Session: Student Poster Competition

Submission Number: 185711

Torsional Vibration Damping Using a Modular Control Scheme With Preview 

This poster investigates active torsional vibration damping for a specific drivetrain consisting of two electrical motors, a gearbox, and a turbine. Due to the elastic couplings between the drivetrain components, the system exhibits three distinct torsional natural frequencies (TNFs). In accordance with relevant API specifications, these TNFs must not be excited by the rotational speed of the investigated drivetrain. This results in speed exclusions of approximately ±10 percent around each TNF.
To reduce these speed restrictions to ±2 percent, a modular control scheme is established. The control architecture consists of a nominal PI speed controller augmented by a linear quadratic regulator (LQR) and a model predictive controller (MPC) with preview. The modular structure was deliberately chosen to ensure straightforward integration into an existing control system for a 9 MW plant.
All controllers operate in parallel and can be activated seamlessly during normal operation as extensions to the nominal PI controller. Each controller fulfills a distinct task within the overall control concept. The LQR provides active damping of TNFs excited by disturbances, mass imbalance, and misalignment. The MPC optimizes speed trajectories in order to avoid excitation of TNFs during transients.
The modular control scheme was validated on a scale-down test bench representing the drivetrain under investigation. Experimental results show that the modular control scheme reduces torsional vibration amplitudes by approximately 30 percent while allowing a separation margin of only 2 percent around the TNFs. The experimental results prove that the proposed modular concept is a robust and flexible addition to the PI controller and extends the usable speed range.

Presenting Author: Matthias Geissmann University of Applied Sciences and Arts Northwestern Switzerland

Presenting Author Biography: Matthias Geissmann is a PhD student in control engineering and system identification with a focus on torsional multi-mass systems. He holds a BSc in Systems Engineering and an MSc in Mechatronics and Automation. His research concentrates on modeling, identification, and advanced control methods for torsional vibration damping in drivetrain systems.

Authors:

Matthias Geissmann University of Applied Sciences and Arts Northwestern Switzerland
Pieder Jörg ABB Schweiz AG
Thomas Besselmann University of Applied Sciences and Arts Northwestern Switzerland
Jana Kertzscher Technische Universität Bergakademie Freiberg