基于Rough集理论的薄膜蒸发器产量预测  

Study on thin-film evaporator's output forecasting based on the Rough set theory

作　　者：华洁[1] 李庆生[1] 尹侠[1] 李永生[1] 

机构地区：[1]南京工业大学机械与动力工程学院,江苏南京210009

出　　处：《南京工业大学学报（自然科学版）》2006年第3期94-99,共6页Journal of Nanjing Tech University(Natural Science Edition)

摘　　要：由于薄膜蒸发器生产过程高度复杂,导致产量的预测比较困难且准确率低。采用基于Rough集理论的数据分析工具包Rosetta,实现了薄膜蒸发器的产量预测研究。该研究完成了知识发现的全过程包括数据准备、数据分析以及结果的解释评估。研究结果表明:对条件属性的约简找出了对产量影响较大的因素,如试验时间、流量、蒸发强度、导热油温度以及刮板转速等;对产量的预测基本符合实际,总体正确率达73.7%。The throughput forecasting of thin-film evaporator is quite difficult and the precision is quite low because of high complexity of the process. By the Rosetta data analysis saddlebag based on the Rough set theory, the total process of knowledge discovery in databases （KDD） was accomplished including data preparing, data analysis and the interpreting of the results. The results indicated that the simplification of condition offered possibility to find out the important factors to affect the output such as test time, flux, evaporation intensity, the temperature of the heatconductive oil and rotation rate of scraper. The forecasting was examined with the practice and the total correct ness reached 73.7%.

关 键 词：ROUGH集 薄膜蒸发器 预测 

分 类 号：TQ051.62[化学工程] TP18[自动化与计算机技术—控制理论与控制工程]