基于K-Means算法改进的NTP及其在用电信息采集系统中的应用

An Improved NTP Based on K-Means Algorithm and its Application in Electricity Information Acquisition System

基于K-Means算法改进网络对时协议(network time protocol,NTP)方案,并将其应用于用电信息采集系统中的主站和集中器对时。依据NTP对时模型,推导出对时误差阈值;基于对时误差阈值将多组时钟偏差值中的畸变数据滤除,保留较优的样本数据;针对保留的样本数据,采用一种改进的K-Means算法对其进行聚类迭代,获取实际时钟偏差值;将时钟偏差值补偿于集中器的时间,即完成主站和集中器的时钟同步。采用本文方案对某供电公司提供的主站和集中器对时数据进行实验,验证了本文方案的准确性和实用性。

A scheme of network time protocol(NTP)was improved based on K-Means algorithm,and was applied to synchronize main station with concentrator in the power information collection system.According to the NTP timing model,the clock deviation threshold was deduced.The distortion data of clock deviation values were filtered based on the clock deviation threshold,and better sample data were kept.An improved K-Means algorithm was used to cluster and iterate on the retained sample data,and the actual clock deviation value was obtained.The actual clock deviation value was compensated to the concentrator time,and the clock synchronization between main station and concentrator was completed.The accuracy and practicality of the proposed scheme are verified by experiments on the time data of the main station and the concentrator provided by a power supply company.