Scholarlink

<p>During well planning, accurately determining geomechanical parameters is crucial to ensure wellbore stability during drilling, especially in complex reservoirs. A thorough assessment of these factors optimizes the drilling and well completion processes. Among the primary stresses, minimum horizontal stress plays a key role in hydraulic fracturing design and wellbore stability. However, directly measuring this stress is labor-intensive and costly, highlighting the need for efficient, alternative methods. This study focuses on the Niger Delta Basin, where unconsolidated reservoirs and overpressured shales are common, by presenting a well-log scale 1D geomechanical model and a data-driven approach to predict minimum horizontal stress in seven wells within the Eocene Agbada Formation. Using industry-standard equations, relevant geomechanical parameters were calculated, and six machine learning models were applied to conventional log data to predict minimum horizontal stress. Results showed that the sandstone sediments’ pore pressure gradient values ranged from 0.56 to 0.60 psi/ft, reaching 0.69 to 0.71 psi/ft in overpressured shales. The 1D model revealed a narrower mud window in overpressured zones. Among the models, gradient boosting achieved the highest accuracy, with an R<sup>2</sup> of 0.92 and the lowest MAE of 273.53 on test data. Blind testing on additional wells validated the model’s robustness and low error rate. These machine-learning results can significantly reduce the time, manpower, and resources typically required for direct measurements, enabling cost-effective pre-drilling evaluations of critical geomechanical properties, including stress and rock strength.</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

Parameterization of Geomechanical Properties Through ML Algorithms for Accurate Determination and Prediction of Horizontal Stress: A Case of Niger Delta Basin and Implications on Its Application

  • 期刊论文

  • 发表日期：

    2025年11月28日

查看全文
  • Oluwaseun Daniel Akinyemi,
  • John Oluwadamilola Olutoki,
  • Mohamed Elsaadany,
  • Numair Ahmed Siddiqui,
  • Sami ElKurdy,
  • Muthuvairavasamy Ramkumar

摘要

During well planning, accurately determining geomechanical parameters is crucial to ensure wellbore stability during drilling, especially in complex reservoirs. A thorough assessment of these factors optimizes the drilling and well completion processes. Among the primary stresses, minimum horizontal stress plays a key role in hydraulic fracturing design and wellbore stability. However, directly measuring this stress is labor-intensive and costly, highlighting the need for efficient, alternative methods. This study focuses on the Niger Delta Basin, where unconsolidated reservoirs and overpressured shales are common, by presenting a well-log scale 1D geomechanical model and a data-driven approach to predict minimum horizontal stress in seven wells within the Eocene Agbada Formation. Using industry-standard equations, relevant geomechanical parameters were calculated, and six machine learning models were applied to conventional log data to predict minimum horizontal stress. Results showed that the sandstone sediments’ pore pressure gradient values ranged from 0.56 to 0.60 psi/ft, reaching 0.69 to 0.71 psi/ft in overpressured shales. The 1D model revealed a narrower mud window in overpressured zones. Among the models, gradient boosting achieved the highest accuracy, with an R2 of 0.92 and the lowest MAE of 273.53 on test data. Blind testing on additional wells validated the model’s robustness and low error rate. These machine-learning results can significantly reduce the time, manpower, and resources typically required for direct measurements, enabling cost-effective pre-drilling evaluations of critical geomechanical properties, including stress and rock strength.