Variable interaction empirical relationships and machine learning provide complementary insight to experimental horizontal wellbore cleaning results

Aug 4, 2026

Upload by Finn

Long horizontal wellbore sections are essential for developing tight oil and gas reservoirs but introduce distinct borehole-cleaning and cuttings-transport challenges. Experimental work typically reports borehole cleaning outcomes as multivariate empirical relationships using dimensionless cuttings bed thickness/concentration (H%). This study extracts complementary pairwise H% relationships from published data via interpolated trends and optimizers, then applies five machine learning models to a compiled 10-variable dataset. Results show pairwise optimizer-derived relationships predict H% with RMSE below 1.8%, with extreme gradient boosting yielding the lowest errors. When local information for multiple variables exists, machine learning is more reliable than empirical models; when only limited variables are available, pairwise empirical relationships remain valuable for drilling decisions.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages13

Reading Time33 min

AccessFree to download

Introduction

Borehole cleaning challenges in horizontal wells

Factors and limitations of existing empirical models

Study approach and objective

Advances in Geo-Energy Research Vol. 9, No. 3, p. 172-184, 2023 Original article Variable interaction empirical relationships and machine learning provide complementary insight to experimental horizontal wellbore cleaning results David A. Wood * DWA Energy Limited, Lincoln LN59JP, United Kingdom Keywords: Hole-cleaning factors cuttings carrying performance cuttings transport feature importance optimized empirical relationships cuttings-bed concentrations Cited as: Wood, D. A. Variable interaction empirical relationships and machine learning provide complementary insight to experimental horizontal wellbore cleaning results. Advances in Geo-Energy Research, 2023, 9(3): 172-184 . [https://doi.org/10.46690/ager.2023.09.05](https://doi.org/10.46690/ager.2023.09.05) Abstract: Long horizontal wellbore sections are now a key requirement of oil and gas drilling, particularly for tight reservoirs. However, such sections pose a unique set of borehole-cleaning challenges which are quite distinct from those associated with less inclined wellbores. Experimental studies provide essential insight into the downhole variables that influence borehole cleaning in horizontal sections, typically expressing their results in multivariate empirical relationships with dimensionless cuttings bed thickness/concentration (H %) . This study demonstrates how complementary empirical H % relationships focused on pairs of influential variables can be obtained from published experimental data using interpolated trends and optimizers. It also applies five machine learning algorithms to a compiled multivariate (10-variable) interpolated dataset to illustrate how reliable H % predictions can be derived based on such information. Seven optimizer-derived empirical relationships are derived using pairs of influential variables which are capable of predicting H % with root mean squared errors of less than 1.8% . The extreme gradient boosting model provides the lowest H % prediction errors from the 10-variable dataset. The results suggest that in drilling situations where sufficient, locally-specific, information for multiple influential variables is available, machine learning methods are likely to be more effective and reliable at predicting H % than empirical relationships. On the other hand, in drilling conditions where information is only available for a limited number of influential variables, empirical relationships involving pairs of influential variables can provide valuable information to assist with drilling decisions. 1. Introduction Long horizontal wellbore sections are now recognized asthe most effective way to develop tight reservoirs, and the length of the horizontal sections drilled is increasing particularly in certain shale formations to optimize resource recovery. Hole cleaning and drilling cuttings transport is a challenge for wellbores of all configurations and inclinations due to the large number of variables that influence it (Li and Walker, 2001) . For horizontal wellbore sections, it is a particular challenge because cuttings beds tend to form more readily on the lower side of the wellbore due to gravitational forces and often slowdown cuttings transport (Sun et al., 2013) . Such cuttings beds, if they are allowed to remain in the horizontal sections exert high frictional and torque forces on the drill pipe leading to inefficiency and slow rates of penetration (ROP) (Mahmoud et al., 2020a) . As the length of horizontal sections increases so do the impacts of the cuttings beds on borehole cleaning efficiency and the ability to move logging tools and completion equipment in and out of a wellbore (Nazari et al., 2010) . Inefficient borehole cleaning leading to thick cuttings beds accumulating in horizontal sections is a major cause of drilling problems and non-productive time (Power et al., 2000) . Potential problems include the drill string becoming stuck and “packoffs” blocking the circulation of the drilling f

FAQs

Why are horizontal wellbore sections harder to clean than other wellbores?

Long horizontal sections promote cuttings bed formation on the lower side due to gravity, which can slow cuttings transport and increase friction and torque, reducing efficiency and ROP.

How does the study obtain complementary empirical H% relationships from existing experimental data?

It uses interpolated trends and optimizers to generate pair-focused empirical H% relationships from published experimental datasets.

When are machine learning methods likely to outperform empirical relationships for predicting H%?

Machine learning is likely more effective and reliable when sufficient, locally specific information is available for multiple influential variables.

Aug 4, 2026

Upload by Finn