论文部分内容阅读
为分析水泥生产过程各环节的实际能耗水平,提出了一种改进的基于密度的带有噪声的空间聚类(density-based spatial clustering of applications with noise,DBSCAN)算法,并结合线性最小方差融合算法建立了水泥生产过程各能耗变量的实际标杆.针对大时间尺度下能耗数据的聚类数目因工况变化和噪声数据而无法直接获得的问题,采用改进的DBSCAN算法对水泥生产过程各环节的能耗历史数据分别进行聚类分析,获得了各能耗变量的分类、聚类中心及其方差.利用各聚类的中心及其方差,采用线性最小方差融合算法分别对各能耗变量的数据进行优化融合,得到包含综合影响因素的各环节实际能耗标杆值.应用实例表明:改进的DBSCAN算法能减少核心对象的查询次数,有效降低算法的执行时间;通过数据融合得到的能耗标杆能够合理反映水泥生产过程实际能耗水平,揭示企业节能潜力.
In order to analyze the actual energy consumption of each stage of cement production process, an improved density-based spatial clustering of applications with noise (DBSCAN) algorithm is proposed. Combined with linear minimum variance fusion The algorithm establishes the actual benchmark of each energy consumption variable in cement production process.Aiming at the problem that the number of clusters of energy consumption data in large time scale can not be obtained directly due to the change of working conditions and noise data, The energy consumption historical data of each link were clustered to obtain the classification of each variable of energy consumption, the center of clustering and its variance.Using the center of each cluster and its variance, linear least square variance fusion algorithm was used to calculate the energy consumption variables , The actual energy consumption benchmark of each link including the comprehensive influencing factors is obtained.Application examples show that the improved DBSCAN algorithm can reduce the number of queries of the core objects and reduce the execution time of the algorithm effectively.The energy consumption through data fusion Benchmark can reasonably reflect the actual energy consumption of cement production process, revealing the potential of energy-saving enterprises.