Leveraging reduced-order models for state estimation using deep learning

State estimation is key to both analyzing physical mechanisms and enabling real-time control of fluid flows. A common estimation approach is to relate sensor measurements to a reduced state governed by a reduced-order model (ROM). (When desired, the full state can be recovered via the ROM). Current methods in this category nearly always use a linear model to relate the sensor data to the reduced state, which often leads to restrictions on sensor locations and has inherent limitations in representing the generally nonlinear relationship between the measurements and reduced state. We propose an alternative methodology where a neural network architecture is used to learn this nonlinear relationship. Neural network is a natural choice for this estimation problem, as a physical interpretation of the reduced state-sensor measurement relationship is rarely obvious. The proposed estimation framework is agnostic to the ROM employed, and can be incorporated into any choice of ROMs derived on a linear subspace (e.g., proper orthogonal decomposition) or a nonlinear manifold. The proposed approach is demonstrated on a two-dimensional model problem of separated flow around a flat plate, and is found to outperform common linear estimation alternatives.

[1]  Clarence W. Rowley,et al.  Linearly-Recurrent Autoencoder Networks for Learning Dynamics , 2017, SIAM J. Appl. Dyn. Syst..

[2]  Charbel Farhat,et al.  Nonlinear model order reduction based on local reduced‐order bases , 2012 .

[3]  T. Colonius,et al.  A fast immersed boundary method using a nullspace approach and multi-domain far-field boundary conditions , 2008 .

[4]  Emmanuel J. Candès,et al.  Near-Optimal Signal Recovery From Random Projections: Universal Encoding Strategies? , 2004, IEEE Transactions on Information Theory.

[5]  Jeff D. Eldredge,et al.  Machine-Learning-Based Detection of Aerodynamic Disturbances Using Surface Pressure Measurements , 2019 .

[6]  Nathan E. Murray,et al.  Modified quadratic stochastic estimation of resonating subsonic cavity flow , 2007 .

[7]  Steven L. Brunton,et al.  Shallow Learning for Fluid Flow Reconstruction with Limited Sensors and Limited Data , 2019, ArXiv.

[8]  Stéphane Mallat,et al.  Understanding deep convolutional networks , 2016, Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences.

[9]  B. Podvin,et al.  On combining linear stochastic estimation and proper orthogonal decomposition for flow reconstruction , 2018 .

[10]  P. Schmid,et al.  Dynamic mode decomposition of numerical and experimental data , 2008, Journal of Fluid Mechanics.

[11]  J. Kutz,et al.  Compressive Sensing Based Machine Learning Strategy For Characterizing The Flow Around A Cylinder With Limited Pressure Measurements , 2013 .

[12]  Kookjin Lee,et al.  Model reduction of dynamical systems on nonlinear manifolds using deep convolutional autoencoders , 2018, J. Comput. Phys..

[13]  S. Obayashi,et al.  Assessment of probability density function based on POD reduced-order model for ensemble-based data assimilation , 2015 .

[14]  Steven L. Brunton,et al.  Greedy Sensor Placement With Cost Constraints , 2018, IEEE Sensors Journal.

[15]  Steven L. Brunton,et al.  Nonlinear model reduction for dynamical systems using sparse sensor locations from learned libraries. , 2015, Physical review. E, Statistical, nonlinear, and soft matter physics.

[16]  T. Colonius,et al.  Data-assimilated low-order vortex modeling of separated flows , 2018, Physical Review Fluids.

[17]  Mark N. Glauser,et al.  Towards practical flow sensing and control via POD and LSE based low-dimensional tools , 2004 .

[18]  Steven L. Brunton,et al.  Data-Driven Sparse Sensor Placement for Reconstruction: Demonstrating the Benefits of Exploiting Known Patterns , 2017, IEEE Control Systems.

[19]  Lawrence Sirovich,et al.  Karhunen–Loève procedure for gappy data , 1995 .

[20]  Steven L. Brunton,et al.  Sparse reduced-order modelling: sensor-based dynamics to full-state estimation , 2017, Journal of Fluid Mechanics.

[21]  Steven L. Brunton,et al.  Robust flow reconstruction from limited measurements via sparse representation , 2018, Physical Review Fluids.

[22]  Petros Koumoutsakos,et al.  Machine Learning for Fluid Mechanics , 2019, Annual Review of Fluid Mechanics.

[23]  Ronald Adrian,et al.  On the role of conditional averages in turbulence theory. , 1975 .

[24]  T. Başar,et al.  A New Approach to Linear Filtering and Prediction Problems , 2001 .

[25]  J. Peraire,et al.  Balanced Model Reduction via the Proper Orthogonal Decomposition , 2002 .

[26]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[27]  N. Gordon,et al.  Novel approach to nonlinear/non-Gaussian Bayesian state estimation , 1993 .

[28]  Clarence W. Rowley,et al.  Integration of non-time-resolved PIV and time-resolved velocity point sensors for dynamic estimation of velocity fields , 2013 .

[29]  Steven L. Brunton,et al.  Optimal Sensor Placement and Enhanced Sparsity for Classification , 2013, ArXiv.