Scenario Generation and Planning Method for Park Integrated Energy Systems Considering Multiple Uncertainties

Park Integrated Energy Systems (PIES) planning faces significant uncertainties arising from internal renewable energy generation, load fluctuations, and external energy procurement prices. These uncertainties can lead to variations in equipment capacity configurations, which in turn cause fluctuations in total planning costs. Existing approaches often assume known probability distributions for uncertainty factors; however, limitations in historical data and fitting errors make it challenging to accurately estimate these distributions. This paper presents a novel approach for modeling the risk of cost fluctuations in PIES planning without relying on specific probability distributions. We propose using a Dirichlet model to construct a fuzzy set based on a cumulative probability distribution function derived from limited historical data, effectively capturing uncertainty in probability distributions. Furthermore, the Distributionally Robust Conditional Value-atRisk (DRCVaR) method is employed to quantify the risk of planning cost fluctuations under worst-case probability distributions within the fuzzy set. The proposed model mitigates the conservatism typically seen in traditional methods and effectively reduces the risk of cost fluctuations. Comparative results demonstrate that the DRCVaR-based approach outperforms CVaR-based models in minimizing risk-induced losses.