基于分布式边缘计算的智能配电网规划策略研究

Research on Smart Distribution Network Planning Strategy Based on Distributed Edge Computing

配电网规划、调度能有效提升能源的经济效益,但传统ADMM算法在求解规划问题时易陷入局部最优,造成目标收益差、系统成本高的不良后果。为解决上述问题,基于边缘节点并行更迭决策量的优化,提出一种EN-ADMM改进交替方向乘子法最优求解算法。首先需保证调度的稳定性与安全性,将柔性负荷、分布式源与电网总和作为配电网的稳定约束条件;然后构建配电网最低运行成本目标函数,并对目标函数进行优化,构建出32节点分布式电网模型;最后分别采用传统ADMM算法与EN-ADMM算法对配电网模型进行性能参数与经济参数的最优求解。实验结果显示出,较传统ADMM算法相比,EN-ADMM算法的计算耗时降低了41.9%,且目标成本降低了4.21%,即EN-ADMM算法通过边缘节点优化提升了收敛时效性与聚拢性,同时提高了配电网目标经济效益。综上,EN-ADMM算法在智能配电网规划中具有较为重要的仿真价值。

Distribution network planning and dispatching can effectively improve the economic benefits of energy,but the traditional ADMM algorithm is easy to fall into local optimum when solving the planning problem,resulting in poor target income and high system cost.In order to solve the above problems,an improved alternating direction method of multipliers(EN-ADMM)algorithm is proposed based on the optimization of parallel alternation decision quantity of edge nodes.In order to ensure the stability and security of dispatching,the sum of flexible load,distributed source and power grid is taken as the stability constraint of distribution network,and then the objective function of the lowest operating cost of distribution network is constructed,and the objective function is optimized to construct a 32-node distributed power grid model.Finally,the traditional ADMM algorithm and EN-ADMM algorithm are used to optimize the performance parameters and economic parameters of the distribution network model.The experimental results show that compared with the traditional ADMM algorithm,the computational time of EN-ADMM algorithm is reduced by 41.9%,and the target cost is reduced by 4.21%,that is,the EN-ADMM algorithm improves the convergence timeliness and convergence through edge node optimization,and improves the target economic benefits of distribution network.To sum up,EN-ADMM algorithm has important simulation value in smart distribution network planning.