基于在线支持张量机分类算法的中性点接地方式研究（英文）

Research on neutral grounding method based on online supporting tensor classification algorithm

以往关于中压配电网中性点接地方式决策的研究,只考虑了不同接地方式下可靠性等单方面因素,忽略了可靠性、安全性和经济性三者对接地决策的综合影响。针对此问题,提出在线支持张量机分类算法对可靠性、安全性和经济性进行三维综合分析。分别采用层次分析法、模糊综合评价法和模糊聚类法构建可靠性、安全性和经济性的评价模型,利用支持张量机有效核函数优化模型求解三维数据所对应的权张量,进而得到最优的中性点接地方式。数据分析表明:所提算法更能保留影响因素的结构信息,提高了数据时效性。

In the conventional works on neutral grounding modes for medium-voltage distribution system,the single factor such as reliability is only considered in various grounding modes,which neglects the comprehensive effect of reliability,security and economy on grounding modes. Therefore,online support tensor machine classification algorithm is proposed to analyze the reliability,security and economy in three-dimensional space. Specifically,the hierarchical analysis algorithm,fuzzy synthetic evaluation algorithm and fuzzy clustering method are proposed to establish the evaluation model of reliability,security and economy,respectively. The weight tensor is derived by solving the effective kernel function,then the optimal neutral grounding mode is obtained. Simulation results show that,compared with the above algorithms,the proposed algorithm could keep the structural information of data,and the data timeless is improved.