计及碳排放配限额的虚拟电厂多目标风险规避优化模型

Multi-objective Risk Avoidance Optimization Model for Virtual Power Plants Considering the Carbon Emission Allowance

将风电场、光伏发电、生物质发电、储能和燃气轮机及柔性负荷聚合为虚拟电厂(Virtual power plant,VPP).进一步,为刻画风光不确定性风险,分别利用条件风险价值方法(Conditional risk at value,CVaR)构造最小化运营风险目标函数及利用鲁棒随机优化理论转化含不确定性变量约束条件,并选取最大化运营收益和最小化碳排放总量,构建VPP多目标风险规避优化模型.最后,选取改进IEEE30节点系统进行算例分析,结果表明:1)所提风险规避模型能够兼顾效益、风险和碳排放多方诉求;特别是,当鲁棒系数Γ≤0.85,较小的不确定性会带来较大的风险,表明决策者风险态度会影响VPP调度方案;2)预测误差e较高时,相同的Γ增长幅度会带来更高的CVaR增长幅度,表明较低的预测精度会放大不确定性风险,意味着决策者需通过提升预测精度以降低VPP运营风险;3)META能凸显清洁能源环境友好特性,实现VPP整体的最优均衡运行.综上,所提模型能够为决策制定最优VPP调度策略提供决策支撑.

The paper integrated wind power plant(WPP),photovoltaic power generation(PV),biomass power generation(BPG),energy storage system(ESS),conventional gas turbines(CGT)and flexible load into a virtual power plant(VPP).Then,in order to analyze uncertainties,the conditional risk at value(CVaR)is introduced to construct the objective of the minimum operation risk and the robust optimization theory is used to convert the constraint conditions into stochastic constraint conditions.Thirdly,the objective functions of the maximum operation revenue and the minimum carbon emission are chosen to construct the multi-objective risk aversion model for VPP operation.Finally,the improved IEEE 30 bus system is chosen for example analysis,the results show:1)The proposed risk aversion model can balance the benefits,risks and carbon emissions;in particular,when the robustness coefficientΓ<0.85,the smaller uncertainty will bring greater risks,indicating that the risk attitude of decision makers will influence VPP’s scheduling scheme.2)When the prediction error e is high,the same?growth rate ofΓwill bring a higher growth rate of CVaR,indicating that lower prediction accuracy will amplify the uncertainty risk,meaning that the decision maker needs to improve the predictive accuracy to reduce VPP operational risk;3)META can highlight the environmentally friendly characteristics of clean energy and achieve optimal optimal operation of VPP.Therefore,the proposed multi-objective and solution algorithm model can be used to provide reliable decision support for decision makers to develop optimal VPP operation plans.