Non-Destructive Quality Assessment of Permanent Magnets inPedelec Drive Systems via Magnetic Field Mapping and SupportVector Machine Classification

This paper presents a non-destructive methodology for quantifying irreversible permanent magnet degradation inpermanent magnet synchronous motors (PMSMs) used in pedelec drive systems, using spatially resolved magnetic-field-mapping [MFM] combined with machine learning classification. Fourteen NdFeB magnets from a daum electronicImpulse2 e-bike drive system were characterized before and after 5,000 km of endurance testing on a biomechanicalHardware-in-the-Loop test bench at Aalen University. Spatially resolved magnetic field mapping was used to obtain fluxdensity distributions for comparison of fully magnetized and degraded magnet states of the permanent magnets. Theresulting difference field maps indicate small but measurable field deviations in the order of a few milliTesla (0.3 to 7.6 mT)across the magnet selection. While these deviations are modest relative to the nominal flux density of over 300 mT, theobserved spatial patterns, concentrated at peripheral magnet regions, show localized demagnetization effects.Twenty-five statistical, spatial, and texture features were extracted from each grayscale field map, including intensitydistribution moments, gradient magnitude statistics, Gray-Level Co-Occurrence Matrix (GLCM) texture parameters, andpercentile distributions. A comparative classification study employing five machine learning algorithms (SVM with linearand RBF kernels, Decision Trees, k-Nearest Neighbors, and Bagged Tree Ensembles) was evaluated using Leave-One-Outcross-validation. The results demonstrate that the proposed methodology is capable of detecting and classifying these smallfield deviations, establishing a baseline framework for non-destructive permanent magnet quality assessment.

Authors

Institutions

Publication Details

Journal
Research Square
Published
2026-11-18
DOI
https://doi.org/10.21203/rs.3.rs-10053810/v1
Primary Topic
Electric Motor Design and Analysis
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Non-Destructive Quality Assessment of Permanent Magnets inPedelec Drive Systems via Magnetic Field Mapping and SupportVector Machine Classification

Maximilian Lanz, Sebastian Feldmann, Matthias Wiedenmann
1 links
Research Square
Electric Motor Design and Analysis
preprint

Non-Destructive Quality Assessment of Permanent Magnets inPedelec Drive Systems via Magnetic Field Mapping and SupportVector Machine Classification

Maximilian Lanz, Sebastian Feldmann, Matthias Wiedenmann
preprint en

Abstract

This paper presents a non-destructive methodology for quantifying irreversible permanent magnet degradation inpermanent magnet synchronous motors (PMSMs) used in pedelec drive systems, using spatially resolved magnetic-field-mapping [MFM] combined with machine learning classification. Fourteen NdFeB magnets from a daum electronicImpulse2 e-bike drive system were characterized before and after 5,000 km of endurance testing on a biomechanicalHardware-in-the-Loop test bench at Aalen University. Spatially resolved magnetic field mapping was used to obtain fluxdensity distributions for comparison of fully magnetized and degraded magnet states of the permanent magnets. Theresulting difference field maps indicate small but measurable field deviations in the order of a few milliTesla (0.3 to 7.6 mT)across the magnet selection. While these deviations are modest relative to the nominal flux density of over 300 mT, theobserved spatial patterns, concentrated at peripheral magnet regions, show localized demagnetization effects.Twenty-five statistical, spatial, and texture features were extracted from each grayscale field map, including intensitydistribution moments, gradient magnitude statistics, Gray-Level Co-Occurrence Matrix (GLCM) texture parameters, andpercentile distributions. A comparative classification study employing five machine learning algorithms (SVM with linearand RBF kernels, Decision Trees, k-Nearest Neighbors, and Bagged Tree Ensembles) was evaluated using Leave-One-Outcross-validation. The results demonstrate that the proposed methodology is capable of detecting and classifying these smallfield deviations, establishing a baseline framework for non-destructive permanent magnet quality assessment.

Research Square
Hochschule Aalen (DE)
Electric Motor Design and Analysis
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon