Cement Bond Quality Prediction Based on Wide and Deep Neural Networks
Bibliographic record
Abstract
Abstract Cementing quality is critical for the safety, efficiency, and cost-effectiveness of oil and gas wells. However, predicting cementing quality remains highly challenging due to the numerous influencing factors and their interdependencies. Current cementing quality prediction methods are still in the developmental stage, making it difficult to build robust models that meet the stringent requirements of cementing design. This study collected cementing data from an oilfield in western China. Various data-cleaning techniques were applied to the raw data, effectively addressing issues with missing values and anomalies. Subsequently, correlation analysis was conducted to eliminate irrelevant variables, and the dataset for model construction was finalized. Based on the correlation analysis and the characteristics of cementing data, a cement bond quality prediction model was developed using a Wide & Deep Neural Network. The model simultaneously predicts the cement bond quality of both the casing-cement sheath and the cement sheath-formation interfaces. Given the significant impact of thickening time on cementing outcomes, this study incorporated the Arrhenius equation to correct thickening time under varying temperature and pressure conditions, embedding this correction into the neural network. This further enhanced the model’s accuracy. The final model achieved a prediction accuracy of 87.5% for the casing-cement sheath interface and 89.5% for the cement sheath-formation interface, enabling the prediction of wellbore quality prior to cementing operations. Finally, the interpretability of the intelligent model was analyzed using the SHAP (SHapley Additive exPlanations) method, which identified the factors with the most significant impacts on cement bond quality after model training. The proposed approach provides valuable guidance towards the optimization of cementing operations.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 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 teacher head, 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".