The thermal conductivity of coir fiber-reinforced PVC composites is influenced by fiber content, particulate size, and chemical treatment. Optimization models, such as Particle Swarm Optimization (PSO), Dragonfly Optimization (DFO), and Cuckoo Search Algorithm (CSA), are used to predict thermal conductivity. The models are validated through experiments and statistical metrics. The results show that CSA achieves the highest thermal conductivity of 0.801 W/mK. The convergence behavior of the algorithms is analyzed, with PSO converging rapidly, DFO balancing exploration and exploitation, and CSA exhibiting a gradual decrease in error. The error reduction plots show that PSO has a faster reduction in error with greater stability, while DFO and CSA require more iterations to reach comparable reductions in error. The scatter plots illustrate the relationship between experimental and predicted thermal conductivity values, with CSA showing the closest predictions to experimental values. The comparative analysis highlights the distinct advantages and trade-offs among the optimization algorithms, with CSA providing the best overall accuracy and PSO exhibiting rapid convergence and stability. A radar chart compares the performance of the algorithms based on mean squared error, mean absolute error, and coefficient of determination, with CSA being the best-performing model.
Utilizing advanced computational methods to enhance thermal conductivity in polyvinyl chloride composites reinforced with coir fibers.
