Prediction of pyrolysis products distribution for end-of-life wind turbine blades using attention-based feature fusion

Zhang Qiliang; Jiang Dekai; Li Chenxu; Zhang Baoxin; Hu Bin; Xu Mingxin; Li Kai; Lu Qiang

doi:null

Your Location：

Home >

Browse articles >

Prediction of pyrolysis products distribution for end-of-life wind turbine blades using attention-based feature fusion

Updated：2026-09-17

- Prediction of pyrolysis products distribution for end-of-life wind turbine blades using attention-based feature fusion
- Chinese Journal of Chemical Engineering Vol. 95, Issue 7, Pages: 1-9(2026)
- Affiliations：
  
  Department of Automation, North China Electric Power University, Baoding 071003, China
  Department of Economic Management, North China Electric Power University, Baoding 071003, China
  National Engineering Research Center of New Energy Power Generation, North China Electric Power University, Beijing 102206, China
- Author bio：
  
  *Corresponding author.E-mail address:qianglu@mail.ustc.edu.cn(Q. Lu).
- Funds：
- DOI：
  CLC：
- Received：10 November 2025，
  
  Revised：2026-01-22，
  
  Accepted：02 February 2026，
  
  Online First：08 April 2026，
  
  Published：2026-07
- Accepted：
Scan QR Code
Zhang Qiliang, Jiang Dekai, Li Chenxu, et al. Prediction of pyrolysis products distribution for end-of-life wind turbine blades using attention-based feature fusion[J]. Chinese Journal of Chemical Engineering, 2026, 95(7): 1-9.
DOI：

Zhang Qiliang, Jiang Dekai, Li Chenxu, et al. Prediction of pyrolysis products distribution for end-of-life wind turbine blades using attention-based feature fusion[J]. Chinese Journal of Chemical Engineering, 2026, 95(7): 1-9. DOI：

摘要

Abstract

Pyrolysis can effectively recover glass fibers from end-of-life wind turbine blades (WTBs) while producing reusable pyrolysis oil and syngas. The nonlinear dynamics involved in the pyrolysis of end-of-life WTBs bring challenges in accurately predicting the product distribution. This paper proposes an attention-based feature fusion prediction framework for pyrolysis product distribution of end-of-life WTBs. Firstly

a physics-informed neural network (PINN) incorporating the Arrhenius equation is proposed. Then

a product distribution prediction model for the fast pyrolysis of end-of-life WTBs is developed by integrating the extreme gradient boosting algorithm

multilayer perceptron

and the PINN

with an attention mechanism introduced to dynamically weight their output. Based on this

a dynamic weighted feature fusion algorithm is proposed to predict the pyrolysis product distribution of end-of-life WTBs. Finally

the superiority of the proposed method is demonstrated through fast pyrolysis experiments of end-of-life WTBs. The experimental results show that the proposed method achieves a notable improvement in prediction accuracy

reducing the root mean square error and mean absolute error by 86% to 99% compared to existing approaches.

关键词

Keywords

references

S.Q. Chen, J.Y. Liu, Z.T. Lin, S. Lin, L.Z. Li, Y. Chen, F. Evrendilek, W.X. Li, W.X. Huang, C.Y. Yang, H.R. Yuan, Y. He, K. Dai, Atmosphere-dependent pyrolytic transformability of glass fiber/epoxy resin composites in waste wind turbine blades, Chem. Eng. J. 505 (2025) 159675.

Z.F. Zhang, Y.Q. Wu, X.D. Ma, Quantum machine learning based wind turbine condition monitoring: state of the art and future prospects, Energy Convers. Manag. 332 (2025) 119694.

Q. Lu, H.W. Ji, Y.Y. Lu, J. Yang, W.W. Chen, J.H. Li, W. Li, M.X. Xu, Transitionmetal-assisted pyrolysis to recover glass fibers from end-of-life wind turbine blades, J. Anal. Appl. Pyrolysis 189 (2025) 107081.

Z.F. Ge, Y.Q. Wu, M.X. Zeng, H.Y. Zhang, A critical review on the structure and recovery technologies of end-of-life wind turbine blades, Energy Fuels 38 (20) (2024) 19393—19413.

Y.X. Zhang, J. Van Caneghem, Y. Pontikes, G. Granata, L. Lessard, A.W. Van Vuure, Envir onmental and economic assessment of mechanical recycling of end-of-life wind turbine blades into rebars and comparison with conventional disposal routes, Compos. Part A Appl. Sci. Manuf. 190 (2025) 108711.

S. Sobek, L. Lombardi, B. Mendecka, H. Mumtaz, M. Sajdak, R. Muzyka, S. Werle, A life cycle assessment of the laboratory: scale oxidative liquefaction as the chemical recycling method of the end-of-life wind turbine blades, J. Environ. Manag. 361 (2024) 121241.

Z.F. Ge, Y.Q. Wu, Z.H. Hou, H.Y. Zhang, Analysis on degradation mechanism and product recycle of ex-service wind turbine blades by hydrothermal liquefaction, ACS Omega 9 (47) (2024) 47031—47041.

K. Kalkanis, C.S. Psomopoulos, S. Kaminaris, G. Ioannidis, P. Pachos, Wind turbine blade composite materials - end of life treatment methods, Energy Proc. 157 (2019) 1136—1143.

H. Jiang, L.C. Ge, H.C. Feng, C.Y. Xu, Q.Y. Yang, X.K. Li, X. Liu, Y. Wang, C. Xu, Thermal transformations during thermal recovery of end-of-life composite carbon fiber beams from wind turbine blades, J. Anal. Appl. Pyrolysis 185 (2025) 106879.

K.Y. Liu, H.G. Fan, L.B. Nie, J. Gu, H.R. Yuan, Kinetics, product distribution and synergistic effects on the co-pyrolysis processes for epoxy resins and balsa wood, J. Anal. Appl. Pyrolysis 187 (2025) 106995.

X.P. Su, Z. Wang, N. Li, L.J. Li, P. Zhang, M. Sun, X.X. Ma, Study on coal pyrolysis characteristics by combining different pyrolysis reactors, Chin. J. Chem. Eng. 76 (2024) 1—9.

J. Zhang, J. Gu, R. Shan, H.R. Yuan, Y. Chen, Advances in thermochemical valorization of biomass towards carbon neutrality, Resour. Conserv. Recycl. 212 (2025) 107905.

Y. Yao, Y. Cao, M.M. Rao, R. Shan, J. Zhang, Fast co-pyrolysis characteristics of polyethylene terephthalate and epoxy resin from waste wind turbine blades, J. Energy Inst. 117 (2024) 101841.

L.F. Li, Z.Y. Luo, L.W. Du, F.T. Miao, L.Y. Liu, Prediction of product yields and heating value of bio-oil from biomass fast pyrolysis: explainable predictive modeling and evaluation, Energy 324 (2025) 136087.

X.Y. Ji, H.Y. Liu, J.T. Chen, X. Zhou, J.B. Li, L. Yang, W.J. Lin, N. Chen, An experimental study into the pyrolysis characteristics of waste tire rubber with catalyst addition, Chin. J. Chem. Eng. 81 (2025) 128—141.

W. Zhang, Y.M. Zhang, H.X. Wu, X.Y. Yang, P. Qiao, J.Z. Li, Z.W. Chen, Y. Wang, Investigation on the pyrolysis behaviors and kinetics of walnut shell lignocellulosic biomass with additives, Chin. J. Chem. Eng. 80 (2025) 303—314.

S. Guo, Y.F. Li, S.W. Tang, T. Zhang, The nitrogen transformation behavior based on the pyrolysis products of wheat straw, Chin. J. Chem. Eng. 71 (2024) 58—65.

Q.C. Yang, D.W. Rong, Q.W. Guo, R.J. Bao, D.W. Zhang, Ensemble learningdriven multi-objective optimization of the co-pyrolysis process of biomass and coal for high economic and environmental performance, Chin. J. Chem. Eng. 84 (2025) 23—34.

E.W. Leng, B. He, J.W. Chen, G.L. Liao, Y.J. Ma, F. Zhang, S. Liu, J.Q. E, Prediction of three-phase product distribution and bio-oil heating value of biomass fast pyrolysis based on machine learning, Energy 236 (2021) 121401.

R.R. Lu, J. Li, X. Zou, A.R. Wang, H.G. Dong, Prediction of product distribution of low-medium rank coal pyrolysis using artificial neural networks model, J. Energy Inst. 107 (2023) 101181.

R. Potnuri, C.S. Rao, D.V. Surya, A. Kumar, T. Basak, Utilizing support vector regression modeling to predict pyro product yields from microwave-assisted catalytic co-pyrolysis of biomass and waste plastics, Energy Convers. Manag. 292 (2023) 117387.

D.R. Cahanap, J. Mohammadpour, S. Jalalifar, H. Mehrjoo, S. Norouzi-Apourvari, F. Salehi, Prediction of three-phase product yield of biomass pyrolysis using artificial intelligence-based models, J. Anal. Appl. Pyrolysis 172 (2023) 106015.

C.C. Ukwuoma, D.S. Cai, O. Bamisile, A.M. Bizi, T.J. Amos, F.L. Delali, D. Thomas, Q. Huang, Hydrogen production prediction from co-gasification of biomass and plastics using attention-gated MLP model, Renew. Energy 249 (2025) 123076.

B. Huang, J.H. Wang, Applications of physics-informed neural networks in power systems - a review, IEEE Trans. Power Syst. 38 (1) (2023) 572—588.

L.K. Chen, Y.F. Wang, W. Sun, L. Shang, X.Z. Dong, B. Wang, Modified hard-constrained PINNs for physically consistent microgrid transient response modeling, Int. J. Electr. Power Energy Syst. 169 (2025) 110726.

R.de O. Teloli, R. Tittarelli, M. Bigot, L. Coelho, E. Ramasso, P. Le Moal, M. Ouisse, A physics-informed neural networks framework for model parameter identification of beam-like structures, Mech. Syst. Signal Process. 224 (2025) 112189.

L.Y. Wang, M. Chen, Z.H. Luo, B.W. Zhang, J. Xu, Z.L. Wang, A.C.C. Tan, Dynamic wake field reconstruction of wind turbine through physics-informed neural network and sparse LiDAR data, Energy 291 (2024) 130401.

K.Q. Wang, L.J. Wang, Q. Meng, C. Yang, Y.S. Lin, J.Y. Zhu, Z.Y. Zhao, C. Zhou, C. H. Zheng, X. Gao, Accurate photovoltaic power prediction via temperature correction with physics-informed neural networks, Energy 328 (2025) 136546.

D. Chen, Y.Z. Li, K. Liu, Y. Li, A physics-informed neural network approach to fatigue life prediction using small quantity of samples, Int. J. Fatig. 166 (2023) 107270.

J. Stiasny, B.S. Zhang, S. Chatzivasileiadis, PINNSim: a simulator for power system dynamics based on Physics-Informed Neural Networks, Elec. Power Syst. Res. 235 (2024) 110796.

A.N. Tian, L.Y. He, T. Ding, K.L. Dong, Y.Q. Wang, J.C. Jiang, A·generic physics-informed neural network framework for lithium-ion batteries state of health estimation, Energy 332 (2025) 137215.

H.Y. Liu, P. Li, Y.G. Zhou, J.S. Zhou, J.Z. Shi, B. Hong, Evolution of characteristic components from retired wind turbine blades in multiple thermal atmospheres: t hermal decomposition characterization and kinetic behavior, J. Environ. Manag. 391 (2025) 126341.

P.R. Yang, B.H. Xie, M.H. Wang, W.J. Guo, X.L. Zhang, X. Chen, W. Chen, Prediction of phenol yield by machine learning based on biomass characteristics, pyrolysis conditions, and catalyst properties, Energy Convers. Manag. 320 (2024) 119001.

H. Wang, F. Zhou, L.L. Zhong, F. Zhong, X.T. Chen, J.X. Guo, Z.Z. Zhang, Screen diffusion battery method for measuring the size distribution of radioactive aerosol employing multi-layer perception (MLP) neural network, J. Aerosol Sci. 189 (2025) 106644.

Views

Downloads

CSCD

Alert me when the article has been cited

Submit

Tools

Publicity Resources

Causal temporal graph attention network for fault diagnosis of chemical processes

Attention-based long short-term memory fully convolutional network for chemical process fault diagnosis

The effect of hydrothermal pretreatment on the structure and fast pyrolysis behaviors of ShengLi lignite

Properties of Bio-oil from Fast Pyrolysis of Rice Husk

A highly robust fluorine-functionalized{Sn2}-organic framework for efficiently catalyzing CO2-Epoxide cycloaddition and Knoevenagel condensation

Related Author

Yiyang Dai

Xu Ji

Qian Li

Heping Jin

Zhehao Jin

Jiaojiao Luo

Li Zhou

Shanwei Xiong

Related Institution

School of Chemical Engineering, Sichuan University

China Three Gorges Corporation

School of Chemical Engineering, Northwest University

International Scientific and Technological Cooperation Base of the Ministry of Science and Technology (MOST) for Clean Utilization of Hydrocarbon Resources

Chemical Engineering Research Center of the Ministry of Education (MOE) for Advanced Use Technology of Shanbei Energy

⁰