基于主动学习的浸出过程pH稳定控制研究  

Study on pH Value Stability Control in Zinc Leaching Process Based on Active Learning

作　　者：陈宇 黄煜栋 刘学斌 李海生 劳佳锋[1] CHEN Yu;HUANG Yudong;LIU Xuebin;LI Haisheng;LAO Jiafeng(School of Intelligent Manufacturing,Hangzhou Polytechnic,Hangzhou 311402,China;School of Automation,Central South University,Changsha 410083,China)

机构地区：[1]杭州科技职业技术学院物联网技术学院,杭州311402 [2]中南大学自动化学院,长沙410083

出　　处：《有色金属(冶炼部分)》2025年第5期27-35,共9页Nonferrous Metals(Extractive Metallurgy)

基　　金：浙江省自然科学基金杭州区域创新发展联合基金资助项目(LHZY24A010006);浙江省教育厅一般项目(Y202455590,Y202352510)。

摘　　要：在湿法炼锌的中性浸出过程中,pH的稳定性直接影响产品的生产质量。针对传统人工控制方法在应对溶剂浓度波动和大时滞所带来的pH稳定控制难题,提出了一种基于主动学习分类的模糊规则提取策略。通过深入分析过程机理,将中性浸出过程细分为多个典型工况。利用近似线性依赖方法和主动学习算法,从大量历史数据中精确筛选出代表典型工况的信息样本。随后,采用支持向量机算法,在不同的工况下提取支持向量,并构建模糊规则集。通过实际工业数据的案例进行仿真验证,结果表明,该方法在不同工况下的pH控制合格率高达93.17%,方差低至0.011,充分验证了算法的有效性与实用性。该成果为中性浸出过程的pH稳定控制提供了坚实而有效的支撑。In the neutral leaching process of zinc hydrometallurgy,the stability of pH value directly affects the quality of the product.A fuzzy rule extraction strategy based on active learning classification was proposed to address the challenge of pH value stability control caused by solvent concentration fluctuations and large time delays in traditional manual control methods.The neutral leaching process was subdivided into multiple typical operating conditions by in-depth analysis of the process mechanism.Employing the approximate linear dependency method and active learning algorithm,the information samples representing typical working conditions were accurately screened from a large number of historical data.Then,the support vector machine algorithm was used to extract support vectors under different operating conditions and a fuzzy rule set was constructed.The results show that the qualified rate of pH value control under different working conditions is as high as 93.17%,and the variance is as low as 0.011,which fully verifies the effectiveness and practicability of the algorithm.The results provide a solid and effective support for pH stability control in neutral leaching process.

关 键 词：中性浸出 PH控制 主动学习 支持向量机 规则提取 

分 类 号：TF813[冶金工程—有色金属冶金]