计及PV-EV时空相关性的自适应线性化概率潮流计算及风险评估

Probabilistic Power Flow Calculation and Risk Assessment with Adaptive Piecewise Linearization Incorporating PV-EV Spatiotemporal Correlation

中文关键词:光伏  电动汽车  相关性  自适应  分段线性化  概率潮流  风险评估

keywords:photovoltaic  electric vehicle  correlation  adaptive  piecewise linearization  probabilistic power flow  risk assessment

中文摘要:光伏（photovoltaic，PV）与电动汽车（electric vehicle，EV）的大量接入对配电网可能带来电压和潮流越限的风险。对此，首先采用混合模型建立PV的边缘分布，用路网-配电网耦合建立EV的边缘分布，并通过Vine-Copula函数建立目前研究中考虑不足的PV-EV时空相关性，从而获取更符合实际的联合样本数据用于概率潮流分析。对于概率潮流计算，提出了一种自适应分段线性化方法处理非线性的潮流方程，改善了传统等距分段进行线性化精度差的问题，得到了更精确的节点电压与支路潮流概率分布情况。最后通过定义电压越限、潮流越限严重度函数来量化PV-EV接入对配电网造成的风险。IEEE 118系统测试表明，自适应分段线性化方法相比于传统的等距分段线性化误差更小，可以更加精确地反映实际带有相关性的PV-EV接入配电网所造成的风险，为新能源接入节点的选取提供实际指导意义。

Abstract:The large-scale integration of photovoltaic （PV） systems and electric vehicle （EV） into distribution networks poses significant risks of voltage violations and power flow overloading to distribution networks. To address this issue，a mixed model is first adopted to establish the marginal distribution of PV generation，while a road network-distribution network coupling method is used to establish the marginal distribution of EV charging demand. A Vine-Copula function is then utilized to capture the PV-EV spatiotemporal correlation （which has been inadequately considered in existing studies），thereby obtaining more realistic joint sample data for probabilistic power flow analysis. For probabilistic power flow calculation，an adaptive piecewise linearization method is developed to handle nonlinear power flow equations，overcoming the low accuracy limitation of traditional uniform segmentation linearization and achieving more precise probability distributions of node voltages and branch power flows. Finally，severity functions for voltage violations and power flow overloading are defined to quantify the risks imposed by PV-EV integration on distribution networks. Case studies on the IEEE 118-bus test system demonstrate that the proposed adaptive piecewise linearization method yields smaller errors compared to the traditional uniform segmentation approach. It can more accurately reflect the actual risks of distribution networks with correlated PV-EV integration，providing practical guidance for the selection of renewable energy integration nodes.