金氰化浸出过程实时优化  

Real-time optimization for gold cyanidation leaching process

作　　者：张俊[1] 毛志忠[1,2] 贾润达[1,2] 

机构地区：[1]东北大学信息科学与工程学院,辽宁沈阳110819 [2]东北大学流程工业综合自动化国家重点实验室,辽宁沈阳110819

出　　处：《控制理论与应用》2014年第9期1198-1205,共8页Control Theory & Applications

基　　金：国家"863"计划资助项目(2011AA060204);国家自然科学基金资助项目(61203103);中央高校基本科研业务费资助项目(N110304006)

摘　　要：本文以某湿法冶炼厂金氰化浸出过程为背景,建立了动态机理模型,为提高模型参数辨识精度,提出了基于Tikhonov正则化思想利用含噪声的浓度测量数据估计动力学反应速度的策略.为减小模型与实际过程不匹配对实时优化结果的影响,提出了基于对数–线性闸–罚函数和输出反馈的直接输入自适应方法,加入闸–罚函数后,该方法可以解决含有不等式约束的优化问题,将其应用到金氰化浸出过程实时优化中,仿真结果表明在输出测量值无噪声时,该方法能很快地局部收敛到实际过程的最优设定点;而当测量噪声较小或对目标函数影响较小时,该方法也显示出了优越的性能,而且只需要过程标称模型和实际输出,不需要求实际过程数据梯度,受测量噪声影响较小,更易于实际实施,这为湿法冶金全流程优化控制的顺利实施奠定了重要基础.For the gold cyanidation leaching process in some hydrometallurgy plant,we established a dynamic mech anism model.The Tikhonov regularization method was used to estimate the unavailable kinetic reaction rates with concentration measurements to improve the identification accuracy of model parameters.To reduce the impact resulted from the mismatch between the model and the actual process,we proposed the direct input adaptation method based on the logarithmic linear barrier-penalty function and the output feedback,and applied it to the gold cyanidation leaching process to solve the optimization problem with inequality constraints.Simulation results showed that if the measurement noise was negligibly small,the input could converge locally to the optimal set point of the actual process without the need of estimates of measurement gradients,but only the nominal model and output measurements are required.This facilitates the implementation in practice and even forms an important basis for successfully implementing the plant-wide optimization and control for hydrometallurgy process.

关 键 词：氰化浸出 机理建模 TIKHONOV正则化 模型不确定性 实时优化 邻域极值控制 输出反馈 直接输入自适应 

分 类 号：TF831[冶金工程—有色金属冶金]