Multiple Sweet Spot Detection in Shales Using a Hybrid Data-Driven Machine Learning Technique
Bibliographic record
Abstract
Abstract The objective of this study is to develop a hybrid data-driven machine learning (ML) technique, the XGBoost-MLP, capable of identifying multiple sweet spots defined in this paper as "source rocks simultaneously characterized by suitable values of brittleness index (BI), permeability (k), porosity (ϕ), oil saturation (So), and bulk volume water (BVW). The multiple sweet spots allow to attain commercial production through horizontal wells which are hydraulically fractured in multiple stages." In the proposed method, the BI is determined based on mineralogical data. The ML training is performed using Jarvie's BI estimation equation. The model is validated against BI derived from mechanical properties through well log interpretation. Both methods prove valuable in identifying variables that influence BI. The hybrid data-driven technique, which involves neural networks embedded in the XGBoost model, is effective in identifying brittle intervals and predicting sweet spots through pattern recognition. This methodology addresses the challenges posed by the complexity of target distributions and the variability within the training dataset, making it suitable for heterogeneous reservoirs. Results are illustrated using source-rock data from the heterogeneous Eagle Ford Shale of Texas with the support of the Eagle Ford and Pimienta shales in Mexico, and La Luna shale in the Middle Magdalena Valley and Catatumbo Basins of Colombia. However, the methodology is not case-dependent and consequently could prove valuable in other shale reservoirs around the world. Pattern recognition identifies the multiple sweet spots defined in the first paragraph above. The BVW helps to recognize the maturity of the source. This is so because "BVW is the only parameter that shows an unrelenting, unidirectional reduction during burial history" (Devine, 2014; Olusola and Aguilera, 2018). The BI signals the intervals with the best success probabilities of hydraulic fracturing. The largest permeabilities and porosities point to intervals with natural fractures characterized by nano and micro apertures, and the largest oil saturations to the intervals with the best possibilities of hydrocarbon production. Merging the capabilities of pattern recognition in Pickett plots and machine learning can yield improved insights in tasks related to hydrocarbon exploration as well as production. This ability to identify data patterns represents a breakthrough in both ML and the Pickett plot. The novelty of the paper is the introduction of a model methodology for locating multiple sweet spots in shale petroleum reservoirs through the integration of ML and the power of pattern recognition in Pickett plots. To our knowledge this integration has not been attempted in the past.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".