【摘要】 针对穷举式数据挖掘算法对大样本数据库(如电站SIS系统)的优等隐含信息挖掘效率不高的现状,提出了一种嵌套式数据挖掘算法,该方法融合了划分聚类和双约束相关分析两种数据挖掘技术,首先采用k-means方法对机组运行数据按典型工况进行聚类划分,在各典型工况中抽取一定量的数据构成样本空间;然后采用基于维度约束和用户兴趣度约束的双约束相关分析对样本空间进行"启发式"数据挖掘,探寻机组最优运行方式。通过山西某电厂600 MW亚临界中间再热凝汽式机组实际数据进行验证,结果表明该算法不但可以较穷举挖掘算法显著降低了时间成本,而且挖掘结果能直观反映出机组的优化运行状态,对现场经济运行提供了有效指导。更多还原

【Abstract】 A nested data mining algorithm was proposed to more efficiently obtain superior implicit information from large sample database(such as SIS in power plant) than complete data mining algorithm. The method comprised clustering and double constraints related analysis of two data mining technology. Firstly k-means method was used to divide operating data into typical working conditions, from which a number of operating data were extracted to constitute the sample space. Then the double constraints related analysis based on dimensionality constraint and interest degree constraint was adopted to get optimal working condition by enlightening search. Based on the actual data in a 600 MW plant in Shanxi province, the result shows that the nested data mining algorithm may obtain the optimal operating condition, which can be practical guidance, by the less time cost.更多还原