Amr Ramadan

Amr Ramadan

Auto-generated profile, based on publicly available information

Amr Ramadan

Department of Petroleum Engineering, University of Houston, Houston, Texas, USA

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
4 October 2024 in SPE Journal
Verified
Journal Article
Closed access
AGAmr Gharieb
Amr Gharieb
Ahmed Farid IbrahimAhmed Farid Ibrahim
Ahmed Farid Ibrahim
Mohamed Adel GabryMohamed Adel Gabry
Mohamed Adel Gabry
MEMohamed Elsawy
Mohamed Elsawy
AAAhmed Algarhy
Ahmed Algarhy
NDNihal Darraj
Nihal Darraj
Summary: Reservoir characterization is critical for successful oil and gas exploration, heavily reliant on detailed formation analysis from formation microimager (FMI) logs. However, interpreting these complex logs is time-consuming, subjective, and requires expert-level knowledge. This study addresses this challenge by proposing a novel approach that integrates computer vision (CV) and deep learning (DL) for automated and real-time interpretation of FMI logs. Our methodology leverages CV and DL techniques for automated feature extraction and classification from FMI data. A comprehensive training data set encompassing more than 2,500 images is utilized to train the DL model, enabling the identification of more than 50 distinct geological features, including bedding planes, fractures, and mineral variations. In addition, the study explores the creation of local stress maps from drilling-induced fractures to determine the present-day maximum horizontal stress (SHmax) orientation, crucial for optimizing wellbore stability and hydraulic fracturing strategies. This research presents a groundbreaking advancement in reservoir characterization through the synergy of automated FMI interpretation, CV, and DL. The developed model exhibits exceptional accuracy in geological feature classification, significantly surpassing traditional, human-centric interpretations. For instance, the model achieves a remarkable 92% accuracy in classifying ore than 50 geological features, demonstrably outperforming conventional methods. Furthermore, the developed model was applied to actual field cases to predict the stress field. The model was able to accurately predict the minimum horizontal stress (Shmin) and SHmax based on FMI logs, and the results were used to refine the geomechanical modeling and optimize hydraulic fracture orientation for enhanced hydrocarbon recovery. This work establishes a new benchmark for applying artificial intelligence in subsurface analysis, paving the way for future advancements in reservoir management and geomechanics. The implications are far-reaching, offering greater precision in geological interpretations, improved decision-making for production strategies, and ultimately, a more sustainable approach to hydrocarbon extraction. By automating tedious and subjective tasks, this approach not only reduces reliance on experts but also frees up valuable time for more strategic tasks. The ability to extract critical geological information with such accuracy from complex FMI logs translates to significant improvements in reservoir characterization and production efficiency.

17 April 2025 in SPE Journal
Verified
Journal Article
Closed access
Mohamed Adel GabryMohamed Adel Gabry
Mohamed Adel Gabry
ARAmr Ramadan
Amr Ramadan
MSMohamed Y. Soliman
Mohamed Y. Soliman
Summary: At the end of a hydraulic fracturing treatment, the pump shuts down, causing a water hammer (WH) effect to occur. The WH is influenced by formation properties and fracture complexity. In this study, we use the continuous wavelet transform (CWT) technique to analyze WHs, interpreting damping coefficients of the fracturing stages as indicators of induced fracture complexity within the same reservoir. We introduce a model that treats a WH as a damped harmonic oscillator, leveraging this analogy to understand WH signatures. In this model, the WH signal is viewed as a damped harmonic oscillator. CWT is applied to the post-fracture wellhead signal across multiple scales, using the complex Morlet wavelet. Ridge detection traces dominant frequencies, forming a “ridge” that indicates the system’s natural frequency. Envelope calculation isolates this ridge, capturing amplitude and normalizing it to highlight decay characteristics. The damping coefficient is estimated from the linear regression envelope, where the slope represents the logarithmic decline. In low-permeability reservoirs, fracture complexity and tortuosity strongly influence the damping of the WH signal. The methodology has been applied to various fracture stages, showing a strong correlation between damping coefficients and the number of propagating fractures. This technique automates the estimation of damping coefficients, linking fracture intensity and damping behavior, providing a valuable tool for evaluating induced fracture complexity.

1 June 2026 in SPE Journal
Verified
Journal Article
Closed access
Mohamed Adel GabryMohamed Adel Gabry
Mohamed Adel Gabry
ARAmr Ramadan
Amr Ramadan
MSM. Y. Soliman
M. Y. Soliman
Summary: Stimulated reservoir volume (SRV) remains a critical yet difficult-to-measure parameter in unconventional reservoir development, particularly when microseismic monitoring is unavailable. With this study, we present a data-driven workflow that integrates normalized continuous wavelet transform (CWT) analysis of treating-pressure data, a pretrained deep learning (DL) model, and production-based rate transient analysis (RTA) to estimate and calibrate SRV using stimulation-time measurements only. The workflow builds upon prior work that established a CWT-DL mapping between normalized pressure scalograms and relative microseismic displacement vectors and extends it through a true blind, pad-level validation and production-consistent SRV calibration. The proposed methodology is evaluated using publicly available Marcellus Shale Energy and Environment Laboratory (MSEEL) data. A DL model trained exclusively on the MIP-3H and MIP-5H wells is applied, without retraining, to six Boggess pad wells that were excluded from model development. Predicted event clouds are converted into volumetric SRV proxies using geometric estimators, and the results are independently evaluated against SRV derived from a complete RTA workflow, including flow-regime diagnostics and numerical history matching. Across the Boggess pad, pressure-derived SRV estimates show strong agreement with RTA-derived SRV, with well-level differences ranging from approximately 2% to 11%. To enable meaningful interpretation of these differences, coordinate-level prediction errors from the validation data set were quantified and evaluated at the event-cloud geometry level. The observed discrepancies are therefore interpreted in the context of event-cloud prediction uncertainty, geometric sensitivity in cloud-to-volume conversion, and RTA nonuniqueness, rather than from point estimates alone or as a deterministic truth comparison. The results demonstrate that, within basin-consistent geomechanical conditions and comparable completion practices, treating-pressure data can be used to estimate flow-effective SRV in a production-consistent manner. The proposed workflow provides a practical, pressure-only alternative for SRV calibration when recorded microseismic data are unavailable, while explicitly documenting uncertainty sources, limitations, and requirements for data quality and transferability.

1 September 2026 in SPE Journal
Verified
Journal Article
Closed access
ARAmr Ramadan
Amr Ramadan
CQCole Qamar
Cole Qamar
MGMohamed Gabry
Mohamed Gabry
Summary: Wellbore instability and near-bit rock failure during drilling operations drive significant nonproductive time, stuck-pipe incidents, and costly remedial actions, yet conventional wellbore-stability analyses rely on static geomechanical assumptions that cannot adapt to evolving subsurface conditions ahead of the bit. With this study, we present a data-driven framework that couples an adaptive random forest (ARF) classifier with a multitarget particle swarm optimization (PSO) scheme to deliver horizon-explicit, near-bit rock-failure probabilities at 5-ft, 10-ft, and 20-ft (≈1.5-m, 3.0-m, and 6.1-m) look-ahead distances. Historical wireline logs [gamma ray, bulk density, neutron porosity, and photoelectric factor (PEF)] are harmonized on a 0.25-ft measured-depth grid with surface and downhole drilling channels [weight on bit (WOB), torque, rotary speed, rate of penetration (ROP), standpipe pressure, and flow rate], producing a 5,521-sample data set that covers a 1,400-ft vertical section with 256 labeled breakdown events. A sliding-window feature engineering library generates 20 inputs spanning first- and second-order trend descriptors, physics-informed drilling-efficiency indicators [mechanical specific energy (MSE) computed per the Dupriest-Macpherson convention, and torque-to-WOB ratio], and four lithology proxies. PSO tunes eight ARF hyperparameters against an F1-averaged multitarget fitness, while a depth-based 70/15/15 split and a transfer-learning scenario (pretrain on upper 70% of well and fine-tune with the next 5% of new well samples) are evaluated to emulate realistic deployment. The full-feature ARF-PSO attains macro-F1 scores of 0.91, 0.88, and 0.84 for the three horizons and outperforms a conventional random forest (RF) baseline by 6–11 F1 points; a drilling-only variant that omits logs which are unavailable during drilling (PEF and bulk density) retains 90% of full-feature performance. Tree Shapley additive explanations (TreeSHAP) attributions, a class-imbalance sensitivity study (positive rate swept from 46% down to 5%), and a Dupriest-Macpherson-consistent MSE formulation provide interpretability and robustness evidence that the method generalizes beyond the balanced single-well case. Limitations around single-well validation, image-log labeling uncertainty, and wireline/logging-while-drilling (LWD) depth bias are discussed, and a multiwell extension leveraging Wellsite Information Transfer Standard Markup Language (or WITSML) streaming ingestion is outlined as the next step toward staged field validation. A crosswell generalization study across 10 offset wells totaling 100,000 ft of logged interval confirms that the framework retains 85–88% of its single-well macro-F1 after transfer-learning fine-tuning, providing preliminary evidence of field-level portability. Limitations around label-source reliability, wireline/LWD depth bias, and the need for broader basin-to-basin validation are discussed.