Linear quadratic regulation control for falling liquid films

TL;DR

This paper introduces a novel control methodology using linear-quadratic regulation (LQR) combined with reduced-order models to stabilize falling liquid films, bridging the gap between complex fluid dynamics and practical control techniques.

Contribution

It develops a multi-level control framework applying LQR to reduced-order models and demonstrates its effectiveness on the full Navier-Stokes system for liquid film stabilization.

Findings

Successful stabilization of liquid films using LQR-based control.

Weighted-residual model outperforms Benney model in control accuracy.

Control scheme remains effective beyond initial applicability range.

Abstract

We propose and analyse a new methodology based on linear-quadratic regulation (LQR) for stabilising falling liquid films via blowing and suction at the base. LQR methods enable rapidly responding feedback control by precomputing a gain matrix, but are only suitable for systems of linear ordinary differential equations (ODEs). By contrast, the Navier-Stokes equations that describe the dynamics of a thin liquid film flowing down an inclined plane are too complex to stabilise with standard control-theoretical techniques. To bridge this gap we use reduced-order models - the Benney equation and a weighted-residual integral boundary layer model - obtained via asymptotic analysis to derive a multi-level control framework. This framework consists of an LQR feedback control designed for a linearised and discretised system of ODEs approximating the reduced-order system, which is then applied to the full Navier-Stokes system. The control scheme is tested via direct numerical simulation (DNS), and compared to analytical predictions of linear stability thresholds and minimum required actuator numbers. Comparing the strategy between the two reduced-order models we show that in both cases we can successfully stabilise towards a uniform flat film across their respective ranges of valid parameters, with the more accurate weighted-residual model outperforming the Benney-derived controls. The weighted-residual controls are also found to work successfully far beyond their anticipated range of applicability. The proposed methodology increases the feasibility of transferring robust control techniques towards real-world systems, and is also generalisable to other forms of actuation.