基于画面组态的聚类极值实时计算方法

Real time calculation method of clustering extremum based on graph configuration

分散控制系统(DCS)经常需要监视一些重要测点的极值,传统方法是导入专门的外部计算系统,或者通过组态工具搭建复杂的功能块来计算测点极值,这会影响DCS的稳定性、实时性及可靠性。为提高用户监视效率,降低系统的安全风险,提出了一种基于画面组态的聚类极值实时计算方法。该方法首先在历史站中通过基于距离的k-means算法将测点数据划分为k个初始聚类簇,然后利用基于密度的局部离群因子(LOF)聚类分析计算出k个初始聚类簇的极值点,最后在工程师站构建实时内存索引库对测点信息进行更新,并通过监视画面将测定极值信息展示给用户。该方法已在多个火电厂DCS中应用,实际工程表明该方法能避免外部极值计算系统与DCS通信的安全问题,降低用户配置逻辑组态的复杂性,保障测点极值计算的准确性、实时性和稳定性。

Distributed control system(DCS) often needs to monitor the extreme value of some important measuring points. The traditional approach is to import specialized external computing system, or build complex function blocks through configuration tools to calculate the extreme values of measuring points, which may affect the stability, real-time and reliability of DCS. In order to improve the efficiency of monitoring and reduce the security risk of the system, a real-time calculation method of clustering extreme values based on graphic configuration is proposed. In this method, the measuring points’ data is firstly divided into k initial clusters by the distance-based kmeans algorithm in the historical station, then the density-based local outlier factors(LOF) cluster analysis is used to calculate the extreme points of the k initial clusters. Finally, a real-time memory index library is built on the engineer station to update the information of points, and display the extreme value information in front of the user through the monitoring graphic. At present, this method has been applied in DCS of many thermal power plants. It has been proved by actual engineering that this method can avoid the safety problem of communication between the external extreme value calculation system and the DCS, reduce the complexity of user configuration logic configuration, and ensure the accuracy, real-time and stability of extreme value calculation for measurement points.