Predicting Thermal Behavior in E-Scooters with Neural Network Modeling for Battery Using Phase Change Materials (PCM) Predicting Thermal Behavior in E-Scooters with Neural Network Modeling for Battery Using Phase Change Materials (PCM)

The growing adoption of electric scooters necessitates advanced thermal management strategies to ensure battery safety, efficiency, and longevity. This study investigates the thermal behavior of lithium-ion batteries integrated with beeswax-based phase change material (PCM) under dynamic loading conditions, comparing it against paraffin and air-cooled systems. A custom-designed experimental rig, validated through mechanical analysis, facilitated the evaluation of temperature regulation and voltage stability. Results showed that beeswax PCM consistently reduced peak battery temperatures and preserved voltage integrity. To complement experimental insights, an Artificial Neural Network (ANN) model with a 4-4-1 multilayer perceptron architecture was developed to predict temperature ratios based on RPM, load, PCM type, and voltage ratio. The model exhibited excellent predictive performance with an R² of 0.9415 and low error metrics. This dual approach validates the feasibility of using bio-based PCM materials alongside AI-driven prediction models for enhanced thermal regulation in micromobility battery systems. The findings support the development of sustainable and intelligent battery management frameworks tailored for compact electric vehicles.