基于机理模型的改进氧量软测量方法及软件实现

Improvement for soft-sensing of oxygen content based on mechanism model and software implementation

针对现有氧量分析仪具有的一些缺点,提出一种改进的烟气含氧量软测量方法。根据工艺机理建立关于热量与风量的氧量软测量模型;利用数据挖掘与回归分析方法修正了传感器测得的风量以及标准状态下的单位热量耗空气量系数;给出关于算法的软件实现以及实现过程中的数据预处理、有效时间选取、公式校正3个关键环节。利用某电厂历史数据进行测试,对氧量预测数据与实际值进行比较并对均方差与相关系数两项指标进行分析,结果表明了该方法的有效性。

An improved soft-sensing method of oxygen content in flue gas is presented considering the disadvan-tages of existing oxygen-content analyzers.A soft-sensing model about air flow and heat is proposed through mecha-nism analysis;air flow measured by sensor and coefficient about the amount of air under standard state required for unit quantum of heat are corrected by data mining and regression analysis;software implementation based on the proposed method and three key procedures,namely,data preprocessing,effective time selection and formula modification have been given.Experiment has been performed by history data of certain power plant.Both comparison between predictive and practical oxygen content and analysis of mean square deviation and correlation coefficient show that the method is effective.