New and Highly Accurate Static Young's Modulus Model Using Machine Learning Techniques.

Area of Science:

• Petroleum Geosciences
• Machine Learning Applications
• Rock Mechanics

Background:

• Static Young's modulus (Es) is vital for petroleum calculations.
• Existing models for Es prediction lack precision and are limited in scope.
• Accurate Es determination is crucial for geological assessments and reservoir engineering.

Purpose of the Study:

• To develop and evaluate novel machine learning models for predicting static Young's modulus (Es).
• To compare the performance of an adaptive neuro-fuzzy inference system (ANFIS) against existing models.
• To assess the capability of ANFIS in identifying rock types based on predicted Es.

Main Methods:

• Utilized adaptive neuro-fuzzy inference system (ANFIS) and other machine learning techniques.
• Employed bulk formation density (RHOB), shear wave velocity (DTs), and compressional wave velocity (DTc) as predictor variables.
• Trained and validated models on a diverse dataset of 1853 hydrocarbon reservoir rock samples.

Main Results:

• The ANFIS model achieved an average absolute percentage relative error (AAPRE) of 5.1% and a correlation coefficient (R) of 0.9602.
• ANFIS demonstrated superior predictive accuracy and faster decision-making compared to other machine learning methods.
• Trend analysis confirmed that increasing RHOB increases Es, while increasing DTc and DTs decreases Es.

Conclusions:

• ANFIS is the optimal model for predicting static Young's modulus (Es) with high accuracy.
• The ANFIS model accurately captures the physical relationships between rock properties and Es.
• This research offers significant advancements for geological assessments and petroleum engineering applications.