摘要
为了解决由多种能源调峰无限制导致的配电网运行成本高的问题,提出了基于DBN-RF的配电网多能互补优化调度算法。利用深度置信网络设计多能互补耦合协调结构,并为鲁棒优化DBN模型提供上、下界,输出最优调度方案。使用RF随机森林算法进行总体协调度评价,以样本集信息熵计算结果为依据,统计子决策树的输出结果,将多目标决策问题转化为单目标优化问题,确定最终评价结果。实验结果表明,该算法总成本最低,与基础成本相比少112.8万元,且输出的有功功率基本与理想出力一致,说明利用该算法可以实现配电网经济性与鲁棒性的平衡。
In order to solve the problem of high operation cost of distribution network caused by unlimited multi-energy peak modulation,a multi-energy complementary optimal scheduling algorithm based on DBN-RF was proposed.The multi-energy complementary coupling coordination structure is designed by deep confidence network,and the upper and lower bounds are provided for the robust optimization DBN model,and the optimal scheduling scheme is output.The RF random forest algorithm was used to evaluate the overall coordination degree.Based on the calculation results of the information entropy of the sample set,the output results of the sub-decision tree were counted and the multi-objective decision problem was transformed into a single-objective optimization problem to determine the final evaluation result.Experimental results show that the algorithm has the lowest total cost,which is 1.128 million yuan less than the basic cost,and the output active power is basically consistent with the ideal output,indicating that the algorithm can achieve the balance of distribution network economy and robustness.
作者
汤俊峰
TANG Junfeng(State Grid Zhejiang Electric Power Company,Longquan Power Supply Company,Longquan 323700,China)
出处
《电子设计工程》
2023年第21期128-131,136,共5页
Electronic Design Engineering
关键词
深度置信网络
随机森林算法
配电网多能互补
优化调度
deep confidence network
random forest algorithm
distribution network complement each other
optimal operation
