Investigation on temperature field reconstruction for the forced ignition process of rocket-based combined cycle engine

Abstract

Accurate and reliable combustion state monitoring is a key requirement for the development of future rocket-based combined cycle (RBCC) engines. The rapid advancements in deep learning technology have rendered data-driven combustion state sensing a possibility, thus contributing to the realization of intelligent and efficient combustion organization. This paper proposes a multi-path convolutional neural network model that is suitable for the reconstruction of two-dimensional temperature fields. The impact of diverse model architectures on the precision of reconstruction outcomes is examined. The results of the reconstruction of the entire test set demonstrate that the MPFC-CNN model exhibits superior accuracy and extrapolation generalization ability compared to the SP-CNN and MPU-CNN models. The overall test dataset demonstrates an average reconstruction error of 2.84%, a linear correlation coefficient of 0.9901, and a structural similarity index of 0.8842. Validation of the reconstruction was conducted for additional combustor temperature fields with varying strut placements. The reconstruction results also basically largely satisfied the requirements. Additionally, the MPFC-CNN model has fewer parameters, which can provide a reliable basis for combustion state recognition and monitoring.

Keywords

Temperature field reconstruction / Deep learning / Convolutional neural network / RBCC engines / Turbulent combustion

References

[1] HeGQ, QinF, WeiXG, et al.. Investigation of several fundamental combustion problems in rocket-based combined-cycle engines[J]. J Exp Fluid Mech, 2016, 31: 1-14

[2] IspirAC, CakirBO, SaracogluBH. Design space investigations of scramjet engines using reduced-order modeling[J]. Acta Astronaut, 2024, 217: 349-362

[3] Kodera M, Ogawa H, Tomioka S, et al (2014) Multi-objective design and trajectory optimization of space transport systems with RBCC propulsion via evolutionary algorithms and pseudospectral methods. No. AIAA 2014-0629

[4] TianY, ShiW, GuoM, et al.. Investigation of combustion characteristics in a hydrogen-fueled scramjet combustor[J]. Acta Astronaut, 2021, 186: 486-495

[5] BarnesFW, SegalC. Cavity-based flameholding for chemically-reacting supersonic flows[J]. Prog Aerosp Sci, 2015, 76: 24-41

[6] ZhuX, CaiZ, WuJ, et al.. Convolutional neural network based combustion mode classification for condition monitoring in the supersonic combustor[J]. Acta Astronaut, 2019, 159: 349-357

[7] UrzayJ. Supersonic combustion in air-breathing propulsion systems for hypersonic flight[J]. Annu Rev Fluid Mech, 2018, 50(1): 593-627

[8] WangZ, CaiZ, SunM, et al.. Large Eddy Simulation of the flame stabilization process in a scramjet combustor with rearwall-expansion cavity[J]. Int J Hydrogen Energy, 2016, 41(42): 19278-19288

[9] KnightD, YanH, PanarasAG, et al.. Advances in CFD prediction of shock wave turbulent boundary layer interactions[J]. Prog Aerosp Sci, 2003, 39(2–3): 121-184

[10] FeiQ, XiangLV, PeijinL. Research status and perspective of rocket based combined cycle propulsion system[J]. J Propulsion Tech, 2010, 31(6): 660-665

[11] HernandezR, BallesterJ. Flame imaging as a diagnostic tool for industrial combustion[J]. Combust Flame, 2008, 155(3): 509-528

[12] HanZ, LiJ, ZhangB, et al.. Prediction of combustion state through a semi-supervised learning model and flame imaging[J]. Fuel, 2021, 289: 119745

[13] HaD, RohTS, HuhH, et al.. Development trend of liquid hydrogen-fueled rocket engines (part 1: performance and operation)[J]. Int J Aeronautical and Space Sci, 2023, 24(1): 131-145

[14] SalgadoMCV, BelderrainMCN, DevezasTC. Space propulsion: A survey study about current and future technologies[J]. J Aerosp Technol Manag, 2018, 10: e1118

[15] LeCunY, BengioY, HintonG. Deep learning. Nature, 2015, 521(7553): 436-444

[16] BalogunH, AlakaH, DemirE, et al.. Artificial intelligence for deconstruction: Current state, challenges, and opportunities[J]. Autom Constr, 2024, 166: 105641

[17] YoungT, HazarikaD, PoriaS, et al.. Recent trends in deep learning based natural language processing[J]. ieee Computational intelligenCe magazine, 2018, 13(3): 55-75

[18] KongC, QuanF, LiY, et al.. Prediction model of temperature field in dual-mode combustors based on wall pressure[J]. Acta Astronaut, 2022, 196: 73-84

[19] SunD, WangZ, QuF, et al.. A deep learning based prediction approach for the supercritical airfoil at transonic speeds. Phys Fluids, 2021, 33(8): 086109

[20] LiK, KouJ, ZhangW. Deep learning for multifidelity aerodynamic distribution modeling from experimental and simulation data[J]. AIAA J, 2022, 60(7): 4413-4427

[21] MaH, ZhangY, HaidnOJ, et al.. Supervised learning mixing characteristics of film cooling in a rocket combustor using convolutional neural networks[J]. Acta Astronaut, 2020, 175: 11-18

[22] XuJ, YuanX, ZhangY, et al.. Real-time prediction of drilling forces inside lunar regolith based on recurrent neural networks[J]. Acta Astronaut, 2022, 201: 259-273

[23] XuZ, ChengZ, TangQD, et al.. An encoder-decoder generative adversarial network-based anomaly detection approach for satellite telemetry data[J]. Acta Astronaut, 2023, 213: 547-558

[24] ZhouJ, CuiG, HuS, et al.. Graph neural networks: A review of methods and applications[J]. AI open, 2020, 1: 57-81

[25] LiuY, XueJ, LiD, et al.. Image recognition based on lightweight convolutional neural network: Recent advances. Image Vision Comput, 2024, 146: 105037

[26] FukamiK, HasegawaK, NakamuraT, et al.. Model order reduction with neural networks: Application to laminar and turbulent flows[J]. SN Computer Science, 2021, 2: 1-16

[27] KongC, ChangJ, WangZ, et al.. Prediction model of flow field in an isolator over various operating conditions[J]. Aerosp Sci Technol, 2021, 111: 106576

[28] LiY, ChangJ, WangZ, et al.. Inversion and reconstruction of supersonic cascade passage flow field based on a model comprising transposed network and residual network[J]. Physics of Fluids, 2019, 31(12): 126102

[29] LiY, ChangJ, WangZ, et al.. An efficient deep learning framework to reconstruct the flow field sequences of the supersonic cascade channel[J]. Physics of Fluids, 2021, 33(5): 056106

[30] ZhaoX, DuL, PengX, et al.. Research on refined reconstruction method of airfoil pressure based on compressed sensing[J]. Theor Appl Mech Lett, 2021, 11(2): 100223

[31] ChenH, GuoM, TianY, et al.. Intelligent reconstruction of the flow field in a supersonic combustor based on deep learning. Phys Fluids, 2022, 34(3): 035128

[32] KongC, ChangJT, LiYF, et al.. Deep learning methods for super-resolution reconstruction of temperature fields in a supersonic combustor. AIP Advances, 2020, 10(11): 115021

[33] HuangJ, LiuH, CaiW. Online in situ prediction of 3-D flame evolution from its history 2-D projections via deep learning[J]. J Fluid Mech, 2019, 875: R2

[34] CaiW, LiuH, HuangJ, et al.. Reconstruction of kHz-rate 3-D flame image sequences from a low-rate 2-D recording via a data-driven approach[J]. JOSA B, 2020, 37(12): 3564-3571

[35] LiB, YaoW, LeeY, et al.. Reconstruction model for heat release rate based on artificial neural network[J]. Int J Hydrogen Energy, 2021, 46(37): 19599-19616

[36] Tamaddon-JahromiHR, ChakshuNK, SazonovI, et al.. Data-driven inverse modelling through neural network (deep learning) and computational heat transfer[J]. Comput Methods Appl Mech Eng, 2020, 369: 113217

[37] XueR, HeG, WeiX, et al.. Experimental study on combustion modes of a liquid kerosene fueled RBCC combustor[J]. Fuel, 2017, 197: 433-444

[38] ZhouZF, LuGY, ChenB. Numerical study on the spray and thermal characteristics of R404A flashing spray using OpenFOAM[J]. Int J Heat Mass Transf, 2018, 117: 1312-1321

[39] Ghodke C, Choi J, Srinivasan S, et al (2011) Large eddy simulation of supersonic combustion in a cavity-strut flameholder. No. AIAA 2011-323

[40] Ajmani K, Kundu K, Penko P. A study on detonation of jet-a using a reduced mechanism. No. AIAA 2010-1515

[41] ZhaoM, ZhouT, YeT, et al.. Large eddy simulation of reacting flow in a hydrogen jet into supersonic cross-flow combustor with an inlet compression ramp[J]. Int J Hydrogen Energy, 2017, 42(26): 16782-16792

[42] LiuB, LiY, ZhuS, et al.. Numerical investigation on flame-stabilization mechanism of supersonic-combustor-based high-temperature CO/H2 jet[J]. Int J Hydrogen Energy, 2024, 49: 1192-1204

[43] Ronneberger O, Fischer P, Brox T (2015) U-net: Convolutional networks for biomedical image segmentation. In: Navab N, Hornegger J, Wells W, et al (eds) Medical image computing and computer-assisted intervention – MICCAI 2015, Munich. Lecture notes in computer science. Springer, Cham 9351:234–241

[44] ShanQ, LiZ, JiaJ, et al.. Fast image/video upsampling[J]. ACM Transactions on Graphics (TOG), 2008, 27(5): 1-7

[45] KrizhevskyA, SutskeverI, HintonGE. ImageNet classification with deep convolutional neural networks. Communic ACM, 2017, 60(6): 84-90

Funding

National Natural Science Foundation of China(52202482)

Young Elite Scientists Sponsorship Program of CAST(YESS20220405)

RIGHTS & PERMISSIONS

The Author(s)