Achieving effective energy management in hybridized systems for fuel cell–battery vehicles using stochastic fractal search network

Existing energy management strategies for fuel cell–battery hybrid vehicles commonly ignoring the long-term effects of component degradation, focusing instead on consumption or cost reduction. This oversight results in expedited degradation of essential components, reducing system reliability over time. This paper introduces an innovative energy management system which integrates an improved stochastic fractal search algorithm with a deep neural network to optimize decision-making across changing load scenarios. This work introduces a frequency-decoupling architecture that divides low-frequency and high-frequency power components, allocating them to the fuel cell and battery, respectively, to reduce stress and increase the lifespan of these components. Suggested energy management system employs multi-objective cost function that explicitly integrates hydrogen consumption, system cost and degradation metrics, facilitating a more thorough optimization strategy than traditional approaches. High-gain non-isolated converter and a bidirectional converter are introduced to regulate voltage levels and reduce switching stress. The methodology is validated through simulations throughout many standardized driving cycles, including WLTP, NEDC, and CLTC. The results indicate a 12 % decrease in hydrogen use, a 15 % enhancement in powertrain efficiency and markedly reduced stress levels on the battery and fuel cell. These results validate that the proposed energy management system represents a significant improvement over current control systems by integrating degradation-aware optimization, sophisticated converter design and intelligent power estimate hence enhancing the durability and efficiency of hybrid vehicle operation.