Vertical Vibration Reduction of Maglev Vehicles using Nonlinear MPC
Mario Hermle, Arnim Kargl, Peter Eberhard · 2025
Read Vertical Vibration Reduction of Maglev Vehicles using Nonlinear MPC by Mario Hermle, Arnim Kargl, Peter Eberhard (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-13015.txt.
Introduction
Mario Hermle, Arnim Kargl, and Peter Eberhard give magnetic levitation trains a predictive controller that minds the ride, in this 2025 paper. Standard controllers treat the magnet and the car body apart, which leaves passengers feeling vertical shakes from track bumps. Their nonlinear scheme folds magnet forces and suspension motion into one model and plans ahead.
Simulations on a detailed model of the Transrapid train showed stronger shaking suppression than existing controllers. The scheme also lets engineers trade gap precision against comfort more cleanly. The authors pitch it for future high-speed maglev lines.
Catalog record
| CREATOR | Mario Hermle, Arnim Kargl, Peter Eberhard |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 27 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-13015
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Vertical Vibration Reduction of Maglev Vehicles using Nonlinear MPC Mario Hermle, Arnim Kargl, Peter Eberhard Abstract This work presents a novel Nonlinear Model Predictive Control (NMPC) strategy for high-speed Maglev vehicles that explicitly incorporates mechanical suspension dynamics into the control model. Unlike conventional approaches, which often neglect the interaction between levitation magnet and car body motion, the proposed method enables predictive vibration mitigation by modeling both electromagnetic forces and suspension behavior. This integrated approach significantly improves passenger comfort and ride quality by reducing vertical oscillations caused by track irregularities. Moreover, it allows for a more effective tuning of the trade-off between precise air gap tracking and ride comfort. Simulations based on a detailed multibody model of the Transrapid demonstrate that the method outperforms existing controllers in vibration suppression, making it a promising solution for future high-speed Maglev applications. 1 Introduction Electromagnetic suspension (EMS) Maglev systems, such as the Transrapid, rely on active control to maintain a stable levitation air gap between the vehicle and the guideway. This is essential for system stability, as the open-loop dynamics are inherently unstable. At the same time, the controller must ensure ride comfort by limiting accelerations transmitted to the vehicle, especially to the car body. In this context, low frequency car body movements up to frequencies of 5\,\mathrm{Hz} drastically reduce ride comfort for the passengers [ 1 , 2 ] . While there exist various active control strategies for vibration suppression of high-speed trains, they usually require additional actuators and sensors for the suspension, which increases the system complexity and cost [ 3 ] . In contrast, the approach presented in this study utilizes the existing hardware to achieve a reduction in vertical vibrations solely through adaptation of the levitation control algorithm. The central challenge lies in balancing two objectives, namely precise tracking of the nominal air gap and suppression of vertical vibrations caused by guideway irregularities. These irregularities, resulting from structural tolerances and elastic deflection of the guideway, are unknown to the controller and must be compensated in real time. This study investigates the use of Nonlinear Model Predictive Control (NMPC) to address this challenge. NMPC offers a structured framework for optimizing multiple objectives subject to physical constraints and nonlinear system dynamics. In particular, and this is the novel contribution of this paper, it allows for incorporating both magnet and car body dynamics, including the suspension, into a controller model, which enables a direct influence on ride comfort. The following sections describe the modeling approach, control design, and simulation results with a detailed multibody vehicle model that demonstrate the effectiveness of the proposed NMPC scheme in achieving stability and increased passenger comfort. 2 Air Gap Control of Maglev Systems Figure 1: Schematic of the Transrapid vehicle in the x - y -plane. The Maglev vehicle model considered in this study is based on the Transrapid. A typical vehicle configuration consists of three sections as illustrated in Figure 1 . The car body is connected to the levitation and guidance magnets through the levitation chassis. While the levitation magnets attract the vehicle towards the guideway from below, the guidance magnets keep it aligned within the guideway. A soft suspension between car body and levitation chassis ensures ride comfort by dampening the high frequency vibrations of the magnets. For a more detailed description of the Transrapid vehicle, see [ 4 ] . For this study, the focus is placed on the levitation aspect. Due to their design principle, electromagnetic suspension type Maglev vehicles such as the Transrapid require active control of the magnetic forces to achieve stable levitation, as the open-loop system is inherently unstable. However, this design allows for levitation from standstill [ 5 ] and minimizes issues related to magnetic flux leakage, which are more pronounced in electrodynamic suspension (EDS) systems [ 6 ] . The levitation control algorithm consequently pursues two primary objectives. Firstly, the controller must ensure the stabilization of the vehicle at the nominal air gap, maintaining this equilibrium despite the presence of external disturbances. The nominal air gap is defined as the desired distance between the electromagnets and the guideway. Secondly, the controller aims to minimize accelerations experienced at both the magnet and car body levels to enhance passenger comfort. These two objectives represent conflicting control goals that require careful balancing. The guideway consists of individual girders with fixed length. However, due to structural bending and construction imperfections, the guideway surface is not perfectly smooth but exhibits periodic deviations. These irregularities directly influence the measured air gap for the controller and, consequently, affect the control error. However, the guideway profile is not known to the controller, which means it can only react to such disturbances rather than anticipate them. An ideal tracking controller would perfectly follow the air gap reference, thereby maintaining zero control error. This would lead to high accelerations of both the magnet and the car body, adversely affecting ride comfort. In contrast, an ideal acceleration-minimizing controller would smooth out the effects of the guideway irregularities completely. However, this may result in excessively large variations in the air gap, potentially compromising system stability. Therefore, a trade-off is required, yielding a controller that effectively stabilizes the system while also minimizing accelerations to enhance ride comfort. In this study, the problem is addressed using Nonlinear Model Predictive Control (NMPC). NMPC can explicitly account for the system nonlinearities arising in the electromagnet dynamics. Furthermore, it is able to incorporate constraints on system states and control inputs, which is necessary in this case as the applied voltage to the electromagnets is limited. The predictive capability makes NMPC particularly well-suited for balancing the aforementioned competing objectives through tuning of a cost function. While a soft suspension is able to decouple the motion of the levitation chassis and the car body for high frequencies, the damping of low frequencies is limited. This motivates the development of an algorithm, which is able to dampen these low frequency vibrations. 3 Modeling The modeling of the control subsystem is divided into two main components. First, a mechanical model of the system is necessary to capture the dynamic interaction between the levitation magnets and the vehicle structure. This model serves as the foundation for predicting the vertical motion of the system. Following the principle of the magnetic wheel [ 7 ] , the single mass model as depicted in Figure 2a is derived. (a) Single mass model (b) Two-mass model Figure 2: The single mass model contains one unsprung mass representing one levitation half magnet. The two-mass model contains one additional sprung mass representing part of the car body’s mass. There, the mass m_{1} accounts for one half of a levitation magnet and one quarter of a levitation chassis. It is influenced by gravity and pulled towards the guideway from below by the magnet force F_{\mathrm{mag}} . The car body and the weight of the passengers is modeled with the static load force F_{\mathrm{L}} . The absolute position of mass m_{1} in the inertial frame is z_{1}=d_{\mathrm{gw}}+s , where s is the air gap between magnet and guideway and d_{\mathrm{gw}} the guideway deflection. This single mass model has been successfully implemented and deployed in previous studies [ 8 , 9 ] , however, with limited capabilities on