运行于多变边界条件下给火电机组深度节能理论研究带来新的挑战。针对这一问题,提出了多变条件下火电机组能耗基准状态概念,利用机组的海量实际运行数据,基于模糊粗糙集(fuzzy rough set,FRS)约简和模糊均值聚类(fuzzy C mean,FCM)等数据挖掘方法构建机组性能优化知识库,在可比历史边界条件下动态寻优确定机组能耗基准状态。以某600MW亚临界机组为案例开展了实例研究,结果表明该方法具有快速、自适应性、高度复现性和可动态调整优化等特点,适用于机组在不同工况和边界条件下的能耗分析与节能诊断。
The in-depth energy conservation of thermal power units is confronting new challenges under the varying operation conditions and ambient constraints.Hence,a concept of energy-consumption benchmark state was proposed to describe the economic performance of thermal power units.According to the mass operation data storage in supervisory information system(SIS) of power units,and based on the data mining method,such as fuzzy rough set(FRS)-based decision table reduction and fuzzy C mean(FCM)-based clustering,the energy-consumption benchmark state was determined under the comparable operation boundary.The method was performed on a 600MW subcritical power unit.The results show that the proposed method is fast,adaptive,recurrent and automatically adjustable under different operation conditions and external constraints of power units.