基于融合建模的浮选药剂消耗预测研究

浮选是选矿技术中常用的技术之一,为了使选矿工艺指标达到要求,需要加入捕收剂、抑制剂、活化剂和 pH 调整剂等药剂。药剂消耗在影响浮选工艺指标的同时也影响着选矿的成本,因此预测每天、周、月、年的药剂消耗量,对其进行定额管理是浮选作业中重要的研究问题,基于缺失数据填补技术采用融合建模的方法在原矿性质、浮选指标及药剂消耗之间建立带有混合效应的非线性模型,并预测 7 d（1 周）、30 d（1 月）、365 d（1 年）的浮选药剂消耗量。预测结果与生产数据对比表明,该模型实现了对选厂浮选药剂用量的短期精确预测。实际数据显示该方法对生产实践具有指导意义。

Prediction of Flotation Reagents Consumption Based on Fusion Modeling

Flotation is one of the commonly used technologies in beneficiation technology. In order to meet the requirements of beneficiation process indicators,it is necessary to add agents such as collectors,inhibitors,activators and pH adjusters.The consumption of chemicals affects the cost of beneficiation while affecting the flotation process indicators,therefore,forecasting daily,weekly,monthly and yearly consumption of pharmaceuticals and managing consumption of pharmaceuticals is an important research question. Based on the missing data filling technique,uses fusion modeling to establish a non-linear model with mixed effects on ore properties,flotation index and reagent consumption,and predict 7 d (1 week),30 d (1 month),365 d (1 year) consumption of flotation agent. The comparison between the prediction results and the production data shows that the model achieves short-term accurate prediction of the flotation agent dosage in the concentrator. The actual data show that the method has guiding significance for production practice.