Scholarlink

<p>Predicting compressional slowness (DTCO) from non-sonic logs can reduce acquisition cost, fill data gaps, and support field planning. We evaluate blind cross-well DTCO prediction on two offshore Newfoundland &amp; Labrador wells using a strictly leakage-free, features-only strategy: causal lag windows are built from past non-sonic logs and all sonic/sonic-derived channels are excluded. The pipeline includes deterministic depth conditioning, relative-depth features, multi-scale depth derivatives, rank-aggregated feature selection, and time-aware validation on the training well. We compare three model families: Random Forest (RF), Extreme Gradient Boosting (XGBoost), and a BiLSTM. In this setting, tuned XGBoost with the top 20 predictors and a 10-sample lag attains blind cross-well performance of <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(R^2=0.895\)</EquationSource> </InlineEquation>, MAE<InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(=11.38~\mu \mathrm {s/m}\)</EquationSource> </InlineEquation>, RMSE<InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(=15.12~\mu \mathrm {s/m}\)</EquationSource> </InlineEquation> when trained on Well&#xa0;1 and tested on Well&#xa0;2; the reverse direction is lower, indicating inter-well distribution shift. RF performs competitively in several configurations, whereas BiLSTM underperforms on these data. Overall, rigorous leakage control, depth-aware feature engineering, and principled feature selection are key drivers of performance, and tree-based ensembles provide strong, data-efficient baselines for cross-well pseudo-sonic prediction.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Cross-well machine learning prediction of sonic logs in Newfoundland and Labrador

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月15日

查看全文
  • Bahare Zare,
  • Mohammad Mojammel Huque,
  • Lesley A. James,
  • Hamid Usefi

摘要

Predicting compressional slowness (DTCO) from non-sonic logs can reduce acquisition cost, fill data gaps, and support field planning. We evaluate blind cross-well DTCO prediction on two offshore Newfoundland & Labrador wells using a strictly leakage-free, features-only strategy: causal lag windows are built from past non-sonic logs and all sonic/sonic-derived channels are excluded. The pipeline includes deterministic depth conditioning, relative-depth features, multi-scale depth derivatives, rank-aggregated feature selection, and time-aware validation on the training well. We compare three model families: Random Forest (RF), Extreme Gradient Boosting (XGBoost), and a BiLSTM. In this setting, tuned XGBoost with the top 20 predictors and a 10-sample lag attains blind cross-well performance of \(R^2=0.895\) , MAE \(=11.38~\mu \mathrm {s/m}\) , RMSE \(=15.12~\mu \mathrm {s/m}\) when trained on Well 1 and tested on Well 2; the reverse direction is lower, indicating inter-well distribution shift. RF performs competitively in several configurations, whereas BiLSTM underperforms on these data. Overall, rigorous leakage control, depth-aware feature engineering, and principled feature selection are key drivers of performance, and tree-based ensembles provide strong, data-efficient baselines for cross-well pseudo-sonic prediction.