Magnetic Flux Map Reconstruction With Compressed Sensing for Motor Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing magnetic flux map acquisition methods for electric machines are time-consuming and require numerous measurements, leading to inefficiencies in high-performance control due to the need for dense data sampling and look-up tables, which compromise access speed and accuracy.
Innovation Solution
A compressed sensing approach is employed to reconstruct 2D and 3D magnetic flux maps using fewer randomly sampled data points, leveraging Fourier or wavelet bases to achieve a sparse representation, thereby reducing the number of required measurements and enabling an analytical model that accounts for spatial harmonics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional flux map acquisition methods are used with dense grid-point sampling, then measurement precision is improved, but loss of time increases significantly
Solution Approach 1:
The patent applies partial action by acquiring flux map data at only a subset of grid points rather than all dense grid points. Specifically, it measures data at selected grid points (e.g., boundary points or randomly selected points) and uses compressed sensing algorithms to reconstruct the complete flux map, thereby reducing measurement time while maintaining accuracy.
Solution Approach 2:
The patent replaces the mechanical measurement process (physical experimentation at all grid points) with a computational approach. By substituting physical measurements at every point with mathematical reconstruction algorithms (compressed sensing, interpolation), it dramatically reduces the time required while preserving measurement precision.
2Manufacturing precision
If look-up tables are used to store identified magnetic models, then manufacturing precision is improved, but device complexity increases due to search and interpolation requirements
Solution Approach 1:
The patent extracts only the essential features of the flux map into a compact analytical representation. Instead of storing complete look-up tables with all grid point data, it extracts key parameters and uses analytical functions (polynomial models, Fourier series) to represent the flux map, thereby reducing memory requirements and simplifying the control system.
Solution Approach 2:
The patent changes the representation parameters from discrete grid-point values to continuous analytical functions with fewer parameters. By transforming the flux map data into polynomial coefficients or Fourier series coefficients, it reduces the dimensionality of the stored data while enabling efficient computation and differentiation without complex search and interpolation operations.
3Measurement precision
If dense grid-point measurements are conducted to capture flux variation accurately, then measurement precision is improved, but productivity decreases due to the large number of required tests
Solution Approach 1:
The patent performs partial measurements at strategically selected grid points rather than conducting exhaustive measurements at all grid points. This partial sampling approach, combined with compressed sensing reconstruction, captures the essential flux variation characteristics while dramatically increasing productivity by reducing the number of required tests.
Solution Approach 2:
The patent performs preliminary analysis to identify which grid points provide the most informative measurements for flux map reconstruction. By pre-selecting optimal measurement locations based on the expected flux distribution patterns, it ensures that the reduced set of measurements still captures all necessary information for accurate reconstruction, thereby maintaining precision while improving productivity.
Data Source
AI summary
A magnetic flux mapping system for reconstructing a magnetic flux map efficiently from random measurements of a magnetic flux of an electric motor. The system first generates an intermediate reconstructed flux map dataset by performing a zero-padding method for the measurements of the magnetic flux, and then iteratively updates the flux map by FFT, soft-thresholding, inverse FFT, and data consistency steps until convergence. The flux map is used in motor control system to generate desired motor current, voltage, and speed to drive load.