基于注意力增强梯度的径向基函数神经网络颗粒态有机氮预测研究

宋贞海; 王功明; 赵赫楠; 曹亚新

doi:10.11949/0438-1157.20260881

您当前的位置：

首页 >

文章列表页 >

基于注意力增强梯度的径向基函数神经网络颗粒态有机氮预测研究

智能过程工程 | 更新时间：2026-07-20

- 基于注意力增强梯度的径向基函数神经网络颗粒态有机氮预测研究
- Research on particle organic nitrogen prediction using radial basis function neural network with enhanced attention gradient
- 化工学报 2026年
- 作者机构：
  
  1.曲阜师范大学继续教育学院，山东曲阜 273165
  2.北京工业大学信息科学技术学院，北京 100124
  3.曲阜师范大学图书馆，山东曲阜 273165
- 作者简介：
  
  宋贞海(1975—)，男，硕士，副教授，css_tsg@126.com
  王功明(1987—)，男，博士，教授，wanggm@bjut.edu.cn
- 基金信息：
  
  国家自然科学基金面上项目(62373018);中国博士后科学基金(2025T180479)
- DOI：10.11949/0438-1157.20260881
  中图分类号： TP 183
- 收稿：2026-06-25，
  
  修回：2026-07-20，
  
  录用：2026-07-20，
- 稿件说明：
移动端阅览
宋贞海, 王功明, 赵赫楠, 等. 基于注意力增强梯度的径向基函数神经网络颗粒态有机氮预测研究[J/OL]. 化工学报, 2026.

SONG Zhenhai, WANG Gongming, ZHAO Henan, et al. Research on particle organic nitrogen prediction using radial basis function neural network with enhanced attention gradient[J/OL]. CIESC Journal, 2026.
宋贞海, 王功明, 赵赫楠, 等. 基于注意力增强梯度的径向基函数神经网络颗粒态有机氮预测研究[J/OL]. 化工学报, 2026. DOI： 10.11949/0438-1157.20260881.

SONG Zhenhai, WANG Gongming, ZHAO Henan, et al. Research on particle organic nitrogen prediction using radial basis function neural network with enhanced attention gradient[J/OL]. CIESC Journal, 2026. DOI： 10.11949/0438-1157.20260881.

摘要

针对水环境颗粒态有机氮浓度预测问题，本文提出了一种基于双尺度注意力机制的自适应径向基函数神经网络(Adaptive radial basis function neural network with dual-scale attention

ARBFNN-DSA)；首先设计了一种具备输入特征注意力机制的数据驱动型RBFNN来提高有效特征的提取能力；其次将注意力机制的差异化权重与核函数的中心和宽度相关联，使核函数具备较好的自适应能力，从而实现对输入数据分布和特征重要性的自适应定量评估；同时利用互信息度量方法分析ARBFNN-DSA输入输出之间的敏感性，从而量化关联变量对PON状态演化的因果贡献程度；最后将ARBFNN-DSA模型应用于山东省青岛地区的大沽河流域PON浓度的软测量，结果显示ARBFNN-DSA模型不仅实现了对PON动态特性的精确建模与预测，还能为潜在的污染溯源提供诱因变量的定量分析。相较于对比模型SODBN-AAL和EDBN，ARBFNN-DSA在预测精度和泛化性能上分别提高了16.98%和13.39%以上。

Abstract

For the problem of predicting the particle organic nitrogen (PON) concentration in water environment

this paper proposes an adaptive radial basis function neural network with dual-scale attention (ARBFNN-DSA); Firstly

a data-driven RBFNN with an input feature attention mechanism is designed to enhance the extraction ability of effective features; Secondly

the differentiated weights of the attention mechanism are associated with the center and width of the kernel function

enabling the kernel function to have better adaptive ability

thereby achieving the adaptive quantitative evaluation of the input data distribution and feature importance; At the same time

the mutual information measurement method is used to analyze the sensitivity between the input and output of the ARBFNN-DSA

thereby quantifying the causal contribution degree of the associated variables to the evolution of the PON state; Finally

the ARBFNN-DSA model is applied to the soft measurement of PON concentration in the Dagu River Basin of Shandong province. The results show that the ARBFNN-DSA model not only achieves accurate modeling and prediction of PON dynamics

but also provides quantitative analysis of the potential pollution source variables. Compared with the comparative models SODBN-AAL and EDBN

ARBFNN-DSA has improved the predictive accuracy and generalization performance by 16.98% and 13.39%

respectively.

关键词

Keywords

references

Wang G M , Chen H , Han H G , et al . Predicting water quality with nonstationarity: event-triggered deep fuzzy neural network [J ] . IEEE Transactions on Fuzzy Systems , 2024 , 32 ( 5 ): 2690 - 2699 .

李欣怡 , 王功明 , 王自鹏 , 等 . 基于事件触发的污水处理过程水质智能预测研究 [J ] . 化工学报 , 2025 , 76 ( 6 ): 2828 - 2837 .

Li X Y , Wang G M , Wang Z P , et al . Research on intelligent prediction of water quality in sewage treatment process based on event triggering [J ] . CIESC Journal , 2025 , 76 ( 6 ): 2828 - 2837 .

Alfa U , Tijjani A S , Fabresse L , et al . AI-driven decision support for spatio-temporal water quality monitoring using aquatic drones [J ] . IFAC-PapersOnLine , 2025 , 59 ( 33 ): 25 - 31 .

王功明 , 李文静 , 乔俊飞 . 基于PLSR自适应深度信念网络的出水总磷预测 [J ] . 化工学报 , 2017 , 68 ( 5 ): 1987 - 1997 .

Wang G M , Li W J , Qiao J F . Prediction of effluent total phosphorus using PLSR-based adaptive deep belief network [J ] . CIESC Journal , 2017 , 68 ( 5 ): 1987 - 1997

Wang G M , Jia Q S , Zhou M C , et al . Soft-sensing of wastewater treatment process via deep belief network with event-triggered learning [J ] . Neurocomputing , 2021 , 436 : 103 - 113 .

Bi J , Lin Y Z , Dong Q X , et al . Large-scale water quality prediction with integrated deep neural network [J ] . Information Sciences , 2021 , 571 : 191 - 205 .

Bi J , Wang Z Q , Yuan H T , et al . Long-term water quality prediction with transformer-based spatial-temporal graph fusion [J ] . IEEE Transactions on Automation Science and Engineering , 2025 , 22 : 11392 - 11404 .

Chang P , Shi S Q . Soft-sensor of the key water quality indicators in the wastewater treatment process based on improved mamba [J ] . IEEE Transactions on Instrumentation and Measurement , 2025 , 74 : 2548809 .

Chang P , Zhang S R , Wang Z C . Soft sensor of the key effluent index in the municipal wastewater treatment process based on transformer [J ] . IEEE Transactions on Industrial Informatics , 2024 , 20 ( 3 ): 4021 - 4028 .

Liu Y , Liang Y X , Ouyang K , et al . Predicting urban water quality with ubiquitous data - a data-driven approach [J ] . IEEE Transactions on Big Data , 2022 , 8 ( 2 ): 564 - 578 .

Zhou P , Wang X , Chai T Y . Multiobjective operation optimization of wastewater treatment process based on reinforcement self-learning and knowledge guidance [J ] . IEEE Transactions on Cybernetics , 2023 , 53 ( 11 ): 6896 - 6909 .

Wang G M , Qiao J F . An efficient self-organizing deep fuzzy neural network for nonlinear system modeling [J ] . IEEE Transactions on Fuzzy Systems , 2022 , 30 ( 7 ): 2170 - 2182 .

Wang G M , Yuan G H , Hu Z Q , et al . Complexity-based structural optimization of deep belief network and application in wastewater treatment process [J ] . IEEE Transactions on Industrial Informatics , 2024 , 20 ( 4 ): 6974 - 6982 .

Zhuang K Z , Luan X W , Jiang X T , et al . A self-optimizing deep belief network with adaptive-active learning: dynamic optimization for neural network [J ] . IEEE Systems, Man, and Cybernetics Magazine , 2025 , 11 ( 2 ): 75 - 83 .

Cao Y X , Zhuang K Z , Wang G M , et al . Using radial basis function neural network with weight memory mechanism: nonlinear system identification [J ] . IEEE Systems, Man, and Cybernetics Magazine , 2025 , 11 ( 4 ): 33 - 38 .

Wang L J , Feng M H , Wang G M , et al . An efficient radial basis function neural network with hybrid regularizations for predicting library loan volume: generalization and adaptability are improved [J ] . IEEE Systems, Man, and Cybernetics Magazine , 2026 , 12 ( 1 ): 112 - 117 .

Tan X B , Bai Y L , Yue X X , et al . A Mamba-based method for multi-feature water quality prediction fusing dual denoising and attention enhancement [J ] . Journal of Hydrology , 2025 , 660 : 133424 .

Yang H Y , Ding Y F , Wu X L , et al . An identification model of sludge bulking based on self-organized recurrent fuzzy neural network [J ] . IEEE Transactions on Industrial Informatics , 2025 , 21 ( 1 ): 357 - 365 .

Wang G M , Jia Q S , Qiao J F , et al . A sparse deep belief network with efficient fuzzy learning framework [J ] . Neural Networks , 2020 , 121 : 430 - 440 .

Wang G M , Chen H , Jiang S L , et al . Neurodynamics-driven prediction model for state evolution of coastal water quality [J ] . IEEE Transactions on Instrumentation and Measurement , 2024 , 73 : 2519409 .

Yang C L , Yang S , Tang J , et al . Evolving deep delay echo state network for effluent NH4-N prediction in wastewater treatment plants [J ] . IEEE Transactions on Instrumentation and Measurement , 2023 , 72 : 2513812 .

Li Y B , Wang Z Q , Yuan H T , et al . Hybrid water quality prediction based on attention combined with frequency enhancement and multi-seasonal decomposition [J ] . Journal of Water Process Engineering , 2025 , 78 : 108747 .

Bi J , Chen Z X , Yuan H T , et al . Accurate water quality prediction with attention-based bidirectional LSTM and encoder–decoder [J ] . Expert Systems with Applications . 2024 , 238 : 121807 .

Qiao J F , Lin Y Z , Bi J , et al . Attention-based spatiotemporal graph fusion convolution networks for water quality prediction [J ] . IEEE Transactions on Automation Science and Engineering , 2025 , 22 : 1 - 10 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

水土环境中微生物胞外聚合物对污染物迁移和归宿影响的研究进展

基于RBF神经网络的油藏相对渗透率曲线计算

基于独立分量分析和RBF神经网络的气液两相流流型识别

原位水热法可控合成Bi2MoO6/MoO3复合纤维及其高效光催化产氢性能

基于机器学习的改型歧管式微通道热沉流动传热特性多目标优化