气流干燥过程水份软测量集成建模研究

Research on Integrated Modeling of Soft Sensor for Moisture of Pneumatic Drying

气流干燥是闪速熔炼的关键工序之一,控制干燥后精矿的含水率在0.1%-0.3%之间是稳定熔炼生产的前提;由于水份含量人工检测的时间间隔较长,很难及时反映生产实际,影响干燥过程的优化控制;采用主元分析的方法,建立了水份含量的主元回归模型,并利用BP神经网络模型进行误差修正,实现对干精矿含水率的软测量;实际应用表明,该集成模型精度高,能满足工业生产要求。

Pneumatic drying is one of the key procedures in the process of flash smelter, keeping the moisture content of ore ranging between 0. 1 % and 0. 3% is the precondition of flash smelting process. But the interval of manual detection is too long to reflect the producing condition in time, which infected the optimal control of drying. The soft measurement model is established based on PCA-regression, where the BP neural network is adopted for error correction. The application results show that the integrated model has highl precision and can meet the request of industrial production.