Scholarlink

<p>The study deals with the concept of combining the aspects of fractional calculus and artificial intelligence (AI) in predicting and optimizing the process of refining oil quality under different environmental and production parameters. The analysis is a synthesis of AI-driven algorithms and models that are defined by fractional differential equations and are relevant to study the dynamic processes, including hydrolysis and other mechanisms that affect the quality of oils, including palm oil, soybean oil, and olive oil. The effects of memory are captured by the fractional model hence making it more predictive than traditional models. The quantitative assessment of the offered framework demonstrated that the prediction error was decreased by 30 percent, and the mean absolute error (MAE) was 0.041, the score of the <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(R^2\)</EquationSource> </InlineEquation> was 0.94, which was better compared to the traditional regression and Random Forest models. These enhancements validate the strength and accuracy of the integration of the fractional-AI in the process of modeling nonlinear and memory-dependent oil degradation. The temperature, humidity and storage time are factors that are critical towards measuring the quality of oil. More precise predictive models are obtained and refining is optimized with advanced AI methods such as supervised learning algorithms, Physics-Informed Neural Networks (PINNs). The novelty of the proposed study is that the coupling of fractional differential equations with PINNs is direct, which creates a hybrid model that combines data-driven learning with the effect of physical memory. Unlike the former researches that have used the individual methods, this integration enables the model to portray the underlying process physics as well as the dynamic environmental effects, which is a significant advance in the area of predicting and optimizing oil quality. Even though it could be a solution, issues like data acquisition problems, expensive computing resources, or complex setups of models still exist and they have to go through some adjustments before getting accepted for more widespread factory use. This research identifies very promising possibilities to increase oil production, organize the storage space better, and control the quality through the introduction of new modeling techniques.</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

Analysis of integrated fractional calculus and artificial intelligence methods in optimizing oil refinement

  • 期刊论文

  • 发表日期：

    2026年5月22日

查看全文
  • Prabakaran Raghavendran,
  • Tharmalingam Gunasekar,
  • Kottakkaran Sooppy Nisar

摘要

The study deals with the concept of combining the aspects of fractional calculus and artificial intelligence (AI) in predicting and optimizing the process of refining oil quality under different environmental and production parameters. The analysis is a synthesis of AI-driven algorithms and models that are defined by fractional differential equations and are relevant to study the dynamic processes, including hydrolysis and other mechanisms that affect the quality of oils, including palm oil, soybean oil, and olive oil. The effects of memory are captured by the fractional model hence making it more predictive than traditional models. The quantitative assessment of the offered framework demonstrated that the prediction error was decreased by 30 percent, and the mean absolute error (MAE) was 0.041, the score of the \(R^2\) was 0.94, which was better compared to the traditional regression and Random Forest models. These enhancements validate the strength and accuracy of the integration of the fractional-AI in the process of modeling nonlinear and memory-dependent oil degradation. The temperature, humidity and storage time are factors that are critical towards measuring the quality of oil. More precise predictive models are obtained and refining is optimized with advanced AI methods such as supervised learning algorithms, Physics-Informed Neural Networks (PINNs). The novelty of the proposed study is that the coupling of fractional differential equations with PINNs is direct, which creates a hybrid model that combines data-driven learning with the effect of physical memory. Unlike the former researches that have used the individual methods, this integration enables the model to portray the underlying process physics as well as the dynamic environmental effects, which is a significant advance in the area of predicting and optimizing oil quality. Even though it could be a solution, issues like data acquisition problems, expensive computing resources, or complex setups of models still exist and they have to go through some adjustments before getting accepted for more widespread factory use. This research identifies very promising possibilities to increase oil production, organize the storage space better, and control the quality through the introduction of new modeling techniques.