First EAGE Workshop on Surface Logging

This document presents techniques for supporting the prediction of pore pressure and fracture gradient in real-time within non/shale formations (such as sandstones and carbonates), where traditional shalebased methods are not applicable. Using SLB’s PreVue service along with LWD/MWD tools, drilling parameters, and gas logging data, the approach helps to interpret indicators like mud losses, fluid gains and specially connection gas peaks.

The document emphasizes the importance of properly adjusting monitoring graph scales to detect events that might otherwise go unnoticed. A standardized workflow and plotting template are proposed to aid real-time interpretation. This methodology has been successfully applied in the Gulf of Mexico projects, enhancing understanding of the pressure regime during drilling operations-even without direct measurements.

This study highlights the transformative value of automated drill cuttings analysis as a real-time geoscience tool, especially when downhole tools like MWD and LWD are limited by cost, failure, or extreme environments. By integrating robotics and elemental analysis techniques—namely XRF and LIBS—the system extracts geochemical and mechanical insights from surface-collected cuttings. The process involves automatic collection, cleaning, and pulverization of cuttings, followed by high-resolution elemental analysis. Outputs include modeled mineralogy, brittleness, TOC, elemental gamma ray (EGR), and mechanical proxies, with quality control through standards and barium contamination checks.

Machine learning models, trained using PCA and DFA, classify chemofacies with >90% accuracy. Results from over 864 wells demonstrate broad utility: EGR curves replaced failed MWD in the Haynesville and Eagle Ford; geosteering improved through accurate facies classification; redox proxies mapped TOC-rich intervals; silica spikes identified bit wear zones; clay shifts predicted overpressure hazards; and elemental trends helped map subsalt faults and Jurassic boundaries in the Gulf of Mexico.

This workflow significantly enhances subsurface understanding, reduces non-productive time (NPT), and enables data-driven decisions in real-time. Its field-proven success makes it valuable not only in hydrocarbon development but also in carbon storage and other subsurface applications.

This study presents an empirically derived formula for estimating water saturation (Sw) in oil reservoirs using surface logging data. Specifically, Sw is proportional to both mud log total gas analysis (TGA) and reservoir gas-oil ratio (GOR), and inversely proportional to drilling mud viscosity in overbalanced drilling systems. The approach enables real-time screening of reservoir productivity and early evaluation of reservoir quality where wireline data is limited or unavailable.

Algeria’s Silurian formation in a western field poses significant challenges in reservoir characterization and production. These challenges include deep and unstable boreholes, heterogeneous reservoirs with low resistivity pay zones, and substantial uncertainties in determining fluid contacts. The instability and complexity of the reservoirs, combined with low-resistivity pay zones, make it difficult to accurately identify and characterize hydrocarbon-bearing intervals. Accurate determination of the gas/water contact is crucial for reserves estimation, completion design, and development planning but remains challenging due to geological and petrophysical complexities.

Addressing these challenges requires a multidisciplinary and coordinated strategy. Overcoming these issues involves optimizing drilling operations, utilizing advanced logging and interpretation techniques, and integrating multiple data sources. An effective approach includes using advanced mud gas (AMG) technology for early formation fluid characterization and fluid contact identification, followed by a wireline formation tester (WFT) for precise contact determination and formation water identification.

Porosity and permeability are important parameters reflecting flow potential in subsurface. In Standardized practices quantify these parameters via core analysis, well logging, and well testing. Nonetheless, these methods are generally costly and mainly applied to reservoir sections. The emphasis on the identification of flow potential in overburden sections is commensurate with concerns about drilling safety and efficient well plugging operations. The study aims to explore an alternative method to quantifying the abovementioned parameters, specifically porosity, in the overburden sections. By leveraging our in-house database and machine learning (ML) techniques, we established a workflow that could yield a machine learning model for estimating the porosity in the overburden sections. We gathered data for 23 wells from the Norwegian Sea with similar geological properties. After preprocessing, we divided the data into training and testing datasets, with the training dataset being input into the SparkBeyond discovery platform for feature engineering. Through several iterations, we refined our ML model for further testing. Our findings exhibit the significant potential of utilizing ML to quantify the parameters and lay the groundwork for future research in this domain.

The study area is located at Block #26 within the Salinas Basin, specifically in Deep Water at the southeastern Gulf of Mexico, where geological complexities in the Upper Eocene formations were observed. In this stage, the region has a distinctive geological history, marked by the interaction between the Caribbean and the southern Gulf of Mexico. The main challenges faced in the region were caused by sediment provenance, diagenesis, authigenic mineral formation, and hydrothermal alteration.

The methodology used a multi-faceted approach, integrating elemental ratios from XRF analysis, mineralogical identification through XRD, and chemostratigraphical evaluation. Our study emphasises a robust Quality Control framework during sample acquisition to prevent mixing and ensure accurate depth allocation. Our research monitored the potential contribution of radioactive minerals in non-shaly rocks, particularly in high Th, U, and K₂O values, a critical limitation associated with downhole GR logs. The results reveal a distinctive signature in the elemental ratios, with an increase in MgO, Cr, Ni, and Co, coupled with a depletion in K₂O, Al₂O₃, and CaO. Mineralogical shifts identified through XRD have confirmed a marked decrease in the Total Clays and Total Carbonates, along with a rise in Quartz, Feldspars Minerals, and the occurrence of Enstatite and Chrysotile.

In southern Kuwait, a complex carbonate reservoir within the Lower Cretaceous formation has been characterized by micritic and wackestone facies with variable dolomitization, resulting in heterogeneous porosity and permeability. Its compartmentalized nature—divided by impermeable barriers such as tight carbonates or shale—poses challenges for hydrocarbon extraction ( Khan et al., 2013 ). To enhance production, horizontal drilling and acid fracturing are employed to connect isolated reservoir sections. Effective fracturing depends on the accurate identification of open natural fractures, which are typically detected using expensive wireline logging tools, such as formation micro-imagers or acoustic logs. This study presents a cost-effective alternative: using surface logging to monitor drilling fluid dynamics through a high-accuracy sensor. By analyzing micro-losses in conjunction with drilling parameters (e.g., rate of penetration and weight on bit), this method can detect open fractures, induced fractures, or permeable zones in real time. Field trials conducted in South Kuwait demonstrated a strong correlation between the surface-logging-derived fracture indicators and wireline log data. This technique offers a cost-effective solution, improves drilling efficiency, improve drilling safety, and supports more targeted completion designs—potentially enhancing production outcomes in a tight carbonate reservoir by optimizing acid fracturing stage placement.

Surface formation evaluation (SFE) acts as a continuous geochemical log of the encountered fluids along a well and these data are crucial for making real-time decisions such as geosteering, optimizing downhole sampling and well testing programs. The goal of the Mud Gas Advisor (MGA) workflow is to provide a real-time Interpretation Confidence Index (ICI) and to suggest data corrections or hardware solutions to effectively manage various influencing factors. Different from hardware QC, MGA’s novelty is that it enriches the user with the knowledge of confidence level of mud gas interpretation of the downhole fluid type, regardless the type of hardware used and the changing impact of the drilling and environmental conditions. Hence, the ICI offers a new real-time continuous SFE industry standard for mud gas logging across all service levels.

Given the absence of universal calibration standards for GPC, careful selection and control of system parameters are essential to minimize measurement errors. Prior to initiating large-scale data acquisition, it was essential to optimize the GPC-UV methodology. In early system configuration, the void time was excessively long, approximately 54 minutes, before any signal was detected. The primary objective was to reduce elution time and establish optimal operating conditions. Extensive laboratory testing was conducted to refine the setup, including evaluation of parameters such as the number of columns, mobile solvent type, solvent flow rate, operating temperature, sample injection volume and amount. A total of sixty stock tank oil samples, with API gravities ranging from 11° to 50°, were analyzed alongside with their corresponding wet drill cutting samples, collected from various fields across the Norwegian Continental Shelf (NCS). After observing the fluid response from GPC, the associated cutting samples were compared against the fluid sample. And both data exhibits a remarkable degree of similarity, indicating a strong correlation between the fluid and cutting sample profiles.

This study demonstrates how the integration of AI-driven techniques can transform drill cuttings into a powerful tool for reservoir characterization. A proprietary workflow that builds a detailed lithotype model by combining X-ray fluorescence (XRF) elemental data and high-resolution image analysis was used to analyze 278 samples from two European wells that targeted Mesozoic reservoirs.

Standardized imaging in both white light and ultraviolet light was part of the laboratory workflow, and 32 elements were subjected to XRF analysis. RGB and YUV color information, as well as particle metrics, were extracted from the images. Combining XRF and Image data, the results were used to classify samples into silicarich or carbonate-rich domains (via Si/Ca ratio) and further subdivided into 14 lithotypes based on brightness values.

The identification of depositional cycles, patterns, and Gross Sedimentary Packages (GSPs) was made possible by this classification. For example in Well B, cleaner intervals in the middle GSP suggested higher reservoir quality, whereas dark-colored, clay-rich, and reducing conditions in the upper GSP may be advantageous for hydrocarbon preservation.

With the aim of improving well planning and reservoir management throughout the region, this workflow provides a quick, quantitative, and repeatable method for understanding depositional environments and refining geological models.