Dynamic Input Mapping Inversion to Eliminate Algebraic Loops in Hydraulic Actuator Control

TL;DR

This paper introduces a dynamic input-mapping inversion method for hydraulic actuator control that eliminates algebraic loops, enhances robustness, and improves tracking performance without causing chatter or oil degradation.

Contribution

It proposes a novel nonlinear control architecture with a dynamic input-mapping inversion module to avoid algebraic loops in hydraulic actuators.

Findings

Effective loop elimination demonstrated on high-fidelity wind turbine simulator.

Improved tracking performance compared to state-of-the-art methods.

Validated robustness and stability on full-scale hydraulic system.

Abstract

The application of nonlinear control schemes to electro-hydraulic actuators often requires several alterations in the design of the controllers during their implementation. This is to overcome challenges that frequently arise in such control algorithms owing to model nonlinearities. Moreover, advanced control solutions for this type of systems often introduce input algebraic loops that pose significant design and tuning difficulties. Conventional methods to avoid such loops introduce chatter, which considerably degrade tracking performance and has oil degradation and wear as side effects. This study presents a nonlinear control architecture for hydraulic actuators that comprises low-complexity modules that facilitate robust high performance in tracking and avoids the drawbacks of chatter. The salient feature is a dynamic input-mapping inversion module that avoids algebraic loops in the control input and is followed by dedicated position control. The stability of the closed-loop system is analyzed using arguments from Lyapunov theory for cascaded non-autonomous nonlinear systems. The effectiveness of the proposed solution is evaluated on a high-fidelity simulator of a wind turbine pitch system, and validated on a full-scale laboratory setup that includes a hydraulic pitch system and blade bearing. Appropriate quantitative metrics are used to evaluate the closed-loop system performance in comparison to a state-of-the-art nonlinear design.