摘要
针对铝土矿连续磨矿过程球磨机节能降耗问题以及铝土矿来源复杂、品位差异大等特点,提出了球磨机多目标多模型预测控制方法.该方法首先建立状态空间浓度预测模型和粒级质量平衡加权多模型细度预测模型.然后构建了包含磨机排矿浓细度区间控制和经济性能指标的多目标优化结构的多模型预测控制策略.最后采用乘子罚函数法求解控制器局部最优解.仿真及现场试验结果表明了该方案的有效性.
Considering the reduction of power consumption of ball-mill, we propose a multi-objective multi-model predictive control for the continuous grinding process of bauxite with bauxite ores coming from different mine sources and with different qualities. In this method, we first build the state-space concentration-predictive model and the finenessprediction model based on the weighted multi-model of size-mass balance; and then, we develop an optimal multi-model predictive control scheme for optimizing multiple objectives including the interval control of concentration and fineness of the discharged ore pulp from the ball-mill, along with economic indices. The local optimal control law of the controller is obtained by minimizing a multiplier penalty function. The simulation and the field test results show the effectiveness of this method.
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
2012年第10期1339-1347,共9页
Control Theory & Applications
基金
国家自然科学基金资助项目(61134006
61273187)
新世纪优秀人才支持计划资助项目(NCET 08 0576)
关键词
磨矿过程
多模型预测控制
多目标优化
区间控制
乘子罚函数
mineral grinding process
multiple model predictive control
multiple objective optimization
interval control
multiplier penalty function
