Machine-learned control-oriented flow estimation for multi-actuator multi-sensor systems exemplified for the fluidic pinball

Abstract

We propose the first machine-learned control-oriented flow estimation for multiple-input, multiple-output plants. The starting point is constant actuation with open-loop actuation commands leading to a database with simultaneously recorded actuation commands, sensor signals and flow fields. A key enabler is an estimator input vector comprising sensor signals and actuation commands. The mapping from the sensor signals and actuation commands to the flow fields is realized in an analytically simple, data-centric and general nonlinear approach. The analytically simple estimator generalizes linear stochastic estimation for actuation commands. The data-centric approach yields flow fields from estimator inputs by interpolating from the database – similar to Loiseau, Noack & Brunton (J. Fluid Mech., vol. 844, 2018, pp. 459–490) for unforced flow. The interpolation is performed with $k$ nearest neighbours ( $k$ NN). The general global nonlinear mapping from inputs to flow fields is obtained from a deep neural network (DNN) via an iterative training approach. The estimator comparison is performed for the fluidic pinball plant, which is a multiple-input, multiple-output wake control benchmark (Deng et al., J. Fluid Mech., vol. 884, 2020, A37) featuring rich dynamics under steady controls. We conclude that the machine learning methods clearly outperform the linear model. The performance of $k$ NN and DNN estimators are comparable for periodic dynamics. Yet the DNN performs consistently better when the flow is chaotic. Moreover, a thorough comparison regarding the complexity, computational cost and prediction accuracy is presented to demonstrate the relative merits of each estimator. The proposed method can be generalized for closed-loop flow control plants.

Elimination of lock-in phenomenon in vortex-induced vibration by passive modal control

“…Compared with passive control methods, active control techniques such as blowing and suction (Dong, Triantafyllou & Karniadakis 2008;Qu et al 2017) can directly inject energy into the flow field, changing the stability of the wake mode. Furthermore, machine learning methods provide a brand new approach to flow control (Ren, Wang & Tang 2019;Maceda et al 2021;Li, Li & Noack 2022). However, the location of the actuation, the velocity of the blowing or suction and other factors have crucial influences on the control effect.…”

“…Furthermore, machine learning methods provide a brand new approach to flow control (Ren, Wang & Tang 2019; Maceda et al. 2021; Li, Li & Noack 2022). However, the location of the actuation, the velocity of the blowing or suction and other factors have crucial influences on the control effect.…”

Dynamics-augmented cluster-based network model

“…Increasingly powerful machine learning methods can make data-driven ROMs more automated. Examples include the sparse identification of nonlinear dynamics (SINDy) aim at human interpretable models (Brunton, Proctor & Kutz 2016), ROMs with artificial neural networks (San & Maulik 2018;San, Maulik & Ahmed 2019;Zhu et al 2019;Kou & Zhang 2021), turbulence modelling and flow estimation with multi-input multi-output by deep neural networks (Kutz 2017;Li, Li & Noack 2022) and manifold learning methods (Farzamnik et al 2023).…”

Pymodconn: A Python Package for Developing Modular Sequence-to-Sequence Control-Oriented Deep Neural Networks

“…Black-box models, sometimes referred to as data-driven models, predominantly leverage statistical or machine learning techniques to predict system responses [11]. Examples like AutoRegressive Integrated Moving Average (ARIMA) based control-oriented models for electric vehicle energy management [12], Support Vector Machines (SVMs) based control-oriented models for diesel engine emission control [13], and Deep Neural Networks (DNNs) based control-oriented models for building control [14] and control-oriented flow estimation [15] fall under this category.…”