引用本文:孙凯,隋璘,张芳芳,杨根科.基于非负绞杀与长短期记忆神经网络的动态软测量算法[J].控制理论与应用,2023,40(1):83~93.[点击复制]
SUN Kai,SUI Lin,ZHANG Fang-fang,YANG Gen-ke.Dynamic soft sensor algorithm based on nonnegative garrote and long short-term memory neural network[J].Control Theory and Technology,2023,40(1):83~93.[点击复制]
基于非负绞杀与长短期记忆神经网络的动态软测量算法
Dynamic soft sensor algorithm based on nonnegative garrote and long short-term memory neural network
摘要点击 965  全文点击 367  投稿时间：2021-06-20  修订日期：2022-04-21
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/CTA.2021.10529
  2023,40(1):83-93
中文关键词  神经网络  软测量  长短期记忆  动态建模  变量选择  模型简化
英文关键词  neural networks  soft sensor  long short-term memory  dynamic modeling  variable selection  model reduction
基金项目  山东省自然科学基金项目(ZR2021MF022), 国家重点研发计划项目(2019YFB1705702, 2020YFB1711204), 山东省重点研发计划项目(公益类专 项)(2019GGX104037)资助.
作者单位E-mail
孙凯 齐鲁工业大学(山东省科学院） sunkai79@qlu.edu.cn 
隋璘 齐鲁工业大学(山东省科学院）  
张芳芳 齐鲁工业大学(山东省科学院）  
杨根科* 上海交通大学 gkyang@sjtu.edu.cn 
中文摘要
      现代工业过程建模中, 生产过程的多变量、非线性及动态性会导致模型复杂度增高且建模精度降低. 针对 这一问题, 将非负绞杀算法(NNG)嵌入长短期记忆(LSTM)神经网络, 提出一种基于LSTM神经网络及其输入变量选 择的动态软测量算法. 首先, 通过参数优化生成训练好的LSTM神经网络, 利用其出色的历史信息记忆能力处理工 业过程中的动态、时滞等问题; 其次, 采用NNG算法对LSTM网络输入权重进行压缩, 剔除冗余变量, 提高模型精度, 并采用网格搜索法与分块交叉验证对其超参数寻优; 最后, 将算法应用于某火电厂脱硫过程排放烟气SO2浓度软测 量建模, 并与其它先进算法进行性能比较. 实验结果表明所提算法能有效剔除冗余变量, 降低模型复杂度并提高其 预测性能.
英文摘要
      In modern industrial process modeling, the multivariable, nonlinearity and dynamics of the production process increase the model complexity and reduce the model accuracy. In response to this problem, a dynamic soft-sensing algorithm based on the long short-term memory (LSTM) neural network and its input variable selection is proposed by embedding the nonnegative garrote (NNG) into the LSTM neural network. First, a well-trained LSTM neural network is generated with parameter optimization, in which the dynamics and time-delay of industrial processes are handled by its excellent memory capacity of historical information. Then, the NNG algorithm is used to compress the input weights of the LSTM network to eliminate the redundant variables and improve the model accuracy. Grid search and blocked crossvalidation are used to find the optimal hyperparameter of the algorithm. Finally, the algorithm is applied to the soft-sensing modeling of SO2 concentration in the flue gas that is discharged from the desulfurization process of a thermal power plant, and the performance of the algorithm is compared with other state-of-the-art algorithms. Experimental results demonstrate that the proposed algorithm can effectively delete the redundant variables, reduce the model complexity and improve the prediction performance of the model.