| 引用本文: | 张俊,毛志忠,贾润达.金氰化浸出过程实时优化[J].控制理论与应用,2014,31(9):1198~1205.[点击复制] |
| ZHANG Jun,MAO Zhizhong,JIA Runda.Real-time optimization for gold cyanidation leaching process[J].Control Theory and Technology,2014,31(9):1198~1205.[点击复制] |
|
| 金氰化浸出过程实时优化 |
| Real-time optimization for gold cyanidation leaching process |
| 摘要点击 1985 全文点击 2138 投稿时间：2013-05-09 修订日期：2014-06-04 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI编号 10.7641/CTA.2014.30453 |
| 2014,31(9):1198-1205 |
| 中文关键词 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 |
| 英文关键词 cyanidation leaching mechanism modeling Tikhonov regularization model uncertainty real-time opti- mization neighboring-extremal control output feedback direct input adaptation |
| 基金项目 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). |
|
| 中文摘要 |
| 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. |
| 英文摘要 |
| 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 con- centration 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. |
|
|
|
|
|