Decision Making and Energy Storage System Management

Abstract

This study evaluates the effectiveness of various machine learning strategies in managing energy in Fuel Cell Electric Vehicles (FCEVs), focusing on fuel cell and battery inverter behaviour. The analysis compares four methods Gaussian Naive Bayes (NB), Random Forest, k-NN, and AdaBoost using key metrics: Recall, f1-score, and precision. Gaussian NB and Random Forest achieve identical performance for the fuel cell (Recall: 0.87, f1-score: 0.82, precision: 0.89) and battery (Recall: 0.66, f1-score: 0.57, precision: 0.5). In contrast, k-NN achieves a precision of 0.74, while AdaBoost excels with a fuel cell precision of 0.98 and a battery precision of 0.94. AdaBoost also outperforms the other methods in f1-score (0.98 for the fuel cell, 0.90 for the battery) and recall (0.95 for the fuel cell, 0.84 for the battery), highlighting its superior performance in inverter behaviour control.