基于数据驱动的浓密-压滤过程协调优化控制

Data driven coordinated optimization control of thickening-filter process

针对某选矿厂由于浓密-压滤过程关键变量没有实现在线检测,导致该工序生产操作无序、生产指标难以达标、能耗经济指标高等问题,利用浓密-压滤过程的生产运行数据,提出一种基于数据驱动的浓密-压滤过程协调优化控制方法.首先,通过偏最小二乘(PLS)方法建立浓密-压滤过程的数据模型;然后,在阶梯电价、浓密机运行安全、生产指标的约束下,以浓密-压滤过程能耗经济指标最小为目标,建立浓密-压滤过程协调优化模型,规划放矿压滤时间序列.离线实验与现场应用表明,所建立优化模型能提高浓密-压滤过程底流浓度生产指标、降低过程能耗经济指标、减少滤布损耗、降低生产异常次数等应用效果.

Since the key variables in the thickening-filter process of a concentrator can not be measured online, there are problems in the production process such as improper operation, substandard production index and high energy economic index(EEI). Based on the production data of the thickening-filter process, a data-driven coordination optimization control method for the thickening-filter process is proposed. Firstly, the partial least squares(PLS) method is used to establish the data driven model of thickening-filter process. Then, under the constraints of stepped electricity price, operation safety of thickeners and production indexes, a coordinated optimization model aiming at the minimum EEI in the thickening-filter process is established to plan the time series of ore drawing and filter-press. An off-line experiment and field application show that the optimization model can improve the production index of the underflow concentration, reduce the EEI in the process of thickening-filter, the loss of press cloth and the number of abnormal production.