Performance analysis of a hybrid energy system under weather variability using probability-based multi-objective optimization

초록

This study presents a probability-based multi-objective optimization framework for a hybrid distributed energy system that integrates photovoltaic generation, wind turbine generation, phosphoric acid fuel cells, gas turbine combined cycle units, and battery energy storage. The framework simultaneously minimizes the daily operating cost per unit of electricity generated and the electricity purchase rate while accounting for renewable resource variability through seasonal and cloud-cover-based probability scenarios. Unlike previous deterministic approaches, this study combines combined probabilistic modeling with Shapley additive explanation analysis to quantify the influence of the major design variables and enhance the interpretability of optimization results. Sensitivity analysis revealed that the photovoltaic and wind power capacities were the most influential variables for both objectives, highlighting a clear cost-reliability trade-off between investment expansion and grid dependence. Comparative analysis of the annual, seasonal, and probabilistic seasonal scenarios showed that as the weather variability increased, the operating cost and the electricity purchase rate rose, indicating that higher resource uncertainty leads to overall performance degradation. The probabilistic seasonal configuration, validated through this sequential case structure-(i) annual, (ii) seasonal, and (iii) probabilistic seasonal based on the cloud-cover probability-provided a realistic but computationally feasible framework for system design. The integration of probabilistic modeling with Shapley-based analysis improves the interpretability of the optimization outcomes and provides a practical basis for determining the renewable capacities under variable weather conditions.

키워드