Optimal Stochastic Scheduling of a Hybrid Photovoltaic/Hydrogen Storage-Based Fuel Cell System in Distribution Networks Using Dynamic Point Sampling and Adaptive Weighting

Abstract

This paper proposes a novel stochastic scheduling and energy management framework for a hybrid photovoltaic/fuel cell energy system integrated with hydrogen storage (HBES) in a radial distribution network. The goal is to minimize the total annual energy losses cost (TAELC), voltage deviation (VD), energy not-supplied (ENS), annual emission cost (AEC), HBES cost (HBESC). The study recommends an Improved Weighted Average Algorithm (IWAA) based on an adaptive exploration-exploitation trade-off (AEETO) to prevent premature convergence, which enables optimal scheduling and placement of the HBES. A dynamic point sampling with Adaptive Weighting (DPSAW) methodology is also recommended to model photovoltaic power and network loading uncertainties in the stochastic scheduling framework. The strategy is tested on a 33-bus radial distribution network, and scenarios Case I (PV alone), Case II (PV + FC), and Case III (energy management and stochastic scheduling of PV + FC) are studied. Simulation results are presented to find that Case III outperforms Case II to improve network performance, reduce TAELC, VD, ENS, AEC, and HBESC by 18.26 %, 21.91 %, 15.10 %, and 0.75 %, respectively. The IWAA also demonstrates better performance and convergence rates compared to conventional WAA and other optimizers. Furthermore, the stochastic model incorporating uncertainties results in increases in TAELC, VD, ENS, AEC, and HBESC by 9.01 %, 8.70 %, 7.73 %, and 0.43 %, respectively, when compared to the deterministic model in Case I. The effectiveness of the DPSAW-based stochastic approach is further confirmed, emphasizing the significance of accounting for uncertainties in real-world energy management systems. This approach offers a more realistic reflection of operational challenges, facilitating improved decision-making for network operators.