摘要
配电网的"低电压"问题严重影响社会经济的发展和人民的生活,因此建立高效的电网低电压成因诊断模型,可以优化低电压投资方案、明确低电压投资方向并为低电压的治理提供决策支持。文中研究了基于大数据挖掘的电网低电压成因诊断方法。从电网低电压现象成因的实际情况出发,提出了一个较为完备的低电压诊断模型,模型执行流程主要包括聚类和分类两部分。首先,基于密度策略选择初始聚类中心,并利用DBI指标选择最优确定聚类个数,对K-means聚类算法进行改进。之后利用SVM对成因进行分类,并利用粒子群算法对核宽度参数和函数拟合误差进行优化筛选。算例仿真分析证明,该方法具有较高的可用性和良好的准确率,可以满足电力企业对于低电压诊断的需求。
The low voltage of the distribution network problems affect the development of social economy and people's life seriously,aiming at the above problem,this paper established the low voltage caused diagnosis model which could optimize the investment scheme,put the investment orientation and provide decision support for the management of low voltage. The low voltage causes diagnostic method based on big data mining was studied on in the paper,and a comparatively complete low voltage of diagnosis model was proposed from the actual situation of low power grid voltage phenomenon causes,which mainly included two parts of the clustering and classification. Firstly,the K means clustering algorithm was improved based on the density of initial clustering center strategy choice and using the DBI index for choosing the optimal clustering number.Then,it used SVM to classify,and the particle swarm optimization algorithm to optimize the parameters of wide and function fitting error. The example simulation indicated that the method had high availability and good accuracy,which could meet the needs of diagnosis for low voltage electric power enterprises.
出处
《信息技术》
2017年第4期174-177,共4页
Information Technology
