsurface geothermal reservoirs in magmatic environments. DAS technology was used to mitigate the deficiency of downhole seismic survey results for subsurface geothermal exploration. The fiber optic cables were equipped with an SMF for DAS and MMF for DTS measurement, operating at temperatures above 400 °C; hence, the cable was coated with copper. The conceptual acquisition design of the DAS cable installation was based on a surface layout (buried in a trench) of optical cable deployed in the x and y directions and along the wellbore installation in the z direction to achieve a three-dimensional cable array. The cable installation result indicated that the main effect on the SNR of the measurement was caused by the coupling between the wellbore and fiber optics. DAS data were recorded along fiber optic cables with a 5.9 km cable length and settings of spatial resolution of 1 m and measured as strain rate. A seismic vibrator truck was used for DAS measurement as its active source. Subsequently, conventional VSP used a swept-impact seismic technique as an active source. Further data preprocessing steps such as shift and stack correlation were performed to generate estimated SNR and processed along with data recorded by a seismic station network. Measured DAS and geophone vertical components were compared in terms of SNR at certain depth intervals, indicating that significant result in terms of quantitative SNR values. The measured signals of surface optical fiber and a 3C geophone were compared to detect the signal generated by the seismic vibrator with an appropriate SNR value. Combining the DAS data with data from a seismic network deployed on the surface increased the resolution of the seismic survey imaging. These research findings also highlight the use of DAS as an option to overcome the operating temperature limitation of conventional geophones, as well as the specifications within a magmatic geothermal environment required for fiber optic cable material to be able to tolerate the geothermal reservoir temperature. When used to image a subsurface geothermal reservoir, the result proved that DAS is well-suited to be used for seismic applications in high-temperature wells.
The research reported in [
87] discusses the sensing sensitivity improvement of a surface-deployed DAS fiber configuration to detect steep-angle P-wave reflections. The aim of this DAS measurement was to enhance the sensitivity of the existing surface DAS tests conducted at the site. The sensitivity was improved by augmenting the total vertical alignment of the fiber inside experimental setups consisting of both helical and straight fiber segments. The test was conducted using fiber optic cables that consisted of two SMFs and two MMFs placed inside a stainless-steel tube and wrapped with 0.61 mm stainless-steel armor wires and a polyethylene jacket. For the horizontal configuration, the cable was placed along a shallow trench. The specified DAS VSP settings of the acquisition parameters were a gauge length of 10 m, spatial sampling of 1.02 m, and a fiber cable length up to 4288 m. The DAS recording was measured as strain rate. During the 3D seismic survey conducted at the designated site for the DAS test configurations, data were collected simultaneously. A total of 401 dynamite shots used as active sources were captured by the DAS test cable. The raw data for single dynamite shots and NMO-corrected receiver gathers were compared to assess their similarity to geophone data. The variability in the responses of the various configurations along the test cable was observed to proceed with significant data processing, NMO correction, and trace stacking. The DAS configurations of these measurements, the buried geophone, and a receiver from the permanent DAS fiber were compared, indicating surface coupling within amplitude variation offset (AVO) plots referring to the data. Subsequently, numerical modeling was applied to assess the DAS response of the asymmetric helical fiber. This modeling result indicated that the recorded field response was mostly influenced by the fiber shape rather than the borehole coupling. The results of the sensitivity analysis showed that the asymmetric helix and vertical straight fiber configurations had the maximum sensitivity increases with steep-angle P-wave reflections, which were attained by decreasing the surface-wave response and increasing the signal response. However, the research result indicated that it was still associated with uncertainties in locating channels along the fiber and gauge length within the shot configuration. For future DAS testing, increasing the configuration cable length to the gauge length is necessary to associate the seismic measurement observed on multiple channels and reduce the uncertainties of channel authorization using tap tests to obtain the best cable coupling of the DAS fiber.
The research reported in [
88] involved a time lapse of DAS VSP imaging in the Aquistore CO
2 reservoir. The Aquistore project utilized 4D VSP to perform volumetric reservoir assessment and locate the injected CO
2 plume over time. Further processing and resulting imaging discrepancies in the VSP data for extremely closely spaced images were analyzed to estimate the time-lapse noise from field recordings. During the data imaging process, the reduced SNR of the DAS data was mitigated by a noise attenuation algorithm as measured by NRMS values. Nevertheless, it is clear that the SNR of the DAS data was lowered by the NRMS attenuation of noise. The DAS measurements used fiber optic cables consisting of two MMFs for DTS and two SMFs for DAS measurement. The cable was deployed inside stainless-steel production tubing clamped to the outside of the well casing and cemented in place, as illustrated on
Figure 8. The DAS data were recorded using a 10 m gauge length and 2.036 m sensor spacing and measured as strain rate. The obtained DAS VSP data were further processed and converted to high time-lapse NRMS values and an image of the horizon of the reservoirs. The processing and imaging results indicated that the reservoir baseline exhibited fluctuating SNR characteristics. However, it still achieved an acceptable level of repeatability to identify time-lapse noise. The research outcome highlights that DAS VSP is considered a useful technique for time-lapse monitoring of CO
2 reservoirs. For future enhancements in DAS VSP results, more attempts should be made with respect to the processing step and the development of its framework.
All relevant literature discussed in this section is summarized in
Table 6 below.