计及风电不确定性的区域互联动态经济优化调度方法

结合含风电场的区域互联电力系统运行特点,考虑风电出力、负荷的不确定性因素以及电力系统安全运行约束,建立计及系统燃料费用、机组运行维护成本、风险成本的区域互联动态经济调度优化模型,并以一种融合拉丁超立方采样、场景缩减法和自学习差分算法的优化方法对所提模型进行求解。该方法根据风电和负荷预测误差采用拉丁超立方采样技术生成大量样本,并对所生成样本结合场景缩减法进行缩减,再由自学习差分算法进行全局寻优,得到各场景所对应的最优调度方案。结果表明:采用该方法既能模拟出风电、负荷的不确定性特点,又能避免建立过于复杂的随机性模型,降低了建模和求解的难度,同时所提自学习差分算法具有良好的收敛特性及鲁棒性。因此,所提优化方法对区域互联电力系统优化调度具有参考价值。

Uncertainty of wind electric power-considered dynamic-economic optimal dispatching method for regional interconnected power system

Combined with the operation characteristics of the regional interconnected power system that includes wind farms，a dynamic-economic optimal dispatching model for the regional interconnection with the consideration of the fuel cost，power-generating unit operation and maintenance costs and risk cost of the system is established by taking the factors of the uncertainties of the output and load of wind farm as well as the restriction on the safety operation of power system into account. Moreover，the model proposed herein is solved with an optimized method that integrates the Latin hypercube sampling，scenario-reduction method and self-learning differential algorithm. By this method，a large amount of samples are created with the Latin hypercube sampling technique at first in accordance with the wind electric power and load prediction errors，and then the created samples are reduced with the scenario-reduction method. Secondarily，global optimization is carried out with the self-learning differential algorithm，and then the optimal dispatching schemes corresponding to all the scenarios concerned are obtained. The result shows that the adoption of this method can not only simulate the characteristics of the uncertainties of wind electric power and load，but can also avoid establishing those more complicated random models，thus the modelling and solving difficulties are to be lowered as well. Meanwhile，the self-learning differential algorithm proposed herein has better convergence characteristics and robustness. Therefore，the optimal method proposed herein has a referential value for the optimal dispatching of the regional interconnected power system.