典型赤铁矿磨矿过程智能运行反馈控制

Intelligent operational feedback control for typical hematite grinding processes

冶金磨矿是典型的高能耗、低效率过程,其控制与优化不仅仅是使常规过程控制系统尽可能好地跟踪期望设定值,而且要控制整个过程运行,实现表征磨矿整体运行性能的磨矿粒度与生产效率等运行指标的优化.针对我国广泛使用的赤铁矿两阶段全闭路磨矿,由于其原矿石性质与成份复杂且不稳定、粒级波动大,磨矿运行指标不能在线测量,工况时变,难以建立过程数学模型,提出基于数据与知识的智能运行反馈控制方法,包括基于案例推理的控制回路预设定、磨矿粒度动态神经网络软测量以及多变量模糊动态调节器.为了验证所提方法的有效性,将所题方法应用于中国某大型赤铁矿选厂,取得显著应用成效.

Metallurgical grinding is a typical process of high energy consumption and low efficiency; its control andoptimization will make the outputs of the controlled processes best follow their setpoints, and optimize the entire plant sothat the grinding particle size and efficiency during the production phase being maintained within their desired ranges. Forthe complex hematite grinding processes which are widely used in China, the processed hematite has complex compositionand properties, and the operational indices cannot be measured online. Moreover, their dynamic characteristics are veryintricate, making the accurate dynamic models of them are difficult to be built. An intelligent operational feedback controlapproach is proposed which is based on data and knowledge. This approach includes a case-based reasoning loop pre-setting controller, an artificial neural network （ANN）-based particle size soft sensor module, and a multivariable fuzzydynamic adjustor. The proposed method has been successfully applied to the grinding process of a large hematite mineralprocessing plant in China.