Uncertainty and Explainable Analysis of Machine Learning Model for Reconstruction of Sonic Slowness - Ensemble Learning and SHAP-Based Quality Control

Aug 3, 2026

Upload by Patrick

Borehole logs underpin lithology identification and estimation of subsurface oil and gas reserves, yet critical logs are often missing in horizontal or aged wells. The study reconstructs missing compressional and shear wave slowness (DTC, DTS) logs from other logs acquired in the same borehole using NGBoost ensemble learning, which outputs both predictions and uncertainty. SHAP analysis provides interpretability, quantifying input-log importance and coupling. NGBoost is compared with Random Forest, GBDT, XGBoost, and LightGBM to evaluate reconstruction quality and confidence intervals.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages41

Reading Time1 h 43 min

AccessFree to download

Summary

Keywords

Introduction

Uncertainty and Explainable Analysis of Machine Learning Model for Reconstruction of Sonic Slowness Logs Hua Wang1*, Yuqiong Wu 1, Yushun Zhang 1, Fuqiang Lai2, Zhou Feng3, Bing Xie4, Ailin Zhao4 1. School of Resources and Environment, University of Electronic Science and Technology of China, Chengdu, Sichuan Province, P.R. China, 611731； 2. School of Petroleum Engineering, Chongqing University of Science and Technology, Chongqing, P.R. China, 401331 3. CNPC Research Institute of Petroleum Exploration and Development 4. Exploration and Development Research Institute, PetroChina Southwest Oil and Gas Field Company, Chengdu, 610041 Summary Logs are valuable information for oil and gas fields as they help to determine the lithology ofthe formations surrounding the borehole and the location and reserves of subsurface oil and gas reservoirs. However, important logs are often missing in horizontal or old wells, which poses a challenge in field applications. To address this issue, conventional methods involve supplementing the missing logs by either combining geological experience and referring data from nearby boreholes or reconstructing them directly using the remaining logs in the same borehole. Nevertheless, there is currently no quantitative evaluation for the quality and rationality of the constructed log. In this paper, we utilize data from the 2020 machine learning competition of the Society of Petrophysicists and Logging Analysts (SPWLA), which aims to predict the missing compressional wave slowness (DTC) and shear wave slowness (DTS) logs using other logs in the same borehole. We employ the natural gradient boosting (NGBoost) algorithm to construct an Ensemble Learning model that can predicate the results as well as their uncertainty. Furthermore, we combine the SHAP (SHapley Additive exPlanations) * Correspoding author, [Email: huawang@uestc.edu.cn](Email: huawang@uestc.edu.cn) method to investigate the interpretability of the machine learning model. We compare the performance of the NGBosst model with four other commonly used Ensemble Learning methods, including Random Forest, GBDT, XGBoost, LightGBM. The results show that the NGBoost model performs well in the testing set and can provide a probability distribution for the prediction results. This distribution allows petrophysicists to quantitatively analyze the confidence interval of the constructed log. In addition, the variance of the probability distribution of the predicted log can be used to justify the quality of the constructed log. Using the SHAP explainable machine learning model, we calculate the importance of each input log to the predicted results as well as the coupling relationship among input logs. Our findings reveal that the NGBoost model tends to provide greater slowness prediction results when the neutron porosity (CNC) and gamma ray (GR) are large, which is consistent with the cognition of petrophysical models. Furthermore, the machine learning model can capture the influence of the changing borehole caliper on slowness, where the influence of borehole caliper on slowness is complex and not easy to establish a direct relationship. These findings are in line with the physical principle of borehole acoustics. Finally, by using the explainable machine learning model, we observe that although we did not correct the effect of borehole caliper on the neutron porosity log through preprocessing, the machine learning model assigned a greater importance to the influence of the caliper, achieving the same effect as caliper correction. Keywords: Missing Log Construction, Ensemble Learning, NGBoost, Quality Control 1. Introduction Ensuring control over a country's oil and gas resources requires reducing exploration cost and increasing production efficiency. One effective means of achieving this is borehole logging, a widely used technique for determining the properties of subsurface reservoirs(Asquith et al. 2004) . However, borehole lo

FAQs

Why are missing borehole logs a problem in oil and gas field applications?

Missing logs hinder determination of surrounding formation properties and reservoir location and reserves, increasing uncertainty in interpretation and leading to suboptimal decisions for production and management.

How does the proposed method reconstruct missing sonic slowness logs?

It reconstructs compressional and shear wave slowness (DTC, DTS) using NGBoost ensemble learning trained on other logs from the same borehole.

How is uncertainty and model interpretability assessed?

NGBoost provides a probability distribution so confidence intervals and variance can justify reconstruction quality, while SHAP quantifies the importance of each input log and reveals coupling relationships among inputs.

Aug 3, 2026

Upload by Patrick