摘要
湿法冶金过程具有反应机理复杂、工艺流程长、工序众多等特点,由于模型误差等因素,基于模型得到的生产过程最优工作点不是实际生产过程的最优工作点.如何保持湿法冶金生产流程运行在经济效益最优的状态成为生产优化控制的难点.本文提出了一种基于数据的湿法冶金过程操作量优化设定补偿方法.该方法在基于模型得到的最优工作点基础上,采用即时学习(Just-in-time learning,JITL)的思想,在当前工作点附近利用历史数据建立操作量补偿值和经济效益增量的相关模型,优化求解在当前工作点下,使经济效益增量最大化的操作量补偿值,施加到生产流程,并在新工作点进行迭代补偿.将所提出的方法仿真应用于某精炼厂的湿法冶金生产流程,仿真结果验证了所提出方法的有效性.
Hydrometallurgical process has the characteristics of complicated reaction mechanism, long process flow and many sub-processes. The optimal setpoint of a model-based process based on the model is not the optimal working point of the actual process due to model error and so on. How to keep the hydrometallurgical process running in the state with optimal economic efficiency has become the difficulty of production optimization control. In this paper, a method based on data is proposed to optimize the operation of hydrometallurgical process. Based on the optimal setpoint of the model,the just-in-time learning（JITL） idea is used to establish a relevant model of the manipulative compensation value and the economic benefit increment in the vicinity of the current working point. At the work point, the amount of compensation that maximizes the economic gain is applied to the production process and iterated at the new work point. Simulation is carried out to verify the proposed method with the hydrometallurgical production process of a refinery, and the results show its effectiveness of the proposed method.
出处
《自动化学报》
EI
CSCD
北大核心
2017年第6期1047-1055,共9页
Acta Automatica Sinica
基金
国家自然科学基金(61533007)
流程工业综合自动化国家重点实验室基础科研业务费(2013ZCX02-04)资助~~
关键词
湿法冶金
基于数据
优化补偿
即时学习
Hydrometallurgy
data-based
optimized compensating
JITL（Just-in-time learning）
