Scholarlink

Abstract <p>Nanoemulsions are increasingly recognised for their superior interfacial properties and stability, making them highly valuable in various industrial sectors, including oilfield applications. This study presents a novel machine learning approach for predicting the interfacial tension (IFT) of crude oil-in-water nanoemulsions in contact with diesel, incorporating key formulation parameters such as pH, surfactant dosage and crude oil content. Unlike prior research that primarily focuses on IFT measurements between pure crude oil and aqueous phases, this work investigates more realistic multiphase systems involving pre-formulated nanoemulsions. A series of nanoemulsions were prepared under systematically varied conditions and their dynamic IFT behaviour was experimentally measured. To model this behaviour, three artificial neural network (ANN) models based on multilayer perceptron architecture were developed, each trained using different input combinations: (i) sodium dodecylbenzenesulfonate surfactant concentration and time, (ii) crude oil volume percentage and time, and (iii) pH and time. The models were trained and validated using the Levenberg–Marquardt algorithm and evaluated using coefficient of determination (<i>R</i>), mean squared error and margin of deviation metrics. All models demonstrated excellent predictive performance, with <i>R</i> values exceeding 0.9 and minimal deviation from experimental data. The results establish ANN as a powerful tool for accurately predicting interfacial properties of complex emulsion systems. This data-driven machine learning framework has the potential to reduce the need for extensive experimentation and to provide a robust platform for rapid formulation screening and optimisation in oilfield applications and beyond.</p> Graphical abstract <p>Crude oil–in–water nanoemulsions were prepared across diff erent pH, surfactant, and oil fraction. Pendant-drop tensiometry recorded dynamic interfacial tension (IFT) of the nanoemulsions against diesel. Multilayer-perceptronmodels using time and formulation variables accurately predicted IFT (R0.9), enabling rapid screening and guidelinesfor optimization while reducing experimental workload.</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

Effects of pH, crude oil content and surfactant dosage on interfacial tension of crude oil nanoemulsions against diesel: A combined experimental and machine learning investigation

  • 期刊论文

  • 发表日期：

    2026年3月12日

查看全文
  • Andaç Batur Çolak,
  • Sagheer A Onaizi

摘要

Abstract

Nanoemulsions are increasingly recognised for their superior interfacial properties and stability, making them highly valuable in various industrial sectors, including oilfield applications. This study presents a novel machine learning approach for predicting the interfacial tension (IFT) of crude oil-in-water nanoemulsions in contact with diesel, incorporating key formulation parameters such as pH, surfactant dosage and crude oil content. Unlike prior research that primarily focuses on IFT measurements between pure crude oil and aqueous phases, this work investigates more realistic multiphase systems involving pre-formulated nanoemulsions. A series of nanoemulsions were prepared under systematically varied conditions and their dynamic IFT behaviour was experimentally measured. To model this behaviour, three artificial neural network (ANN) models based on multilayer perceptron architecture were developed, each trained using different input combinations: (i) sodium dodecylbenzenesulfonate surfactant concentration and time, (ii) crude oil volume percentage and time, and (iii) pH and time. The models were trained and validated using the Levenberg–Marquardt algorithm and evaluated using coefficient of determination (R), mean squared error and margin of deviation metrics. All models demonstrated excellent predictive performance, with R values exceeding 0.9 and minimal deviation from experimental data. The results establish ANN as a powerful tool for accurately predicting interfacial properties of complex emulsion systems. This data-driven machine learning framework has the potential to reduce the need for extensive experimentation and to provide a robust platform for rapid formulation screening and optimisation in oilfield applications and beyond.

Graphical abstract

Crude oil–in–water nanoemulsions were prepared across diff erent pH, surfactant, and oil fraction. Pendant-drop tensiometry recorded dynamic interfacial tension (IFT) of the nanoemulsions against diesel. Multilayer-perceptronmodels using time and formulation variables accurately predicted IFT (R0.9), enabling rapid screening and guidelinesfor optimization while reducing experimental workload.