2.624

2020影响因子

(CJCR)

  • 中文核心
  • EI
  • 中国科技核心
  • Scopus
  • CSCD
  • 英国科学文摘

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于改进随机森林算法的工业过程运行状态评价

常玉清 孙雪婷 钟林生 王福利 刘英娇

常玉清, 孙雪婷, 钟林生, 王福利, 刘英娇. 基于改进随机森林算法的工业过程运行状态评价. 自动化学报, 2021, 47(9): 2214−2225 doi: 10.16383/j.aas.c190066
引用本文: 常玉清, 孙雪婷, 钟林生, 王福利, 刘英娇. 基于改进随机森林算法的工业过程运行状态评价. 自动化学报, 2021, 47(9): 2214−2225 doi: 10.16383/j.aas.c190066
Chang Yu-Qing, Sun Xue-Ting, Zhong Lin-Sheng, Wang Fu-Li, Liu Ying-Jiao. Industrial operation performance evaluation of industrial processes based on modified random forest. Acta Automatica Sinica, 2021, 47(9): 2214−2225 doi: 10.16383/j.aas.c190066
Citation: Chang Yu-Qing, Sun Xue-Ting, Zhong Lin-Sheng, Wang Fu-Li, Liu Ying-Jiao. Industrial operation performance evaluation of industrial processes based on modified random forest. Acta Automatica Sinica, 2021, 47(9): 2214−2225 doi: 10.16383/j.aas.c190066

基于改进随机森林算法的工业过程运行状态评价

doi: 10.16383/j.aas.c190066
基金项目: 国家自然科学基金(61673092, 61533007, 61304121, 61973057, 61873053), 创新研究群体科学基金(61621004), 中央高校基础科研业务费(N150404017), 矿冶过程自动控制技术国家重点实验室开放基金(BGRIMM-KZSKL-2018-08)资助
详细信息
    作者简介:

    常玉清:东北大学教授. 2002年于东北大学获得博士学位. 主要研究方向为复杂工业过程建模、监测、运行状态评价及过程优化. E-mail: changyuqing@ise.neu.edu.cn

    孙雪婷:东北大学硕士研究生. 2017年于大连交通大学获得学士学位. 主要研究方向为复杂工业过程的建模控制与优化. 本文通信作者. E-mail: xuetingsun111@163.com

    钟林生:东北大学信息科学与工程学院博士研究生. 主要研究方向为复杂工业过程运行状态评价和故障诊断. E-mail: zhonglinsheng_neu@163.com

    王福利:东北大学信息科学与工程学院教授. 主要研究方向为复杂工业过程建模、优化与故障诊断. E-mail: wangfuli@ise.neu.edu.cn

    刘英娇:东北大学硕士研究生. 2015年于东北大学秦皇岛分校获得学士学位. 主要研究方向为复杂工业过程的建模控制与优化. E-mail: liuyingjiao1992@163.com

  • 收稿日期 2019-01-27 录用日期 2019-09-09 Manuscript received January 27, 2019; accepted September 9,2019 国家自然科学基金 (61673092, 61533007, 61304121, 61973057, 61873053), 创新研究群体科学基金 (61621004), 中央高校基础科研业务费 (N150404017), 矿冶过程自动控制技术国家重点实验室开放基金 (BGRIMM-KZSKL-2018-08) 资助 Supported by National Natural Science Foundation of China (61673092, 61533007, 61304121, 61973057, 61873053), Science Fou-ndation for Innovative Research Groups (61621004), Fundamental Research Funds for the Central Universities (N150404017), andOpen Foundation of State Key Laboratory of Process Automationin Mining and Metallurgy (BGRIMM-KZSKL-2018-08) 本文责任编委 伍洲 Recommended by Associate Editor WU Zhou 1. 东北大学信息科学与工程学院 沈阳 110819 2. 流程工业综合自动化国家重点实验室 (东北大学) 沈阳 110819 1. College of Information Science and Engineering, Northeast-
  • ern University, Shenyang 110819 2. State Key Laboratory ofSynthetical Automation for Process Industries (NortheasternUniversity), Shenyang 110819

Industrial Operation Performance Evaluation of Industrial Processes Based on Modified Random Forest

Funds: Supported by National Natural Science Foundation of China (61673092, 61533007, 61304121, 61873053), Science Foundation for Innovative Research Groups (61621004), Fundamental Research Funds for the Central Universities (N150404017), and Open Foundation of State Key Laboratory of Process Automation in Mining and Metallurgy (BGRIMM-KZSKL-2018-08)
More Information
    Author Bio:

    CHANG Yu-Qing Professor at Northeastern University. She received her Ph.D. degree from Northeastern University in 2002. Her research interest covers industrial process modeling, monitoring, operation performance evaluation, and optimization

    SUN Xue-Ting Master student at Northeastern University. She received her bachelor degree from Dalian Jiaotong University in 2017. Her research interest covers modeling control and optimization of complex industrial processes. Corresponding author of this paper

    ZHONG Lin-Sheng Ph.D. candidate at the College of Information Science and Engineering, Northeastern University. His research interest covers operating performance assessment and fault diagnosis of complex industrial processes

    WANG Fu-Li Professor at Northeastern University. His research interest covers modeling control, optimization, and fault diagnosis of complex industrial processes

    LIU Ying-Jiao Master student at Northeastern University. She received her bachelor degree from Northeastern University at Qinhuangdao in 2015. Her research interest covers modeling control and optimization of complex industrial processes

  • 摘要: 运行状态评价是指在过程正常生产的前提下, 进一步判断生产过程运行状态的优劣. 针对复杂工业过程定量信息与定性信息共存的情况, 本文提出了一种基于随机森林的工业过程运行状态评价方法. 针对随机森林中决策树信息存在冗余的问题, 基于互信息将传统随机森林中的决策树进行分组, 并选出每组中最优的决策树组成新的随机森林. 同时为了强化评价精度高的决策树和弱化评价精度低的决策树对最终评价结果的影响, 使用加权投票机制取代传统众数投票方法, 最终构成一种基于互信息的加权随机森林算法(Mutual information weighted random forest, MIWRF). 对于在线评价, 本文通过计算在线数据处于各个等级的概率, 并且结合提出的在线评价策略, 判定当前样本运行状态等级. 为了验证所提算法的有效性, 将所提方法应用于湿法冶金浸出过程, 实验结果表明, 相对于传统随机森林算法, MIWRF 降低了模型的复杂度, 同时提高了运行状态评价精度.
    1)  收稿日期 2019-01-27 录用日期 2019-09-09 Manuscript received January 27, 2019; accepted September 9,2019 国家自然科学基金 (61673092, 61533007, 61304121, 61973057, 61873053), 创新研究群体科学基金 (61621004), 中央高校基础科研业务费 (N150404017), 矿冶过程自动控制技术国家重点实验室开放基金 (BGRIMM-KZSKL-2018-08) 资助 Supported by National Natural Science Foundation of China (61673092, 61533007, 61304121, 61973057, 61873053), Science Fou-ndation for Innovative Research Groups (61621004), Fundamental Research Funds for the Central Universities (N150404017), andOpen Foundation of State Key Laboratory of Process Automationin Mining and Metallurgy (BGRIMM-KZSKL-2018-08) 本文责任编委 伍洲 Recommended by Associate Editor WU Zhou 1. 东北大学信息科学与工程学院 沈阳 110819 2. 流程工业综合自动化国家重点实验室 (东北大学) 沈阳 110819 1. College of Information Science and Engineering, Northeast-
    2)  ern University, Shenyang 110819 2. State Key Laboratory ofSynthetical Automation for Process Industries (NortheasternUniversity), Shenyang 110819
  • 图  1  基于互信息的加权随机森林算法

    Fig.  1  Weighted random forest algorithm based on mutual information

    图  2  运行状态等级概率

    Fig.  2  Probability of grade of running state

    图  3  运行状态评价等级

    Fig.  3  Grade of running state

    图  4  优到次优的转换

    Fig.  4  Transformation from optimal to suboptimal

    图  5  次优到非优的转换

    Fig.  5  Transformation from suboptimal to non-optimal

    图  6  次优到优的转换

    Fig.  6  Transformation from suboptimal to optimal

    图  7  非优到次优的转换

    Fig.  7  Transformation from non-optimal to suboptimal

    图  8  优、次优到非优的转换

    Fig.  8  Transformation from optimal, suboptimal to non-optimal

    表  1  浸出过程变量列表

    Table  1  Key variables affecting leaching efficiency

    分割方法单位属性
    矿石初始来料量${\rm{kg}}$定量
    浸出调浆后矿浆浓度%定量
    矿石初始金品位${\rm{g/t}}$定性
    一浸浸出槽1氰化钠添加量${\rm{kg/ h}}$定量
    一浸浸出槽2氰化钠添加量${\rm{kg/ h}}$定量
    一浸浸出槽4氰化钠添加量${\rm{kg/ h}}$定量
    一浸浸出槽1槽空气流量${\rm{m^3/ h}}$定量
    一浸浸出槽2槽空气流量${\rm{m^3/ h}}$定量
    一浸浸出槽3槽空气流量${\rm{m^3/ h}}$定量
    一浸浸出槽4槽空气流量${\rm{m^3/ h}}$定量
    二浸前矿浆浓度%定量
    一浸后金品位${\rm{g/ t}}$定性
    二浸浸出槽1氰化钠添加量${\rm{kg/ h}}$定量
    二浸浸出槽2氰化钠添加量${\rm{kg/ h}}$定量
    二浸浸出槽4氰化钠添加量${\rm{kg/ h}}$定量
    二浸浸出槽1槽空气流量${\rm{m^3/ h}}$定量
    二浸浸出槽2槽空气流量${\rm{m^3/ h}}$定量
    二浸浸出槽3槽空气流量${\rm{m^3/ h}}$定量
    二浸浸出槽4槽空气流量${\rm{m^3/ h}}$定量
    下载: 导出CSV

    表  2  浸出过程实验设计

    Table  2  Experiment of leaching process

    数据等级描述
    1~3041~160个样本点, 过程运行状态等级为“优”; 自第161个样本点, 一浸槽4氰化钠添加量逐渐减少, 运行状态等级优性逐渐减弱, 但始终保持“优”等级, 直到第304个样本点.
    305~621次优314~479个样本点, 保持一浸槽4氰化钠添加量不变, 过程稳定运行于等级“次优”; 480~621个样本点, 持续减少槽4 氰化钠添加量, 运行状态等级优性再次减弱.
    622~850非优保持一浸槽4氰化钠添加量不再变化, 过程稳定运行于等级“非优”
    下载: 导出CSV

    表  3  RF与MIWRF实验结果对比

    Table  3  Comparison of experimental results of RF and MIWRF

    实验编号RF决策树数量MIWRF决策树数量RF评价精度 (%)MIWRF评价精度 (%)决策树减少量 (%)
    1501992.994.962.0
    2602493.095.060.0
    3703193.095.155.8
    4803293.995.360.0
    5903693.595.760.0
    61003993.696.261.0
    71104393.695.960.9
    81204693.595.761.7
    91305193.596.060.8
    101405493.695.961.4
    111505893.595.761.3
    121606293.695.860.6
    131706393.696.162.9
    141806593.596.263.9
    151906993.696.163.7
    162007193.796.264.5
    下载: 导出CSV

    表  4  6种评价方法的运行状态评价性能

    Table  4  Performances of 6 evaluation methods

    方法KNNANNRFMIRFMIWRFFDbD
    精度 (%)88.891.293.695.296.293.3
    建模时间 (s)051.151927.450723.087123.683128.7613
    测试时间 (s)0.690110.119331.31230.90620.96391.1534
    下载: 导出CSV
  • [1] Ye L, Liu Y, Fei Z. Online Probabilistic Assessment of Operating Performance Based on Safety and Optimality Indices for Multimode Industrial Processes. Industrial and Engineering Chemistry Research, 2009.48(24): 10912−10923 doi: 10.1021/ie801870g
    [2] Zou X, Wang F, Chang Y. Process operating performance optimality assessment and non-optimal cause identification under uncertainties. Chemical Engineering Research and Design, 2017, 120(Complete): 348−359
    [3] Chang Y Q, Zou X Y, Wang F, Zhao L, Zheng W. Multimode plant-wide process operating performance assessment based on two-level multi-block hybrid model. Chemical Engineering Research and Design, 2018, 136: 721−733 doi: 10.1016/j.cherd.2018.05.023
    [4] 王东, 刘怀亮, 徐国华. 案例推理在故障诊断系统中的应用. 计算机工程, 2003, 29(12): 10−12 doi: 10.3321/j.issn:1002-8331.2003.12.004

    Wang Dong, Liu Huai-Liang, Xu Guo-Hua. Application of case-based reasoning in faulty dignoses system. Computer Engineering, 2003, 29(12): 10−12 doi: 10.3321/j.issn:1002-8331.2003.12.004
    [5] 姜涛, 韩富春, 范卫星. 基于粗糙集理论的架空输电线路运行状态评估. 电气技术, 2007(12): 58−60

    Jiang Tao, Han Fu-Chun, Fan Wei-Xing. Evaluating the condition of overhead transmission lines based on rough sets theory. Computer Engineering, 2007(12): 58−60
    [6] Guo L, Gao J, Yang J, et al. Criticality evaluation of petrochemical equipment based on fuzzy comprehensive evaluation and a BP neural network. Journal of Loss Prevention in the Process Industries, 2009, 22(4): 469−476 doi: 10.1016/j.jlp.2009.03.003
    [7] Xu G, Yang Y P, Lu S Y, et al. Comprehensive evaluation of coal-fired power plants based on grey relational analysis and analytic hierarchy process. Energy Policy, 2011, 39(5): 2343−2351 doi: 10.1016/j.enpol.2011.01.054
    [8] Liu Y, Wang F, Chang Y. Online Fuzzy Assessment of Operating Performance and Cause Identification of Nonoptimal Grades for Industrial Processes. Industrial and Engineering Chemistry Research, 2013, 52(50): 18022−18030 doi: 10.1021/ie402243s
    [9] 邹筱瑜, 常玉清, 王福利, 周阳. 基于 GMM 和贝叶斯推理的多模态过程运行状态评价. 自动化学报, 2016, 33(2): 164−171

    Zou Xiao-Yu, Chang Yu-Qing, Wang Fu-Li, Zhou Yang. Operation performance assessment for multimode processes based on GMM and Bayesian inference. Control Theory and Applications, 2016, 33(2): 164−171
    [10] Liu Y, Wang F, Chang Y. Operating optimality assessment and nonoptimal cause identification for multimode industrial process with transitions. The Canadian Journal of Chemical Engineering, 2016, 94(7): 1342−1353 doi: 10.1002/cjce.22513
    [11] Breiman Leo. Bagging predictors. Machine Learning, 1996, 24(2): 123−140
    [12] Tuv E, Borisov A, Runger G, Torkkola K. Feature Selection with Ensembles, Artificial Variables, and Redundancy Elimination. Journal of Machine Learning Research, 2009, 10(3): 1341−1366
    [13] Paul A, Mukherjee D P, Das P, et al. Improved Random Forest for Classification. IEEE Transactions on Image Processin, 2018, 27(8): 4012−4024 doi: 10.1109/TIP.2018.2834830
    [14] Schulter S, Wohlhart P, Leistner C, Saffari A, Roth P M, Bischof H. Alternating decision forests. In: Proceedings of the 2013 IEEE Conference on Computer Vision and Pattern Recognition. Portland, Oregon, USA: IEEE, 2013. 508−515
    [15] Liu Y, Ge Z. Weighted random forests for fault classification in industrial processes with hierarchical clustering model selection. Journal of Process Control, 2018, 64: 62−70 doi: 10.1016/j.jprocont.2018.02.005
    [16] Breiman L. Random Forests. Machine Learning, 2001, 45(1): 5−32 doi: 10.1023/A:1010933404324
    [17] Tao S. Steve H. Unsupervised learning with random forest predictors. Journal of Computational and Graphical Statistics, 2006, 15(1): 118-138
    [18] Ho T K. The random subspace method for constructing decision forests. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1998, 20(8): 832−844 doi: 10.1109/34.709601
    [19] 王丽婷, 丁晓青, 方驰. 一种鲁棒的全自动人脸特征点定位方法. 自动化学报, 2006, 35(1): 9−16

    Wang Li-Ting, Ding Xiao-Qing, Fang Chi. A novel method for robust and automatic facial features localization. Acta Automatica Sinica, 2006, 35(1): 9−16
    [20] 方匡南, 吴见彬, 朱建平. 随机森林方法研究综述. 统计与信息论坛, 2011, 26(3): 32−38 doi: 10.3969/j.issn.1007-3116.2011.03.006

    Fang Kuang-Nan, Wu Jian-Bin, Zhu Jian-Ping. A review of technologies random forests. Statistics and Information Forum, 2011, 26(3): 32−38 doi: 10.3969/j.issn.1007-3116.2011.03.006
    [21] Krogh A, Vedelsby J. Neural network ensembles, cross validation and active learning. In: Proceedings of the 1995 International Conference on Neural Information Processing Systems. MIT Press: 1995. 231−238
    [22] Cover T, Thomas J, Wiley J. Elements of Information Theory. Tsinghua University Press, 2003.
    [23] Robnik-Šikonja M. Improving random forests. In: Proceedings of the 2004 European Conference on Machine Learning (ECML 2004). Berlin, Germemy: Springer Press, 2004. 359−370
    [24] 陈家镛. 湿法冶金的研究与发展. 冶金工业出版社, 1998.

    Chen Jia-Yong. Research and Development of Hydrometallurgy. Beijing: Metallurgical Industry Press. 1998.
    [25] 刘维平, 邱定蕃, 卢惠民. 湿法冶金新技术进展. 矿冶工程, 2003, 23(5): 39−42 doi: 10.3969/j.issn.0253-6099.2003.05.014

    Liu Wei-Ping, Qiu Ding-Fan, Lu Hui-Min. New advances in hydrometallurgical technology. Mining and Metallurgical Engineering, 2003, 23(5): 39−42 doi: 10.3969/j.issn.0253-6099.2003.05.014
    [26] 马荣骏. 湿法冶金新发展. 湿法冶金, 2007, 26(1): 1−12 doi: 10.3969/j.issn.1009-2617.2007.01.001

    Ma Rong-Jun. New development of hydrometallurgy. Hydrometallurgy of China, 2007, 26(1): 1−12 doi: 10.3969/j.issn.1009-2617.2007.01.001
    [27] 常玉清, 许弟伍. 含不确定信息的湿法冶金金泥品位软测量. 仪器仪表学报, 2018, 39(10): 18−26

    Chang Yu-Qing, Xu Di-Wu. New development of hydrometallurgy. Chinese Journal of Scientific Intrument, 2018, 39(10): 18−26
    [28] 黄礼煌. 金银提取技术. 冶金工业出版社, 2001.

    Huang Li-Huang. Gold and Silver Extraction Technology. Beijing: Metallurgical Industry Press. 2001.
  • 加载中
图(8) / 表(4)
计量
  • 文章访问数:  226
  • HTML全文浏览量:  58
  • PDF下载量:  101
  • 被引次数: 0
出版历程
  • 收稿日期:  2019-01-27
  • 录用日期:  2019-09-09
  • 刊出日期:  2021-10-13

目录

    /

    返回文章
    返回