基于FBD法的配电网设计指标智能检测方法仿真  

Simulation of Intelligent Detection Method for Distribution Network Design Index Based on FBD Method

作　　者：胡诗尧 荆志朋 柴林杰[1] CHAI Lin-jie;JING Zhi-peng;HU Shi-yao(State Grid Hebei Economic Research Institute,Shijiazhuang Hebei 050021,China;Tianjin University,Tianjin 300072,China)

机构地区：[1]国网河北省电力有限公司经济技术研究院,河北石家庄050021 [2]天津大学,天津300072

出　　处：《计算机仿真》2024年第11期167-172,共6页Computer Simulation

基　　金：国家电网公司科技项目(5024JY20000B)。

摘　　要：传统配电网设计指标检测系统与分布式能源的高速发展相驳,安全、稳定的指标体系无法完全反映出新形势下电网的经济性与可持续性优势,为提高配电网设计指标的完整性与检测稳定性,基于自适应卷积网络与FBD等效配电网构建出FBD-ACNN配电网指标智能检测模型。模型首先对采用FBD算法将设计好的配电网系统进行等效,提高配电网指标的获取能力;然后从安全、可靠、环保与运行四个方向,分别构建配电网设计一、二级指标待检测体系接着对获取的配电网四类指标进行去量纲化处理,提升指标检测模型的建模效率;最后将Adam算法与退火算法有机融合,提高卷积网络的指标权重与网络学习率的自适应调整性,构建出配电网指标智能检测模型。基线算法仿真分析结果显示,在PGI指标检测数据上,较其它三类基线检测算法相比,FBD-ACNN算法的单项指标与综合指标均为最优值,这表明提出的算法在实际数据中有良好的应用效果,且最优值说明该算法的稳定性与泛化性均较好。综上,提出的FBD-ACNN算法能有效的将新能源体系纳入配电网设计指标检测中,且提升了检测的稳定性与泛化性,在智能计算机领域具有重要的研究价值。The traditional distribution network design index detection system contradicts the rapid development of distributed energy,and the safe and stable index system can not fully reflect the economic and sustainable advantages of the power grid under the new situation.In this paper,based on adaptive convolutional network and FBD equivalent distribution network,an FBD-ACNN intelligent detection model of distribution network indicators is constructed.Ac-cording to the model,firstly,the designed distribution network system is equivalent by adopting an FBD algorithm,so that the acquisition capability of the distribution network index is improved;secondly,from the four directions of safe-ty,reliability,environmental protection and operation,a distribution network design primary index system and a distribution network design secondary index system to be detected are respectively constructed;thirdly,the four types of obtained distribution network indexes are subjected to de-dimensionalization processing,so that the modeling efficiency of an index detection model is improved;and fourthly,the model is subjected to dedimensionalization processing;finally,Adam algorithm and annealing algorithm are integrated to improve the adaptive adjustment of index weight and network learning rate of convolutional network,and the intelligent detection model of distribution network index is con-structed.The simulation results of the baseline algorithm show that compared with the other three types of baseline detection algorithms,the single index and the comprehensive index of the FBD-ACNN algorithm are the optimal values on the PGI index detection data,which shows that the algorithm proposed in this paper has a good application effect in the actual data,and the optimal value shows that the algorithm has good stability and generalization.To sum up,the FBD-ACNN algorithm proposed in this paper can effectively incorporate the new energy system into the detection of distribution network design indicators,and improve the stability and generalizatio

关 键 词：配电网设计 指标检测体系 自适应卷积网络 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]