摘要
文中针对配变重过载现象,研究了基于改进Logistic回归模型的短期配变重过载预警方法。分析了影响配网发生重过载的各种因素,并对因素样本进行了预处理,给出了合理样本量的计算方法。之后,构建短期预警模型,对其进行了多元共线性改进。最后给出了预警分析的处理过程,并进行了模型适用性评价,算法实用有效,可为电力运检部门提前调整检修、巡检计划提供参考。
Aiming at the weight overload phenomenon, this paper researches on the short-term early warning improved algorithm based on the Logistic regression model. The various factors influencing the distribution network occurs weight overload were analyzed,and the factors sample were pretreated before giving the calculation method of reasonable sample size. Then,the short-term warning models were built and carried on the multivariate linear improvement. Finally, the warning analysis process and model applicability evaluation were proposed. The algorithm protested it was practical and effective,and could provide the reference for adjusting the maintenance and inspection plan of the power maintaining sectors.
出处
《信息技术》
2017年第2期130-132,136,共4页
Information Technology
关键词
重过载
短期预警
LOGISTIC回归模型
模型检验
variable weight overload
short-term early warning
Logistic regression model
model testing
