Application of Artificial Neural Networks to geological classification: porphyry prospectivity in British Columbia and oil reservoir properties in Iran
Bibliographic record
Abstract
Seismic facies analysis aims to classify oil and gas reservoirs into geologically and petrophysically meaningful rock groups, or classes. An artificial neural network (ANN) is a versatile and efficient tool for classifying data or estimating subsurface properties from large geophysical datasets. This tool can provide critical information for oilfield development and reservoir characterization. This study includes application of artificial neural networks on two different datasets: 1) geophysical characterization of an oil reservoir in Iran and 2) geological prospectivity for porphyry in British Columbia, Canada. In the first case study, I utilize seismic attributes, well-log data, and core data analysis and use supervised machine learning techniques to efficiently estimate the acoustic impedance and porosity of the reservoir and to classify it into four lithological classes. Seismic attributes as inputs for our techniques capture the lithological patterns or structural characteristics in the seismic amplitude, phase, frequency, and other complex seismic properties that cannot be directly seen in the original seismic images. Selection of an optimal set of input features from the vast number of possible mathematical transformations of seismic data is a critical task for reservoir property prediction and classification. This selection is performed by standard as well as innovative procedures employing properties of the target classes. Three different supervised approaches to non-linear classification are used: 1) the so-called probabilistic neural network (PNN), 2) conventional ANN, and 3) an ANN with the new approach of optimal attribute selection. For each of these approaches, images of classification confidence levels and confidence-filtered class images are produced. Assessments of the robustness and accuracy of seismic facies classification is performed for each of these algorithms. The ANN classifiers are validated using validation and test data subsets. The proposed algorithm shows a higher performance, particularly in comparison with the PNN algorithm. Several visualization techniques are used to examine and illustrate the power of the ANN-based approaches to classify the seismic facies with high accuracy. However, the three approaches still provide significantly different levels of lateral continuity, frequency content, and classification accuracy. Therefore, some level of expert assessment is still required when using machine learning for reservoir interpretation. In the second case study, I use an ANN to explore the prospectivity for porphyry within the Quesnel Terrane, BC, Canada. A purely data-driven approach based on geophysical, structural, and volcanic-age data results in a predictive prospectivity map which correlates well with known mineral occurrences and suggests new areas for potential exploration.
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 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.003 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 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".