基于PINN的含低浓度化学物种燃烧正逆问题建模方法

张伟娜; 李港扬; 崔弘杨; 毕忠勤; 戴丹

doi:10.19805/j.cnki.jcspe.2026.250709

欢迎来到能源电力期刊网！

English Version 期刊登录

您当前的位置：

首页 >

文章列表页 >

基于PINN的含低浓度化学物种燃烧正逆问题建模方法

更新时间：2026-10-06

- 基于PINN的含低浓度化学物种燃烧正逆问题建模方法
- PINN-Based Modeling Approach for Combustion of Low-concentration Chemical Species Forward and Inverse Problems
- 动力工程学报 2026年46卷第3期页码：95-105
- 作者机构：
  
  1. 上海电力大学计算机科学与技术学院,上海,201306
  2. 中国联合重型燃气轮机技术有限公司,北京,100016
  3. 阿斯顿大学计算机科学与数字技术学院,伯明翰,英国,B4 7ET
- 作者简介：
  
  张伟娜(1983—),女,河北栾城人,博士,讲师,物理信息神经网络及其应用、大模型推理、综合能源调度等,E-mail:mszhangwn@shiep.edu.cn
- 基金信息：
- DOI：10.19805/j.cnki.jcspe.2026.250709
  中图分类号：
- 网络首发：2026-03-10，
  
  纸质出版：2026-03-10
- 稿件说明：
移动端阅览
张伟娜,李港扬,崔弘杨,毕忠勤,戴丹. 基于PINN的含低浓度化学物种燃烧正逆问题建模方法动力工程学报, 2026, 46(3): 95-105 https://doi.

org/10.19805/j.cnki.jcspe.2026.250709
张伟娜,李港扬,崔弘杨,毕忠勤,戴丹. 基于PINN的含低浓度化学物种燃烧正逆问题建模方法动力工程学报, 2026, 46(3): 95-105 https://doi. DOI： 10.19805/j.cnki.jcspe.2026.250709.

org/10.19805/j.cnki.jcspe.2026.250709 DOI：

摘要

随着燃气轮机向更高效、更清洁方向发展

基于高保真模型的燃烧室设计需求日益迫切。提出基于变分自编码器火焰(variational auto-encoder flame

VAEF)的物理信息神经网络(physics-informed neural networks

PINN)对含低浓度化学物种的层流预混火焰正逆问题进行建模:通过基于自适应上下界或观测数据的约束映射;通过平滑窗划分计算域与局部加权;在稀疏观测条件下的逆问题中将变分自编码器重构误差与相对熵项加入观测条件损失计算中。结果表明:以上措施显著提升了训练稳定性与低浓度物种预测精度

PINN模型的决定系数均大于0.95

相对误差绝对值均小于15%

展现良好的全局拟合精度和参数反演能力

为复杂化学机理下的燃烧模拟与排放监测提供了鲁棒性更好的PINN方案。

Abstract

As gas turbines progress towards higher efficiency and cleaner performance

the demand for combustion chamber designs grounded in high-fidelity models is becoming increasingly pressing. To meet this demand

a physics-informed neural network (PINN) based on variational auto-encoder flame (VAEF) was proposed for modeling both the forward and inverse problems of laminar premixed flames containing low-concentration chemical species. Following strategies were employed in this method: constrained mapping was implemented based on adaptive upper and lower boundaries or observational data; the computational domain was partitioned using smoothing windows with regional weighting; for inverse problems under sparse observation conditions

the variational auto-encoder reconstruction error and the Kullback-Leibler (KL) divergence term were incorporated into the calculation of the observation condition loss. Results indicate that these measures substantially enhance training stability and prediction accuracy for low-concentration species. A coefficient of determination greater than 0.95 and an relative error below 15% are achieved by PINN model

demonstrating excellent global fitting precision and parameter inversion capability. A PINN solution with better robust performance is thus provided for combustion simulation and emission monitoring under complex chemical mechanisms.

关键词

Keywords

references

Energy Institute. Statistical Review of World Energy 2025[R]. London: Energy Institute, 2025. https://www.energyinst.org/statistical-review.

AGGARWAL S K. Extinction of laminar partially premixed flames[J]. Progress in Energy and Combustion Science, 2009, 35(6): 528-570.

TSUJI H. Counterflow diffusion flames[J]. Progress in Energy and Combustion Science, 1982, 8(2): 93-119.

RAVIKRISHNA R V, SAHU A B. Advances in understanding combustion phenomena using non-premixed and partially premixed counterflow flames: a review[J]. International Journal of Spray and Combustion Dynamics, 2018, 10(1): 38-71.

MITTAL V, PITSCH H, EGOLFOPOULOS F. Assessment of counterflow to measure laminar bur-ning velocities using direct numerical simulations[J]. Combustion Theory and Modelling, 2012, 16(3): 419-433.

NATARAJAN J, LIEUWEN T, SEITZMAN J. Laminar flame speeds of H 2 /CO mixtures: effect of CO 2 dilution, preheat temperature, and pressure[J ] . Combustion and Flame, 2007, 151(1/2): 104-119.

ZHANG Guihua, WU Yuxin, WU Jiahao, et al. State of the art and challenges of flamelet method in gas turbine combustor simulation[J]. Journal of Tsinghua University (Science and Technology), 2023, 63(4): 505-520.

DENT M R, ROSE J J, TEJERO J, et al. Carbon monoxide poisoning: from microbes to therapeutics[J]. Annual Review of Medicine, 2024, 75(1): 337-351.

LI Lei, CUI Chunhua. Bulk water redox chemistry enables radical-mediated C-C coupling in CO 2 electroreduction[J ] . Nature Chemistry, 2025. https://www.nature.com/articles/s41557-025-01948-z.

刘法志, 张鹏威, 刘涛, 等. Sb改性钒钛SCR脱硝催化剂抗CO中毒性能[J]. 化工进展, 2025, 44(2): 1129-1137.LIU Fazhi, ZHANG Pengwei, LIU Tao, et al. Mechanism of anti-CO poisoning of Sb-modified vanadium-titanium SCR denitrification catalysts[J]. Chemical Industry and Engineering Progress, 2025, 44(2): 1129-1137.

WAHEED F, EHSAN N, NASIR R, et al. Geo-spatial distribution of air pollutants in urban area and its potential health risk analysis solutions[J]. Urban Climate, 2025, 61: 102380.

LIEU K, SMOLLIN C G, LESAINT K T. What is carbon monoxide poisoning?[J]. JAMA, 2025, 333(12): 1098.

HAWORTH D C. Progress in probability density function methods for turbulent reacting flows[J]. Progress in Energy and Combustion Science, 2010, 36(2): 168-259.

JI Weiqi, DENG Sili. Autonomous discovery of unknown reaction pathways from data by chemical reaction neural network[J]. The Journal of Physical Chemistry A, 2021, 125(4): 1082-1092.

RAISSI M, PERDIKARIS P, KARNIADAKIS G E, et al. Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations[J]. Journal of Computational Physics, 2019, 378: 686-707.

D'ALESSIO G, PARENTE A, STAGNI A, et al. Adaptive chemistry via pre-partitioning of compo-sition space and mechanism reduction[J]. Combu-stion and Flame, 2020, 211: 68-82.

ZHANG Tianhan, YI Yuxiao, XU Yifan, et al. A multi-scale sampling method for accurate and robust deep neural network to predict combustion chemical kinetics[J]. Combustion and Flame, 2022, 245: 112319.

NIU Zhitian, QI Hong, SUN Antai, et al. Efficient and robust CNN-LSTM prediction of flame temperature aided light field online tomography[J]. Science China Technological Sciences, 2024, 67(1): 271-284.

KARNIADAKIS G E, KEVREKIDIS I G, LU Lu, et al. Physics-informed machine learning[J]. Nature Reviews Physics, 2021, 3(6): 422-440.

OWOYELE O, PAL P. ChemNODE: a neural ordinary differential equations framework for efficient chemical kinetic solvers[J]. Energy and AI, 2022, 7: 100118.

LIU Kai, LUO Kun, CHENG Yuzhou, et al. Surrogate modeling of parameterized multi-dimensional premixed combustion with physics-informed neural networks for rapid exploration of design space[J]. Combustion and Flame, 2023, 258: 113094.

WANG Xutun, WEN Haocheng, HU Tong, et al. Flow-field reconstruction in rotating detonation combustor based on physics-informed neural network[J]. Physics of Fluids, 2023, 35(7): 076109.

WU Jiahao, ZHANG Su, WU Yuxin, et al. Flame PINN-1D: physics-informed neural networks to solve forward and inverse problems of 1D laminar flames[J]. Combustion and Flame, 2025, 273: 113964.

赵暾, 周宇, 程艳青, 等. 基于内嵌物理机理神经网络的热传导方程的正问题及逆问题求解[J]. 空气动力学学报, 2021, 39(5): 19-26.ZHAO Tun, ZHOU Yu, CHENG Yanqing, et al. Solving forward and inverse problems of the heat conduction equation using physics-informed neural networks[J]. Acta Aerodynamica Sinica, 2021, 39(5): 19-26.

张振, 苏欣荣, 袁新. 基于物理信息神经网络的气膜冷却湍流模型反演学习[J]. 动力工程学报, 2024, 44(9): 1459-1465.ZHANG Zhen, SU Xinrong, YUAN Xin. PINN-based inversion learning of turbulence model for film cooling[J]. Journal of Chinese Society of Power Engineering, 2024, 44(9): 1459-1465.

KEE R J, COLTRIN M E, GLARBORG P. Chemically reacting flow: theory and practice[M]. Hoboken: John Wiley & Sons, 2005.

GOODWIN D G, MOFFAT H K, SPETH R L, et al. Cantera: an object-oriented software toolkit for chemical kinetics, thermodynamics, and transport processes[J]. Zenodo, 2018. https://doi.org/10.5281/zenodo.14455267.

邓志平, 任少君, 翁琪航, 等. 基于物理信息神经网络的生物质气化产物分布预测方法[J]. 动力工程学报, 2024, 44(5): 719-726.DENG Zhiping, REN Shaojun, WENG Qihang, et al. Prediction method for biomass gasification product distribution based on physics-informed neural network[J]. Journal of Chinese Society of Power Engineering, 2024, 44(5): 719-726.

LU Lu, MENG Xuhui, MAO Zhiping, et al. Deep XDE: a deep learning library for solving differential equations[J]. SIAM Review, 2021, 63(1): 208-228.

BAYDIN A G, PEARLMUTTER B A, RADUL A A, et al. Automatic differentiation in machine learning: a survey[J]. Journal of Machine Learning Research, 2018, 18(153): 1-43.

王意存, 邢江宽, 罗坤, 等. 基于物理信息神经网络的燃烧化学微分方程求解[J]. 浙江大学学报(工学版), 2022, 56(10): 2084-2092.WANG Yicun, XING Jiangkuan, LUO Kun, et al. Solving combustion chemical differential equations via physics-informed neural network[J]. Journal of Zhejiang University (Engineering Science), 2022, 56(10): 2084-2092.

KINGMA D P, WELLING M. Auto-encoding variational Bayes[C] //Proceedings of the 2nd Inter-national Conference on Learning Represen-tations. Banff: ICLR, 2014.

ZHONG W H, MEIDANI H. PI-VAE: physics-informed variational auto-encoder for stochastic differential equations[J]. Computer Methods in Applied Mechanics and Engineering, 2023, 403: 115664.

FRANZELLI B, RIBER E, GICQUEL L Y M, et al. Large eddy simulation of combustion instabilities in a lean partially premixed swirled flame[J]. Combustion and Flame, 2012, 159(2): 621-637.

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

基于WOA的循环流化床锅炉SO2排放浓度VMD-Transformer预测模型

芦竹与典型煤协同燃烧过程中灰成分与结渣特性研究

基于Leckner模型的氨掺烧高水蒸气浓度烟气对燃煤锅炉传热特性的影响分析

考虑生物质混燃碳捕集与低碳激励机制的虚拟电厂优化调度

计入工业园区电热负荷需求的源网荷储优化调度方法研究