Advances in geoscience are pivotal in revolutionising energy supply, improving resource management, and addressing environmental challenges. By integrating cutting-edge technologies and innovative methods, geoscientists are paving the way for a more sustainable and efficient future in energy production and resource utilisation.
The integration of robotics and artificial intelligence (AI) is transforming the field of geosciences, significantly enhancing the exploration and management of energy resources. This paper illustrates how robotics and AI are revolutionizing exploration, transitioning from conventional oil and gas to emerging energy resources, including geothermal, transition minerals, and innovative carbon management solutions. It will also highlight internally developed robotic technologies that enhance both efficiency and safety in exploration.
One notable technology is the GeoDrone, an advanced Unmanned Aerial Vehicle (UAV) designed to collect and analyze 3D models of geological outcrops in hard-to-reach areas using digital twin capabilities. A recent field trial successfully demonstrated the GeoDrone’s ability to capture detailed models and images of a 150-meter escarpment in central Arabia, navigating complex terrains while ensuring safety from hazards. The high-resolution data showcased the effective imaging and spatial accuracy of advanced robotics and AI solutions in automated data collection and processing.
Alongside GeoDrone, the Autonomous Seismic Acquisition Device (ASAD) enhances seismic data collection with a fleet of UAVs equipped with advanced sensors. ASAD operates autonomously, utilizing robotic sensors that can position themselves to record seismic data. A recent pilot involving 16 ASADs successfully gathered data under real-world conditions in 45°C desert environments, demonstrating its autonomous functionality and effectiveness in challenging fields while also increasing safety by minimizing manual intervention.
Complementing ASAD in marine environments is SpiceRack, an innovative Autonomous Underwater Vehicle (AUV) designed to collect seismic data from the ocean floor efficiently. It is powered by two lithium-ion batteries that provide operational power for up to 70 continuous days. A recent deployment of 200 SpiceRack units demonstrated coordinated command and control as a swarm, ensuring consistent coupling and signaling. These vehicles successfully navigated to assigned locations, recorded seismic data, and returned to a collection point, showcasing improved navigation, positioning, and the application of AI techniques for managing multiple AUVs simultaneously.
These advancements not only improve data acquisition efficiency but also reduce operational costs compared to traditional methods. By enabling autonomous systems to navigate complex environments and scale operations, it becomes ideal for maximizing discovery and unlocking new frontiers. As the energy sector moves towards lower carbon footprints and increased digitalization, robotics and AI will continually expand the role of geosciences through multidisciplinary approaches in addressing current and future exploration challenges.
Co-author/s:
Abdulrahman Alshuhail, Manager of Geophysics Technology Division at EXPEC ARC, Aramco.
One notable technology is the GeoDrone, an advanced Unmanned Aerial Vehicle (UAV) designed to collect and analyze 3D models of geological outcrops in hard-to-reach areas using digital twin capabilities. A recent field trial successfully demonstrated the GeoDrone’s ability to capture detailed models and images of a 150-meter escarpment in central Arabia, navigating complex terrains while ensuring safety from hazards. The high-resolution data showcased the effective imaging and spatial accuracy of advanced robotics and AI solutions in automated data collection and processing.
Alongside GeoDrone, the Autonomous Seismic Acquisition Device (ASAD) enhances seismic data collection with a fleet of UAVs equipped with advanced sensors. ASAD operates autonomously, utilizing robotic sensors that can position themselves to record seismic data. A recent pilot involving 16 ASADs successfully gathered data under real-world conditions in 45°C desert environments, demonstrating its autonomous functionality and effectiveness in challenging fields while also increasing safety by minimizing manual intervention.
Complementing ASAD in marine environments is SpiceRack, an innovative Autonomous Underwater Vehicle (AUV) designed to collect seismic data from the ocean floor efficiently. It is powered by two lithium-ion batteries that provide operational power for up to 70 continuous days. A recent deployment of 200 SpiceRack units demonstrated coordinated command and control as a swarm, ensuring consistent coupling and signaling. These vehicles successfully navigated to assigned locations, recorded seismic data, and returned to a collection point, showcasing improved navigation, positioning, and the application of AI techniques for managing multiple AUVs simultaneously.
These advancements not only improve data acquisition efficiency but also reduce operational costs compared to traditional methods. By enabling autonomous systems to navigate complex environments and scale operations, it becomes ideal for maximizing discovery and unlocking new frontiers. As the energy sector moves towards lower carbon footprints and increased digitalization, robotics and AI will continually expand the role of geosciences through multidisciplinary approaches in addressing current and future exploration challenges.
Co-author/s:
Abdulrahman Alshuhail, Manager of Geophysics Technology Division at EXPEC ARC, Aramco.
Accurate permeability estimation is vital step to reservoir analysis and management. However, it is very challenging in carbonate reservoirs because of their intrinsic heterogeneity, dual porosity systems, and complex pore geometries. The conventional approaches, including empirical porosity–permeability transforms or log-based correlations, typically do not describe the degree of variability in flow units. By advent of advanced Nuclear Magnetic Resonance (NMR) logs, valuable pore-size distribution data provided by means of transverse relaxation times (T2), and permeability is usually determined from models such as the Schlumberger-Doll Research (SDR) and Timur–Coates equations. Despite their proven effectiveness, NMR logs are costly and not always available. In contrast, Dipole Sonic Imager (DSI) tools are routinely logged in comparison with NMR log data, generating monopole, dipole, and Stoneley waveform data. Although Stoneley slowness and attenuation have been studied as permeability indicators, the richness of data within the full DSI waveform has not yet been tapped. In this paper, we propose a new signal-derived property for permeability prediction from the Hilbert envelope under-curve area (UEA) of DSI waveforms.
The workflow includes:
Application to a carbonate oil reservoir demonstrates that the above-described methodology correlates core permeability higher than NMR-SDR prediction. This hybrid approach bridges acoustic and NMR physics and demonstrates the untapped potential of DSI waveform attributes for petrophysical applications in wells where NMR measurements are unavailable.
Keywords: Permeability, Carbonate reservoir, Dipole Sonic Imager (DSI), Acoustic Waveform, Hilbert transform, Envelope under-curve area (UEA), NMR T2-lm, SDR model.
Co-author/s:
D. Hasanvand, Pasargad Energy Development Company (PEDC).
A. Heravi, Pasargad Energy Development Company (PEDC).
The workflow includes:
- DSI monopole, dipole, and Stoneley waveforms extraction;
- Hilbert transform envelope calculation for each waveform;
- Computation of UEA as a proxy of waveform energy dissipation;
- Extraction of logarithmic mean relaxation time (T2-lm) from NMR logs; (5) scaling of UEA versus T2-lm to derive a semi–T2-lm hybrid parameter; and (6) application of a modified SDR equation to calculate permeability values.
Application to a carbonate oil reservoir demonstrates that the above-described methodology correlates core permeability higher than NMR-SDR prediction. This hybrid approach bridges acoustic and NMR physics and demonstrates the untapped potential of DSI waveform attributes for petrophysical applications in wells where NMR measurements are unavailable.
Keywords: Permeability, Carbonate reservoir, Dipole Sonic Imager (DSI), Acoustic Waveform, Hilbert transform, Envelope under-curve area (UEA), NMR T2-lm, SDR model.
Co-author/s:
D. Hasanvand, Pasargad Energy Development Company (PEDC).
A. Heravi, Pasargad Energy Development Company (PEDC).
Objectives/Scope:
Carbonate rocks reservoirs characterized by complex natural fractures system combined with large dissolution karstic features. Moreover, they have extensive and extremely complex secondary porosity system. Acquiring good, reliable, and interpretable image logs that clearly resolve the rock texture is always challenging particularly while encountering partial and complete fluid loses while drilling.
Methods, Procedures, Process:
The conventional way of acquiring image logs when losses are encountered in the well is through conveying the tool down the hole through the drill pipe which is called TLC (tough logging conditions) which is time consuming. It has been observed that the image quality received with using the forementioned process is not up to the expectations. Although the LWD would give quite good results, the need of 6-arm mechanical calipers is important for completion decisions. Therefore, the need for a new way of conveying Wireline tool was crucially needed which is eventually achieved by introducing the new slim hole imaging through bit technology.
Results, Observations, Conclusions:
Both thru the bit imaging technology and high resolution LWD imager allowed to confidently perform fracture analysis. Both technologies allow the quantification of the fracture attributes, which is a key step in the field development of the naturally fractured reservoirs. The fractures attributes quantification includes orientation, density, porosity, and hydraulic fracture aperture. However, the high-resolution micro-resistivity through bit imaging technology allowed to acquire a higher resolution image that can clearly resolve the rock texture particularly, hence providing a more confident image interpretation that will eventually impact the completion design optimization very positively. Moreover, the presence of heterogeneity in carbonates, pose a challenge for the characterization of such rocks. The identification of textural variations advanced techniques in borehole image analysis have been applied. Post-processing advanced porosity spectrum analysis methods delivers new insight into previously established interpretations of the reservoir.
The new imaging technology helped to reduce the time of the static conditions of the wells and hence reduced the possibility of losing more muds to the formation across the fractured and karstified zones before acquiring the image log, which enabled acquiring a high-quality image across the full hole circumference unlike the image acquired on TLC operation.
Novelty/Significance/Additive Information:
In this paper; The results of all imaging technologies will be discussed in details through couple of case studies. It was concluded that both Thru the bit and LWD provide good image in fractured carbonate. However, thru the bit operation improves the efficiency, save rig time, enhanced the completion design and reservoir characterization through the good image quality acquired.
Carbonate rocks reservoirs characterized by complex natural fractures system combined with large dissolution karstic features. Moreover, they have extensive and extremely complex secondary porosity system. Acquiring good, reliable, and interpretable image logs that clearly resolve the rock texture is always challenging particularly while encountering partial and complete fluid loses while drilling.
Methods, Procedures, Process:
The conventional way of acquiring image logs when losses are encountered in the well is through conveying the tool down the hole through the drill pipe which is called TLC (tough logging conditions) which is time consuming. It has been observed that the image quality received with using the forementioned process is not up to the expectations. Although the LWD would give quite good results, the need of 6-arm mechanical calipers is important for completion decisions. Therefore, the need for a new way of conveying Wireline tool was crucially needed which is eventually achieved by introducing the new slim hole imaging through bit technology.
Results, Observations, Conclusions:
Both thru the bit imaging technology and high resolution LWD imager allowed to confidently perform fracture analysis. Both technologies allow the quantification of the fracture attributes, which is a key step in the field development of the naturally fractured reservoirs. The fractures attributes quantification includes orientation, density, porosity, and hydraulic fracture aperture. However, the high-resolution micro-resistivity through bit imaging technology allowed to acquire a higher resolution image that can clearly resolve the rock texture particularly, hence providing a more confident image interpretation that will eventually impact the completion design optimization very positively. Moreover, the presence of heterogeneity in carbonates, pose a challenge for the characterization of such rocks. The identification of textural variations advanced techniques in borehole image analysis have been applied. Post-processing advanced porosity spectrum analysis methods delivers new insight into previously established interpretations of the reservoir.
The new imaging technology helped to reduce the time of the static conditions of the wells and hence reduced the possibility of losing more muds to the formation across the fractured and karstified zones before acquiring the image log, which enabled acquiring a high-quality image across the full hole circumference unlike the image acquired on TLC operation.
Novelty/Significance/Additive Information:
In this paper; The results of all imaging technologies will be discussed in details through couple of case studies. It was concluded that both Thru the bit and LWD provide good image in fractured carbonate. However, thru the bit operation improves the efficiency, save rig time, enhanced the completion design and reservoir characterization through the good image quality acquired.
Mineralogy-based petrophysical evaluation in near real-time has been traditionally utilized for optimal well placement in deep gas clastic reservoirs. The minimum required logs were Spectral Gamma-Ray, Density, Neutron, Resistivity and Sonic. Due to the presence of silt, which reduces permeability, intervals of similar porosity from Multimin (MM) vary significantly in reservoir performance. This paper will demonstrate a new workflow that requires less logging tools to provide a robust petrophysical evaluation.
A Shaly-Sand Analysis (SSA) model for the area is developed using Gamma-Ray, Density, Neutron and Resistivity only. The petrophysical results such as porosity and water saturation were calibrated to the MM petrophysical analysis. The SSA petrophysical results are also correlated to the Formation Tester (FT) pressure test analysis to verify the permeability zones. The lithology from SSA is comparable to Mudlogs lithology interpretation. The established model will be used across the silty clastic reservoirs for a specific area.
The correlation between porosity, lithology, and FT results shows that the most significant factor that reduces permeability is the volume of silt in the tested layers. The SSA allows the quantification of silt for each layer which results in improving the FT operation by identifying the best zone to test and reducing the time of screening. The SSA total porosity and water saturation values matched the reference values computed using Multimin (MM) approach. Total porosity and water saturation from MM were used as reference since the models were core calibrated for consistency of the basic petrophysical parameters across the studied field. In addition, the calibrated SSA enables the identification of the sweet spot while drilling the horizontal wells by avoiding highly silty clean zones, which can indicate good total porosity.
The proposed workflow identifies sweet spots in horizontal sandstone wells by utilizing minimum logs that LWD companies can provide as inputs to the SSA. The silt volume quantification takes into consideration further factors to characterize permeability than relying only on the porosity since it can be misleading. This approach reduces overall logging costs and potentially improves well productivity. In addition, it leverages the power of big data into creating performance calibrated models that support well placement.
Co-author/s:
Mohammed F. Alzayer, Lead Petroleum Engineer, Saudi Aramco.
Layal N. Alhussain, Petroleum Engineer, Saudi Aramco.
Ali E. Alqunais, Lead Petroleum Engineer, Saudi Aramco.
A Shaly-Sand Analysis (SSA) model for the area is developed using Gamma-Ray, Density, Neutron and Resistivity only. The petrophysical results such as porosity and water saturation were calibrated to the MM petrophysical analysis. The SSA petrophysical results are also correlated to the Formation Tester (FT) pressure test analysis to verify the permeability zones. The lithology from SSA is comparable to Mudlogs lithology interpretation. The established model will be used across the silty clastic reservoirs for a specific area.
The correlation between porosity, lithology, and FT results shows that the most significant factor that reduces permeability is the volume of silt in the tested layers. The SSA allows the quantification of silt for each layer which results in improving the FT operation by identifying the best zone to test and reducing the time of screening. The SSA total porosity and water saturation values matched the reference values computed using Multimin (MM) approach. Total porosity and water saturation from MM were used as reference since the models were core calibrated for consistency of the basic petrophysical parameters across the studied field. In addition, the calibrated SSA enables the identification of the sweet spot while drilling the horizontal wells by avoiding highly silty clean zones, which can indicate good total porosity.
The proposed workflow identifies sweet spots in horizontal sandstone wells by utilizing minimum logs that LWD companies can provide as inputs to the SSA. The silt volume quantification takes into consideration further factors to characterize permeability than relying only on the porosity since it can be misleading. This approach reduces overall logging costs and potentially improves well productivity. In addition, it leverages the power of big data into creating performance calibrated models that support well placement.
Co-author/s:
Mohammed F. Alzayer, Lead Petroleum Engineer, Saudi Aramco.
Layal N. Alhussain, Petroleum Engineer, Saudi Aramco.
Ali E. Alqunais, Lead Petroleum Engineer, Saudi Aramco.
While renewable energies are rapidly expanding, natural gas will remain a key fossil fuel in the coming decades. As a cleaner-burning and relatively low-carbon energy source, it will continue to play an essential role as a transition fuel in the global energy mix. Its importance lies not only in supporting energy security but also in complementing renewables, particularly in regions with rapidly growing demand. Therefore, the discovery and development of new gas reservoirs in the Persian Gulf remain a matter of high scientific and economic significance. The Faraghan and Zakin clastic formations in the Persian Gulf, equivalent to the Pre-Khuff formations in Arabian countries, represent major Paleozoic successions with strong potential as gas reservoirs. While seismic surveys, well logs, core analyses, and production tests form the backbone of modern exploration programs, regional geological studies remain indispensable for interpreting petroleum systems and guiding local decision-making. This study develops a regional geological model by investigating source rocks, reservoir intervals, seal rocks, hydrocarbon-bearing structures, and the tectono-stratigraphic history of the Arabian Plate from the Paleozoic to the present. By linking hydrocarbon traps to structural evolution and basin dynamics, the study proposes exploration strategies aimed at improving the prediction and discovery of hydrocarbons within Paleozoic clastic successions of the Persian Gulf. Evaluation of source rock potential identifies the Qusaiba Shale Member (equivalent to the Sarchahan Formation) as a prolific Silurian source rock widely distributed across the Arabian Plate. Its hydrocarbon generation window, opening about 150 million years ago, coincided with key phases of trap development, ensuring the presence of critical petroleum system elements in the region. Two main phases of trap formation are recognized: the first around 150 Ma, related to tectonic activity along the Qatar-Fars Arch and the emplacement of salt structures, and the second during the Oligocene–Miocene (20–30 Ma), which further enhanced structural complexity and trap integrity. Reactivation of pre-existing north–south faults during the Hercynian orogeny provided migration pathways, while salt tectonics created additional structural traps and conduits for hydrocarbon movement. Comparative assessment of reservoir quality among the Unayzah, Jauf, Jubah, and Tawil formations in Arabian countries, alongside the Faraghan and Zakin formations in the Persian Gulf, highlights intervals with favorable porosity and permeability. Similarly, evaluation of potential seal rocks underscores the need for testing key stratigraphic intervals—particularly in Iranian sectors of the Gulf—to verify their sealing capacity. By integrating tectonic history, structural evolution, and stratigraphic architecture, this study establishes a regional exploration framework that enhances understanding of hydrocarbon distribution. The results provide a scientific basis for designing more effective exploration strategies and maximizing the gas discovery potential of Paleozoic clastic reservoirs in the Persian Gulf region.
Keywords: Persian Gulf, Arabian Plate, Paleozoic clastic reservoirs, Faraghan Formation, Zakin Formation, Sarchahan Formation, hydrocarbon system, regional geology.
Keywords: Persian Gulf, Arabian Plate, Paleozoic clastic reservoirs, Faraghan Formation, Zakin Formation, Sarchahan Formation, hydrocarbon system, regional geology.
Horizontal wells are widely used in oilfields because of its high initial productivity, large drainage area and the ability to connect multiple remaining oil rich areas. Mature oilfields in Bohai Bay are facing with the problem of high-water cut 95%, scattered remaining oil, and the thinner potential sand-bodies, which led to challenges in the implementation of horizontal wells. Detailed reservoir description can boost the confidence and success probability of horizontal well drilling. This paper presents a fine reservoir classification method to improve the accuracy of reservoir description in multi-layer sandstone reservoirs.
We have established a workflow for hierarchical reservoir description. Firstly, according to sand thickness, permeability and oil saturation, the sands are divided into three categories: I, II and III.(Fig.1)
For category I -sand thickness 8m~20m which is basically in the critical region seismic resolution 1/4λ , but the internal configuration of the composite sand body is unclear. With the discontinuous interface analysis technology of the accumulated energy of the ant body, the braided river channel can be accurately located. Moreover, by using the discontinuous interface analysis method of the minimum curvature of the root mean square amplitude, the impermeable boundary channels and the internal small channels can be clearly distinguished.(Fig.2)
For category II -sand thickness 4m~8m which is much less than seismic resolution 1/4λ. The waveform indication simulation method is used to predict the reservoir distribution based on the actual acoustic logging curve and seismic waveform matching.(Fig.3)
For category III-sand thickness 1m~4m, the seismic signals only have a certain response to the thin layer groups. Therefore, an intelligent method driven by massive seismic-geological models is used to predict the favorable areas of the thin layer groups.(Fig.3).
A pilot area for tapping potential of mature field with large-scale maximum reservoir contact (MRC)wells has been developed. 53 wells include horizontal-wells, multilateral-wells and multi-bottom wells have been drilled. The initial water cut of new wells is less than30%. F38H1 is the first high inclination horizontal-well in Bohai Bay Basin with 72°hole inclination angle. F10H1 was the first horizontal well drilled in a category II -sand (4m) with a 246m length with an initial oil production of 300 barrels per-day.(Fig.4).
This paper demonstrates a new method to improve the accuracy of reservoir description by means of multidisciplinary collaboration among seismic, logging, geology, and reservoir disciplines. It can not only broaden engineers' horizons and boost the confidence of horizontal wells, but also increase the recovery factor of mature oilfields and serve as a reference for the similar reservoir.
We have established a workflow for hierarchical reservoir description. Firstly, according to sand thickness, permeability and oil saturation, the sands are divided into three categories: I, II and III.(Fig.1)
For category I -sand thickness 8m~20m which is basically in the critical region seismic resolution 1/4λ , but the internal configuration of the composite sand body is unclear. With the discontinuous interface analysis technology of the accumulated energy of the ant body, the braided river channel can be accurately located. Moreover, by using the discontinuous interface analysis method of the minimum curvature of the root mean square amplitude, the impermeable boundary channels and the internal small channels can be clearly distinguished.(Fig.2)
For category II -sand thickness 4m~8m which is much less than seismic resolution 1/4λ. The waveform indication simulation method is used to predict the reservoir distribution based on the actual acoustic logging curve and seismic waveform matching.(Fig.3)
For category III-sand thickness 1m~4m, the seismic signals only have a certain response to the thin layer groups. Therefore, an intelligent method driven by massive seismic-geological models is used to predict the favorable areas of the thin layer groups.(Fig.3).
A pilot area for tapping potential of mature field with large-scale maximum reservoir contact (MRC)wells has been developed. 53 wells include horizontal-wells, multilateral-wells and multi-bottom wells have been drilled. The initial water cut of new wells is less than30%. F38H1 is the first high inclination horizontal-well in Bohai Bay Basin with 72°hole inclination angle. F10H1 was the first horizontal well drilled in a category II -sand (4m) with a 246m length with an initial oil production of 300 barrels per-day.(Fig.4).
This paper demonstrates a new method to improve the accuracy of reservoir description by means of multidisciplinary collaboration among seismic, logging, geology, and reservoir disciplines. It can not only broaden engineers' horizons and boost the confidence of horizontal wells, but also increase the recovery factor of mature oilfields and serve as a reference for the similar reservoir.
Introduction:
Fractures appear as sinusoidal curves on borehole images (BHI’s) in horizontal wells and provide key insights into reservoir performance, they can indicate highly productive zones or highlight potential concerns for waterflood shortcuts. Traditionally interpretations are manual and is inherently time-consuming, laborious, and subject to interpreter bias and variability. Automating this process could result in significant cost savings allowing geological expertise to be focused elsewhere. However, computer vision algorithms face substantial challenges due to complexities in data quality, particularly in a Logging While Drilling (LWD) environment. LWD Images often experience distortions, noise, missing data patches, and artifacts, exacerbated by downhole shocks and vibrations that tend to increase with drilling depth. These imperfections can obscure true fractures or create patterns that mimic the sinusoidal shape of fractures, leading to high rates of false positives in automated detection. We aimed to develop an AI-assisted interpretation method to address many of these challenges.
Method & Applications:
Recognizing the impact of data quality on fracture interpretation reliability, we address the subjective and meticulous task of image quality labeling. We developed a deep learning-based workflow using an Efficient Net CNN, fine-tuned with borehole images classified as 'good', 'fair', and 'poor'. The resulting model achieves an F1 score of 80%, providing consistent and rapid identification of reliable data sections for interpretation whether manual or automated.
This work also presents a dual approach to improve the accuracy and efficiency of fracture analysis from borehole images. First, we define a set of image quality criteria and design corresponding scoring measures. Emphasis is placed on image quality features, as artifacts and missing data can significantly impact detection either by obscuring actual fractures or introducing false ones. These quality features are leveraged to minimize false positives while preserving true fracture detections, initially identified using a pre-existing sinusoid detection algorithm. This filtering process identifies key discriminative features that help distinguish genuine natural fractures from false detections. We managed to reduce false detections by 98% while keeping 94% of true fractures.
In fields developed with ERD wells Borehole Images are a critical element to understand reservoir performance. Human interpretations can be highly subjective, AI-Assisted methods significantly improves the repeatability and consistency required in field wide reservoir modeling and understanding reservoir performance. In addition, AI-Assisted methods significantly reduce interpretation times from days to less than an hour.
Conclusions:
In conclusion, AI-Assisted methods have been used to significantly improve the timelines and robustness of Borehole Image interpretation.
Fractures appear as sinusoidal curves on borehole images (BHI’s) in horizontal wells and provide key insights into reservoir performance, they can indicate highly productive zones or highlight potential concerns for waterflood shortcuts. Traditionally interpretations are manual and is inherently time-consuming, laborious, and subject to interpreter bias and variability. Automating this process could result in significant cost savings allowing geological expertise to be focused elsewhere. However, computer vision algorithms face substantial challenges due to complexities in data quality, particularly in a Logging While Drilling (LWD) environment. LWD Images often experience distortions, noise, missing data patches, and artifacts, exacerbated by downhole shocks and vibrations that tend to increase with drilling depth. These imperfections can obscure true fractures or create patterns that mimic the sinusoidal shape of fractures, leading to high rates of false positives in automated detection. We aimed to develop an AI-assisted interpretation method to address many of these challenges.
Method & Applications:
Recognizing the impact of data quality on fracture interpretation reliability, we address the subjective and meticulous task of image quality labeling. We developed a deep learning-based workflow using an Efficient Net CNN, fine-tuned with borehole images classified as 'good', 'fair', and 'poor'. The resulting model achieves an F1 score of 80%, providing consistent and rapid identification of reliable data sections for interpretation whether manual or automated.
This work also presents a dual approach to improve the accuracy and efficiency of fracture analysis from borehole images. First, we define a set of image quality criteria and design corresponding scoring measures. Emphasis is placed on image quality features, as artifacts and missing data can significantly impact detection either by obscuring actual fractures or introducing false ones. These quality features are leveraged to minimize false positives while preserving true fracture detections, initially identified using a pre-existing sinusoid detection algorithm. This filtering process identifies key discriminative features that help distinguish genuine natural fractures from false detections. We managed to reduce false detections by 98% while keeping 94% of true fractures.
In fields developed with ERD wells Borehole Images are a critical element to understand reservoir performance. Human interpretations can be highly subjective, AI-Assisted methods significantly improves the repeatability and consistency required in field wide reservoir modeling and understanding reservoir performance. In addition, AI-Assisted methods significantly reduce interpretation times from days to less than an hour.
Conclusions:
In conclusion, AI-Assisted methods have been used to significantly improve the timelines and robustness of Borehole Image interpretation.
Real-time formation evaluation and wellbore stability assessment can be crucial to drilling, logging, well-placement, well-testing and completions operations. High temperature environments limit the utilization of certain logging wireline (WL) and Logging-while drilling (LWD) technologies in extreme high temperature conditions or in hole sizes smaller than 5.5”. Although the technology has advanced to increase the number of measurements available in high and extreme high temperatures, some LWD sensors are not yet available in all sizes and environments. The integration of LWD and advanced mud logging provides a viable real-time petrophysical and wellbore stability evaluation option leading to successful wellbore logging operations.
The expected temperature plays a crucial role in deciding on the optimum tools to be utilized in the logging program. Static or near-static temperature data from offset wells is evaluated to provide an estimate of the expected downhole temperature while drilling. Circulating temperatures, however, can be quite different and are usually lower than the static temperature in vertical wells, in high angle deep wells the temperature increase is subjective to drilling friction and might increase beyond the static temperature. Temperature evaluation in real time for different logging passes provides a key input for selecting the optimum tools to be used. Temperature build-up is time-dependent and in high-temperature regimes, temperature monitoring and projection becomes more critical for avoiding downhole tool failures or temperature stress. Integration of advanced cuttings analysis with LWD provides a solution for evaluating the lithology, porosity, fluid composition, and wellbore stability in real time while drilling high temperature wells.
Temperature monitoring for each hole section and for multiple passes has been used for selecting the right tool rating and contingency plan based on the expected temperature. Variations in temperature gradients and temperature build-up rates have been observed in drilling and circulating temperatures, as well as differences between vertical and horizonal wells. In horizontal wells, the friction causing by drillpipe interaction with the formation has sometimes been seen to drive drilling temperatures higher than static temperatures. X-ray fluorescence (XRF) and X-ray diffraction (XRD) advanced cuttings analyses have been integrated with the downhole measurements to provide a comprehensive formation-evaluation workflow. Natural gamma ray spectroscopy (NGS) from cuttings or Gamma ray logs derived from elemental analysis can be used for depth matching and have served as an alternative measurement in cases where LWD tools could not be used in extreme high temperatures.
This study introduces a customized workflow for evaluating high-temperature wells through monitoring the temperature gradient, the circulating to near-static difference, and alternative surface cuttings solutions to provide a detailed formation-evaluation program.
Co-author/s:
Noor Albasheer, Petroleum Engineer, Saudi Aramco.
Mohamed Fouda, Geoscience Advisor, Halliburton.
Ahmed Taher, Geoscience Manager, Halliburton.
The expected temperature plays a crucial role in deciding on the optimum tools to be utilized in the logging program. Static or near-static temperature data from offset wells is evaluated to provide an estimate of the expected downhole temperature while drilling. Circulating temperatures, however, can be quite different and are usually lower than the static temperature in vertical wells, in high angle deep wells the temperature increase is subjective to drilling friction and might increase beyond the static temperature. Temperature evaluation in real time for different logging passes provides a key input for selecting the optimum tools to be used. Temperature build-up is time-dependent and in high-temperature regimes, temperature monitoring and projection becomes more critical for avoiding downhole tool failures or temperature stress. Integration of advanced cuttings analysis with LWD provides a solution for evaluating the lithology, porosity, fluid composition, and wellbore stability in real time while drilling high temperature wells.
Temperature monitoring for each hole section and for multiple passes has been used for selecting the right tool rating and contingency plan based on the expected temperature. Variations in temperature gradients and temperature build-up rates have been observed in drilling and circulating temperatures, as well as differences between vertical and horizonal wells. In horizontal wells, the friction causing by drillpipe interaction with the formation has sometimes been seen to drive drilling temperatures higher than static temperatures. X-ray fluorescence (XRF) and X-ray diffraction (XRD) advanced cuttings analyses have been integrated with the downhole measurements to provide a comprehensive formation-evaluation workflow. Natural gamma ray spectroscopy (NGS) from cuttings or Gamma ray logs derived from elemental analysis can be used for depth matching and have served as an alternative measurement in cases where LWD tools could not be used in extreme high temperatures.
This study introduces a customized workflow for evaluating high-temperature wells through monitoring the temperature gradient, the circulating to near-static difference, and alternative surface cuttings solutions to provide a detailed formation-evaluation program.
Co-author/s:
Noor Albasheer, Petroleum Engineer, Saudi Aramco.
Mohamed Fouda, Geoscience Advisor, Halliburton.
Ahmed Taher, Geoscience Manager, Halliburton.
