Variable Resonance Frequency Identification in a Magnetic Track System
(2025)Department of Automatic Control
- Abstract
- Identifying resonance frequencies caused by liquid sloshing in modular transport systems is essential for effective vibration suppression and minimising mechanical wear. An example of such a system is the ACOPOStrak by B&R, which utilises a Zero Vibration Filter (ZVF) to mitigate these effects. Accurate tuning of this filter requires knowledge of the system’s resonance frequencies. Obtaining this information can be challenging due to varying loads, segment geometries, and unknown mechanical properties. This thesis examines identification methods that rely solely on standard PLC measurements, aiming to enable real-time tuning without the need for external instrumentation.
Experiments were conducted using two different axisymmetric... (More) - Identifying resonance frequencies caused by liquid sloshing in modular transport systems is essential for effective vibration suppression and minimising mechanical wear. An example of such a system is the ACOPOStrak by B&R, which utilises a Zero Vibration Filter (ZVF) to mitigate these effects. Accurate tuning of this filter requires knowledge of the system’s resonance frequencies. Obtaining this information can be challenging due to varying loads, segment geometries, and unknown mechanical properties. This thesis examines identification methods that rely solely on standard PLC measurements, aiming to enable real-time tuning without the need for external instrumentation.
Experiments were conducted using two different axisymmetric containers filled with either liquid or solid payloads, mounted on shuttles travelling along a closedloop track. Despite variations in load, the identified resonance frequencies remained essentially unchanged, indicating that structural factors, such as segment geometry and mechanical coupling, had a greater impact on system behaviour. Among the tested methods, the bandpass sweep produced the clearest frequency peaks but with a heavy computational load. The Recursive Least Squares (RLS) algorithm showed promise for real-time implementation due to its lower computational cost; however, it required time to converge and was sensitive to segment-specific dynamics. Additionally, the Fast Fourier Transform (FFT) yielded low-resolution results and may not be the most suitable option for this application.
The position error signal was effective in identifying dominant frequencies in specific segments. However, the results indicate that a more reliable identification system may need more comprehensive measurements or model-based methods. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9212274
- author
- Melgar, Christian and Hemark, Felix
- supervisor
- organization
- year
- 2025
- type
- H3 - Professional qualifications (4 Years - )
- subject
- report number
- TFRT-6292
- other publication id
- 0280-5316
- language
- English
- id
- 9212274
- date added to LUP
- 2025-09-18 14:14:54
- date last changed
- 2025-09-18 14:14:54
@misc{9212274,
abstract = {{Identifying resonance frequencies caused by liquid sloshing in modular transport systems is essential for effective vibration suppression and minimising mechanical wear. An example of such a system is the ACOPOStrak by B&R, which utilises a Zero Vibration Filter (ZVF) to mitigate these effects. Accurate tuning of this filter requires knowledge of the system’s resonance frequencies. Obtaining this information can be challenging due to varying loads, segment geometries, and unknown mechanical properties. This thesis examines identification methods that rely solely on standard PLC measurements, aiming to enable real-time tuning without the need for external instrumentation.
Experiments were conducted using two different axisymmetric containers filled with either liquid or solid payloads, mounted on shuttles travelling along a closedloop track. Despite variations in load, the identified resonance frequencies remained essentially unchanged, indicating that structural factors, such as segment geometry and mechanical coupling, had a greater impact on system behaviour. Among the tested methods, the bandpass sweep produced the clearest frequency peaks but with a heavy computational load. The Recursive Least Squares (RLS) algorithm showed promise for real-time implementation due to its lower computational cost; however, it required time to converge and was sensitive to segment-specific dynamics. Additionally, the Fast Fourier Transform (FFT) yielded low-resolution results and may not be the most suitable option for this application.
The position error signal was effective in identifying dominant frequencies in specific segments. However, the results indicate that a more reliable identification system may need more comprehensive measurements or model-based methods.}},
author = {{Melgar, Christian and Hemark, Felix}},
language = {{eng}},
note = {{Student Paper}},
title = {{Variable Resonance Frequency Identification in a Magnetic Track System}},
year = {{2025}},
}