A new maximum-a-posteriori-based gappy method for physical field reconstruction using proper orthogonal decomposition and autoencoder

doi:10.1007/s10483-025-3295-6

摘要/Abstract

Abstract:

A novel gappy technology, gappy autoencoder with proper orthogonal decomposition (Gappy POD-AE), is proposed for reconstructing physical fields from sparse data. High-dimensional data are reduced via proper orthogonal decomposition (POD), and low-dimensional data are used to train an autoencoder (AE). By integrating the POD operator with the decoder, a nonlinear solution form is established and incorporated into a new maximum-a-posteriori (MAP)-based objective for online reconstruction. The numerical results on the two-dimensional (2D) Bhatnagar-Gross-Krook-Boltzmann (BGK-Boltzmann) equation, wave equation, shallow-water equation, and satellite data show that Gappy POD-AE achieves higher accuracy than gappy proper orthogonal decomposition (Gappy POD), especially for the data with slowly decaying singular values, and is more efficient in training than gappy autoencoder (Gappy AE). The MAP-based formulation and new gappy procedure further enhance the reconstruction accuracy.

Key words: data reconstruction, gappy technology, proper orthogonal decomposition (POD), autoencoder (AE), maximum-a-posteriori (MAP)

中图分类号:

O357.1

. [J]. Applied Mathematics and Mechanics (English Edition), 2025, 46(9): 1729-1752.

Wenwei JIANG, Chenhao TAN, Yuntao ZHOU, Kai YANG, Xiaowei GAO. A new maximum-a-posteriori-based gappy method for physical field reconstruction using proper orthogonal decomposition and autoencoder[J]. Applied Mathematics and Mechanics (English Edition), 2025, 46(9): 1729-1752.

图/表 41

参考文献 32

[33]	YAMADA, K., SAITO, Y., NANKAI, K., NONOMURA, T., ASAI, K., and TSUBAKINO, D. Fast greedy optimization of sensor selection in measurement with correlated noise. Mechanical Systems and Signal Processing, 158, 107619 (2021)
[34]	NAGATA, T., YAMADA, K., NONOMURA, T., NAKAI, K., SAITO, Y., and ONO, S. Data-driven sensor selection method based on proximal optimization for high-dimensional data with correlated measurement noise. IEEE Transactions on Signal Processing, 70, 5251–5264 (2022)
[35]	ZHANG, X., JI, T., XIE, F., ZHENG, H., and ZHENG, Y. Unsteady flow prediction from sparse measurements by compressed sensing reduced order modeling. Computational Methods in Applied Mechanics and Engineering, 393, 114800 (2022)
[36]	WU, J., XIAO, D., and LUO, M. Deep-learning assisted reduced order model for high-dimensional flow prediction from sparse data. Physics of Fluids, 35, 103115 (2023)
[37]	OHLBERGER, M. and RAVE, S. Reduced basis methods: success, limitations and future challenges. Proceedings of The Conference Algoritmy, Vysoke Tatry, Slovakia (2016)
[38]	GREIF, C. and URBAN, K. Decay of the Kolmogorov N-width for wave problems. Applied Mathematics Letters, 96, 216–222 (2019)
[39]	KIM, Y., CHOI, Y., and YOO, B. Gappy AE: a nonlinear approach for Gappy data reconstruction using auto-encoder. Computational Methods in Applied Mechanics and Engineering, 426, 116978 (2024)
[40]	JIANG, W., PAN, T., JIANG, G., SUN, Z., LIU, H., ZHOU, Z., RUAN, B., YANG, K., and GAO, X. Data-driven physical fields reconstruction of supercritical-pressure flow in regenerative cooling channel using POD-AE reduced-order model. International Journal of Heat and Mass Transfer, 217, 124699 (2023)
[41]	YE, W., ZHANG, F., WANG, L., and SONG, X. Shale crack identification based on acoustic emission experiment and wavenet data recovery. International Journal of Applied Mechanics, 14, 2250073 (2022)
[42]	MENG, D., ZHAO, Q., and XU, Z. Improve robustness of sparse PCA by L₁-norm maximization. Pattern Recognition, 45, 487–497 (2012)
[1]	AGNEY, A., SINGH, S., and SRINIVAS, G. Structural and fluid flow analysis of subsonic aircraft jet engine centrifugal compressor. Materials Today: Proceedings, 5, 19507–19516 (2018)
[2]	FATEHI, M. and RENZI, M. Modelling and development of ammonia-air non-premixed low NO_X combustor in a micro gas turbine: a CFD analysis. International Journal of Hydrogen Energy, 88, 1–10 (2024)
[3]	DIVAKARAN, A. M., GKANAS, E., JEWKES, J., SHEPHERD, S., JAMAL, W., FARAMEHR, S., and ABO-SERIE, E. 3D CFD modelling and experimental validation of conjugate heat transfer in wheel hub motors in micro-mobility vehicles. Applied Thermal Engineering, 258, 124430 (2025)
[4]	METHODS, C., MECH, A., and GONZÁLEZ-ALBUIXECH, V. F. Analysis of stress intensity factor oscillations in 3D cracks using domain integrals and the extended finite element method. Computer Methods in Applied Mechanics and Engineering, 436, 117739 (2025)
[5]	POPOV, I. S. Space-time adaptive ADER-DG finite element method with LST-DG predictor and a posteriori sub-cell ADER-WENO finite-volume limiting for multidimensional detonation waves simulation. Computers & Fluids, 284, 106425 (2024)
[6]	AHMADI, S. M., BEHBAHANI-NEJAD, M., and SHEKARI, Y. Transient modeling of natural gas in pipeline networks by two non-iterative explicit and implicit finite volume methods. Applied Mathematics and Computation, 478, 128793 (2024)
[7]	VAINIO, E., RUOZZI, A., KINNUNEN, H., LAURÉN, T., and HUPA, L. Low-temperature corrosion in a BFB boiler firing biomass and waste streams — online measurement technique for monitoring deposit corrosivity. Fuel, 371, 131864 (2024)
[8]	JODLBAUER, H., TRIPATHI, S., BACHMANN, N., and BRUNNER, M. Unlocking hidden market segments: a data-driven approach exemplified by the electric vehicle market. Expert Systems with Applications, 254, 124331 (2024)
[9]	NEGAHDARI, V. Integrating physics of the problem into data-driven methods to enhance elastic full-waveform inversion with uncertainty quantification. Physica D: Nonlinear Phenomena, 472, 134518 (2025)
[10]	MIN, G. and JIANG, N. Flow fields prediction for data-driven model of parallel twin cylinders based on POD-RBFNN and POD-BPNN surrogate models. Annals of Nuclear Energy, 199, 110342 (2024)
[11]	BIZON, K., CONTINILLO, G., LEISTNER, K. C., MANCARUSO, E., and VAGLIECO, B. M. POD-based analysis of cycle-to-cycle variations in an optically accessible diesel engine. Proceedings of the Combustion Institute, 32(2), 2809–2816 (2009)
[12]	LIBERGE, E. and HAMDOUNI, A. Reduced order modelling method via proper orthogonal decomposition (POD) for flow around an oscillating cylinder. Journal of Fluids and Structures, 26, 292–311 (2010)
[13]	LI, K., SU, H., CHU, J., and XU, C. A fast-POD model for simulation and control of indoor thermal environment of buildings. Building and Environment, 60, 150–157 (2013)
[14]	XIE, X. and TANG, S. Regularized dynamic mode decomposition algorithm for time sequence predictions. Theoretical and Applied Mechanics Letters, 14, 100555 (2024)
[15]	COALE, J. M. and ANISTRATOV, D. Y. A reduced-order model for nonlinear radiative transfer problems based on moment equations and POD-Petrov-Galerkin projection of the normalized Boltzmann transport equation. Journal of Computational Physics, 509, 113044 (2024)
[16]	FAINSTEIN, F., CATONI, J., ELEMANS, C. P. H., and MINDLIN, G. B. The reconstruction of flows from spatiotemporal data by autoencoders. Chaos, Solitons & Fractals, 176, 114115 (2023)
[17]	LORENZI, S., CAMMI, A., LUZZI, L., and ROZZA, G. POD-Galerkin method for finite volume approximation of Navier-Stokes and RANS equations. Computational Methods in Applied Mechanics and Engineering, 311, 151–179 (2016)
[18]	PEHERSTORFER, B. and WILLCOX, K. Dynamic data-driven reduced-order models. Computational Methods in Applied Mechanics and Engineering, 291, 21–41 (2015)
[19]	GOSSES, M., NOWAK, W., and WÖHLING, T. Explicit treatment for Dirichlet, Neumann and Cauchy boundary conditions in POD-based reduction of groundwater models. Advances in Water Resources, 115, 160–171 (2018)
[20]	TROMEUR-DERVOUT, D. and VASSILEVSKI, Y. POD acceleration of fully implicit solver for unsteady nonlinear flows and its application on grid architecture. Advances in Engineering Software, 38, 301–311 (2007)
[21]	LEE, K. and CARLBERG, K. T. Model reduction of dynamical systems on nonlinear manifolds using deep convolutional autoencoders. Journal of Computational Physics, 404, 108973 (2020)
[22]	KIM, Y., CHOI, Y., WIDEMANN, D., and ZOHDI, T. A fast and accurate physics-informed neural network reduced order model with shallow masked autoencoder. Journal of Computational Physics, 451, 110841 (2022)
[23]	MCBANE, S., CHOI, Y., and WILLCOX, K. Stress-constrained topology optimization of lattice-like structures using component-wise reduced order models. Computational Methods in Applied Mechanics and Engineering, 400, 115525 (2022)
[24]	EFTEKHAR AZAM, S. and MARIANI, S. Investigation of computational and accuracy issues in POD-based reduced order modeling of dynamic structural systems. Engineering Structures, 54, 150–167 (2013)
[43]	JIANG, C., QI, Y., LI, J., MEI, X., LI, W., and SHI, E. Stochastic simulation of magnetically controlled convection in porous media with random porosity via Karhunen-Loève expansion and intrusive polynomial chaos expansion. International Journal of Heat and Mass Transfer, 224, 125387 (2024)
[44]	BOUHOUBEINY, E. and DRUAULT, P. Note on the POD-based time interpolation from successive PIV images. Comptes Rendus-Mécanique, 337, 776–780 (2009)
[45]	POTAPOV, A., POTAPOVA, V., and PETERSON, M. A feasibility study of an autoencoder meta-model for improving generalization capabilities on training sets of small sizes. Pattern Recognition Letters, 80, 24–29 (2016)
[46]	WANG, L., YE, W., ZHU, Y., YANG, F., and ZHOU, Y. Optimal parameters selection of back propagation algorithm in the feedforward neural network. Engineering Analysis with Boundary Elements, 151, 575–596 (2023)
[47]	YANG, G. C., JING, L., KWOK, C. Y., and SOBRAL, Y. D. A comprehensive parametric study of LBM-DEM for immersed granular flows. Computational Geotechnique, 114, 103100 (2019)
[48]	LI, E., HE, Z. C., CHEN, L., LI, B., XU, X., and LIU, G. R. An ultra-accurate hybrid smoothed finite element method for piezoelectric problem. Engineering Analysis with Boundary Elements, 50, 188–197 (2015)
[49]	REYNOLDS, R. W., RAYNER, N. A., SMITH, T. M., STOKES, D. C., and WANG, W. An improved in situ and satellite SST analysis for climate. Journal of Climate, 15, 1609–1625 (2002)
[25]	FRIES, W. D., HE, X., and CHOI, Y. LaSDI: parametric latent space dynamics identification. Computational Methods in Applied Mechanics and Engineering, 399, 115436 (2022)
[26]	HE, X., CHOI, Y., FRIES, W. D., BELOF, J. L., and CHEN, J. S. gLaSDI: parametric physics-informed greedy latent space dynamics identification. Journal of Computational Physics, 489, 112267 (2023)
[27]	EVERSON, R. and SIROVICH, L. Karhunen-Loève procedure for gappy data. Journal of the Optical Society of America A, 12, 1657–1664 (1995)
[28]	YU, B., TONG, Y., HU, P., and GAO, Q. A novel inversion approach for identifying the shape of cavity by combining Gappy POD with direct inversion scheme. International Journal of Heat and Mass Transfer, 150, 119365 (2020)
[29]	WANG, Y. L., LI, C. S., ZHU, Z. L., XU, H., and NI, Y. Z. Real-time stress field reconstruction method using online monitoring parameters for thick-walled power plant components based on gappy proper orthogonal decomposition. International Journal of Pressure Vessels and Piping, 209, 105172 (2024)
[30]	JIANG, G., LIU, H., YANG, K., and GAO, X. A fast reduced-order model for radial integral boundary element method based on proper orthogonal decomposition in nonlinear transient heat conduction problems. Computational Methods in Applied Mechanics and Engineering, 368, 113190 (2020)
[31]	WANG, X., CAO, Z., LIU, Y., and WANG, F. Temperature reconstruction in data centers using gappy POD combined with genetic algorithm and optimization of sensor layout. Energy and Buildings, 326, 115078 (2025)
[32]	JIANG, G., KANG, M., CAI, Z., WANG, H., LIU, Y., and WANG, W. Online reconstruction of 3D temperature field fused with POD-based reduced order approach and sparse sensor data. International Journal of Thermal Sciences, 175, 107489 (2022)