《计算机应用》唯一官方网站 ›› 2022, Vol. 42 ›› Issue (3): 791-796.DOI: 10.11772/j.issn.1001-9081.2021040787

• 2021年中国计算机学会人工智能会议(CCFAI 2021) • 上一篇    

基于集合经验模态分解和长短期记忆网络的催化裂化装置氮氧化物排放预测

陈冲1(), 闫珠1, 赵继轩1, 何为1,2, 梁华庆1   

  1. 1.中国石油大学(北京) 信息科学与工程学院, 北京 102249
    2.中国石油集团安全环保技术研究院有限公司 HSE检测中心, 北京 102206
  • 收稿日期:2021-05-17 修回日期:2021-06-26 接受日期:2021-06-29 发布日期:2021-11-09 出版日期:2022-03-10
  • 通讯作者: 陈冲
  • 作者简介:闫珠(1996—),女,山东菏泽人,硕士研究生,主要研究方向:深度学习、信号处理
    赵继轩(1997—),男,天津人,硕士研究生,主要研究方向:深度学习
    何为(1989—),男,陕西西安人,工程师,硕士,主要研究方向:污染在线监测、人工智能
    梁华庆(1964—),女,广东吴川人,教授,博士,主要研究方向:石油行业的弱信号及测控仪器。
  • 基金资助:
    中国石油天然气集团有限公司直属院所基础科学研究和战略储备技术研究基金资助项目(2017D-5008);中国石油大学(北京)科研基金资助项目(2462020YXZZ025)

Prediction of NOx emission from fluid catalytic cracking unit based on ensemble empirical mode decomposition and long short-term memory network

Chong CHEN1(), Zhu YAN1, Jixuan ZHAO1, Wei HE1,2, Huaqing LIANG1   

  1. 1.College of Information Science and Engineering,China University of Petroleum-Beijing,Beijing 102249,China
    2.HSE Testing Center,Safety and Environmental Protection Technology Research Institute of CNPC,Beijing 102206,China
  • Received:2021-05-17 Revised:2021-06-26 Accepted:2021-06-29 Online:2021-11-09 Published:2022-03-10
  • Contact: Chong CHEN
  • About author:YAN Zhu, born in 1996, M. S. candidate. Her research interests include deep learning, signal processing.
    ZHAO Jixuan, born in 1997, M. S. candidate. His research interest include deep learning.
    HE Wei, born in 1989, M. S., engineer. His research interests include online monitoring of pollution, artificial intelligence.
    LIANG Huaqing, born in 1964, Ph. D., professor. Her research interests include detection of weak signal and measurement and control instrument in petroleum.
  • Supported by:
    CNPC Foundation for Basic Research and Strategic Reserved Technology(2017D-5008);Science Foundation of China University of Petroleum (Beijing)(2462020YXZZ025)

摘要:

氮氧化物(NOx)是催化裂化(FCC)装置再生烟气中的主要污染物之一,准确预测NOx的排放浓度可有效避免炼化企业污染事件的发生。鉴于污染物排放数据具有非平稳、非线性和长记忆等特性,为了提高污染物排放浓度预测精度,提出一种基于集合经验模态分解(EEMD)和长短期记忆网络(LSTM)的耦合模型。将NOx排放浓度数据经过EEMD为若干个固有模态函数(IMF)和一个残差序列;根据IMF子序列与原始数据之间的相关性分析,剔除极弱相关的信号分量,有效减小原信号数据中的噪声;将IMF序列集分为高、低频两部分,分别进入不同深度的LSTM网络;最终,将子序列的预测结果反变换得到NOx排放浓度。实验表明,在催化裂化装置NOx排放预测中,对比LSTM的表现,EEMD-LSTM耦合模型在均方误差(MSE)、平均绝对误差(MAE)分别减小了46.7%、45.9%;在决定系数R2上增大了43%,实现了更高的预测精度。

关键词: 催化裂化, 污染物排放预测, 集合经验模态分解, 长短期记忆网络

Abstract:

Nitrogen oxide (NOx) is one of the main pollutants in the regenerated flue gas of Fluid Catalytic Cracking (FCC) unit. Accurate prediction of NOx emission can effectively avoid the occurrence of pollution events in refinery enterprises. Because of the non-stationarity, nonlinearity and long-memory characteristics of pollutant emission data, a new hybrid model incorporating Ensemble Empirical Mode Decomposition (EEMD) and Long Short-Term Memory network (LSTM) was proposed to improve the prediction accuracy of pollutant emission concentration. The NOx emission concentration data was first decomposed into several Intrinsic Mode Functions (IMFs) and a residual by using the EEMD model. According to the correlation analysis between the IMF sub-sequences and the original data, the IMF sub-sequences with low correlation were eliminated, which could effectively reduce the noise in the original data. The IMFs could be divided into high and low frequency sequences, which were respectively trained in the LSTM networks with different depths. The final NOx concentration prediction results were reconstructed by the predicted results of each sub-sequences. Compared with the performance of LSTM in the NOx emission prediction of FCC unit, the Mean Square Error (MSE), Mean Absolute Error (MAE) were reduced by 46.7%, 45.9%,and determination coefficient (R2) of EEMD-LSTM was improved by 43% respectively, which means the proposed model achieves higher prediction accuracy.

Key words: fluid catalytic cracking, pollutant emission prediction, Ensemble Empirical Mode Decomposition (EEMD), Long Short-Term Memory (LSTM) network

中图分类号: