Comparative Reaction Modelling and k-Nearest Neighbors Analysis of Cocos nucifera Shell Thermal Degradation

Area of Science:

• Biomass valorization
• Thermochemical conversion
• Materials science

Background:

• Coconut shell is a lignocellulosic waste with high energy potential.
• Efficient valorization requires understanding its thermal decomposition behavior.
• Existing models may not fully capture the complex degradation pathways.

Purpose of the Study:

• To establish a framework for coconut shell valorization using thermogravimetry and machine learning.
• To analyze the physicochemical properties and thermal degradation characteristics of coconut shell.
• To develop predictive models for optimizing the thermochemical conversion process.

Main Methods:

• High-resolution thermogravimetry (TG/DTG) for thermal analysis.
• SEM/XEDS and FTIR for physicochemical characterization.
• Kinetic modeling (F3 mechanism) and thermodynamic analysis.
• k-Nearest Neighbors (kNN) machine learning for predictive modeling.

Main Results:

• Coconut shell exhibits high carbon (45.7%) and volatile content (71.5%).
• Distinct degradation windows for hemicellulose, cellulose, and lignin were identified.
• Pyrolysis follows a third-order (F3) mechanism with moderate energy requirements (Ea=165 kJ·mol⁻¹).
• kNN models achieved near-perfect prediction accuracy (R²≈1.000) for TG, DTG, and conversion.

Conclusions:

• Coconut shell is a viable high-energy feedstock for valorization.
• Integrated thermogravimetric and machine learning approaches enable precise prediction of thermal degradation.
• Optimized prediction minimizes experimental waste and enhances process efficiency.