Robust and fast online identification of streamwise vortices properties for closed-loop control purposes

TL;DR

This paper introduces two real-time vortex identification algorithms using PIV data for wind turbine control, with one being robust and the other faster, enabling improved closed-loop control of turbulent boundary layers.

Contribution

The paper develops and compares two online vortex characterization methods for real-time wind turbine control, enhancing robustness and computational efficiency.

Findings

The robust method effectively identifies vortex centers and strengths using full PIV fields.

The faster method uses only horizontal velocity lines to reduce computation time.

Both methods are suitable for real-time closed-loop control in wind turbine applications.

Abstract

We propose to combine the active vortex generators with the particle image velocimetry (PIV) measurements and post-processing streamwise vortex characterization algorithms into a feedback based closed-loop control system for wind turbine applications. We develop two vortex identification and characterization methods that use PIV realizations for the purpose of a real-time (online or on-the-fly) feedback-based control. Both methods can extract centers and strengths of streamwise vortices generated behind active vortex generators in a turbulent boundary layer flow, and we show how to integrate those in a closed-loop control strategy. For demonstration purposes we use stereoscopic PIV measurements at the wind tunnel facility obtained in the transverse-wall-normal plane behind active vortex generators. A robust algorithm is using the $Q$-criteria and the integration of vorticity of each extracted vortex. Results show that a moving window average of a small number of instantaneous fields is nevertheless needed for increased robustness. The robust method requires the full field PIV computation followed by spatial derivatives calculations. A faster method is developed, which, using only horizontal lines of vertical velocity, has a high potential to significantly cut down the computational effort relative to the robust method. We compare the two methods and discuss their shortcomings and the potential for the real-time, online, closed-loop control of turbulent boundary layers of the wind turbine blades.