Integration of Multiple Optimization Algorithms with Machine Learning: Predicting Nitrogen Adsorption Volume of Coal...

Area of Science:

• Interface Science: Focuses on solid-gas interactions and porous material characterization.
• Materials Science: Investigates the physical and chemical properties of coal rocks.
• Computational Science: Utilizes machine learning for predictive modeling.

Background:

• Nitrogen adsorption capacity is vital for characterizing porous materials, carbon sequestration, and coalbed methane development.
• Accurate prediction of adsorption volume is challenging due to complex, nonlinear relationships between coal properties and adsorption behavior.
• Understanding solid-gas interface interactions is key in interface science.

Purpose of the Study:

• To systematically analyze pore structure characteristics of coal rocks with varying particle sizes using low-temperature liquid nitrogen adsorption data.
• To develop and compare machine learning models (Random Forest and Support Vector Machine) optimized with algorithms for predicting nitrogen adsorption volume.
• To identify key factors influencing nitrogen adsorption in coal rocks, particularly particle size.

Main Methods:

• Low-temperature liquid nitrogen adsorption experiments to analyze pore structure.
• Application of Random Forest (RF) and Support Vector Machine (SVM) models.
• Integration of four optimization algorithms: Sparrow Search Algorithm (SSA), Snake Optimization (SO), Chameleon Swarm Algorithm (CSA), and Northern Goshawk Optimization (NGO).

Main Results:

• Optimized models, especially RF enhanced by NGO and CSA, significantly outperformed other approaches.
• The RF model demonstrated superior prediction accuracy over SVM, achieving an R² of 0.9804 with NGO/CSA optimization.
• Coal rock particle size (120-200 mesh) was identified as a critical predictive factor, improving validation set performance by nearly 10%.

Conclusions:

• Integrating optimization algorithms with machine learning models provides an effective quantitative method for predicting nitrogen adsorption capacity in coal rocks.
• Coal particle size is a significant factor influencing nitrogen adsorption volume.
• The study validates the effectiveness of advanced computational approaches for understanding solid-gas interface interactions.