An Improved Cumulant-Based Probabilistic Energy Flow Computation Method for Integrated Electricity and Natural Gas Systems

摘要

Intermittent renewable energies and fluctuant energy-demands introduce uncertainties into planning and operation of the integrated electricity and natural gas system (IENGS). This paper proposes an improved cumulant-based probabilistic energy flow (PEF) computation method. First, a cumulant-based PEF computation model for IENGS is developed with consideration of frequency regulation characteristics of the electricity system, three couplings (gas-fired generators. P2G units and electric-driven compressors), and correlations among input random variables. Second, an improved K-means clustering technique is proposed to reduce linearization errors of the cumulant-based PEF caused by large fluctuations of input random variables. The improved K-means clustering technique incorporates the effects of different scales and sensitivities of different input random variables. Base on the input random variable samples in each cluster, the cumulant-based PEF is conducted with improved accuracy. Finally, the proposed method is examined using an integrated electricity and natural gas test system that is comprised of IEEE 39-bus system and Belgian high-calorific natural gas system. Compared with Monte Carlo simulation (MCS), multi-linear Monte Caro simulation (MLMCS), the cumulant method (CM) without clustering techniques, and the method of combined CM and the conventional K-means (CMCK), the proposed method achieves better performance with consideration of both accuracy and efficiency.