A hybrid twin based on machine learning enhanced reduced order model for real-time simulation of magnetic bearings
Abstract The simulation of magnetic bearings involves highly non-linear physics, with high dependency on the input variation. Moreover, such a simulation is time consuming and can’t run, within realistic computation time for control purposes, when using classical computation methods. On the other ha...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2024-02-01
|
| Series: | Advanced Modeling and Simulation in Engineering Sciences |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s40323-024-00258-2 |
| _version_ | 1797273819595931648 |
|---|---|
| author | Chady Ghnatios Sebastian Rodriguez Jerome Tomezyk Yves Dupuis Joel Mouterde Joaquim Da Silva Francisco Chinesta |
| author_facet | Chady Ghnatios Sebastian Rodriguez Jerome Tomezyk Yves Dupuis Joel Mouterde Joaquim Da Silva Francisco Chinesta |
| author_sort | Chady Ghnatios |
| collection | DOAJ |
| description | Abstract The simulation of magnetic bearings involves highly non-linear physics, with high dependency on the input variation. Moreover, such a simulation is time consuming and can’t run, within realistic computation time for control purposes, when using classical computation methods. On the other hand, classical model reduction techniques fail to achieve the required precision within the allowed computation window. To address this complexity, this work proposes a combination of physics-based computing methods, model reduction techniques and machine learning algorithms, to tackle the requirements. The physical model used to represent the magnetic bearing is the classical Cauer Ladder Network method, while the model reduction technique is applied on the error of the physical model’s solution. Later on, in the latent space a machine learning algorithm is used to predict the evolution of the correction in the latent space. The results show an improvement of the solution without scarifying the computation time. The solution is computed in almost real-time (few milliseconds), and compared to the finite element reference solution. |
| first_indexed | 2024-03-07T14:49:41Z |
| format | Article |
| id | doaj.art-80414fd4ecab4cd2a08a6804ae3e0f31 |
| institution | Directory Open Access Journal |
| issn | 2213-7467 |
| language | English |
| last_indexed | 2024-03-07T14:49:41Z |
| publishDate | 2024-02-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | Advanced Modeling and Simulation in Engineering Sciences |
| spelling | doaj.art-80414fd4ecab4cd2a08a6804ae3e0f312024-03-05T19:46:32ZengSpringerOpenAdvanced Modeling and Simulation in Engineering Sciences2213-74672024-02-0111111510.1186/s40323-024-00258-2A hybrid twin based on machine learning enhanced reduced order model for real-time simulation of magnetic bearingsChady Ghnatios0Sebastian Rodriguez1Jerome Tomezyk2Yves Dupuis3Joel Mouterde4Joaquim Da Silva5Francisco Chinesta6PIMM Lab & ESI Chair, SKF Research Chair, Arts et Metiers Institute of TechnologyPIMM Lab & ESI Chair, Arts et Metiers Institute of TechnologyDepartment of Research and Development Engineering, SKF Magnetic MechatronicDepartment of Research and Development Engineering, SKF Magnetic MechatronicDepartment of Research and Development Engineering, SKF Magnetic MechatronicDepartment of Research and Development Engineering, SKF Magnetic MechatronicPIMM Lab & ESI Chair, CNRS@CREATE, Arts et Metiers Institute of TechnologyAbstract The simulation of magnetic bearings involves highly non-linear physics, with high dependency on the input variation. Moreover, such a simulation is time consuming and can’t run, within realistic computation time for control purposes, when using classical computation methods. On the other hand, classical model reduction techniques fail to achieve the required precision within the allowed computation window. To address this complexity, this work proposes a combination of physics-based computing methods, model reduction techniques and machine learning algorithms, to tackle the requirements. The physical model used to represent the magnetic bearing is the classical Cauer Ladder Network method, while the model reduction technique is applied on the error of the physical model’s solution. Later on, in the latent space a machine learning algorithm is used to predict the evolution of the correction in the latent space. The results show an improvement of the solution without scarifying the computation time. The solution is computed in almost real-time (few milliseconds), and compared to the finite element reference solution.https://doi.org/10.1186/s40323-024-00258-2Spectral methodReduced basisMachine learningMagnetic bearingMagnetic levitationLong-short term memory |
| spellingShingle | Chady Ghnatios Sebastian Rodriguez Jerome Tomezyk Yves Dupuis Joel Mouterde Joaquim Da Silva Francisco Chinesta A hybrid twin based on machine learning enhanced reduced order model for real-time simulation of magnetic bearings Advanced Modeling and Simulation in Engineering Sciences Spectral method Reduced basis Machine learning Magnetic bearing Magnetic levitation Long-short term memory |
| title | A hybrid twin based on machine learning enhanced reduced order model for real-time simulation of magnetic bearings |
| title_full | A hybrid twin based on machine learning enhanced reduced order model for real-time simulation of magnetic bearings |
| title_fullStr | A hybrid twin based on machine learning enhanced reduced order model for real-time simulation of magnetic bearings |
| title_full_unstemmed | A hybrid twin based on machine learning enhanced reduced order model for real-time simulation of magnetic bearings |
| title_short | A hybrid twin based on machine learning enhanced reduced order model for real-time simulation of magnetic bearings |
| title_sort | hybrid twin based on machine learning enhanced reduced order model for real time simulation of magnetic bearings |
| topic | Spectral method Reduced basis Machine learning Magnetic bearing Magnetic levitation Long-short term memory |
| url | https://doi.org/10.1186/s40323-024-00258-2 |
| work_keys_str_mv | AT chadyghnatios ahybridtwinbasedonmachinelearningenhancedreducedordermodelforrealtimesimulationofmagneticbearings AT sebastianrodriguez ahybridtwinbasedonmachinelearningenhancedreducedordermodelforrealtimesimulationofmagneticbearings AT jerometomezyk ahybridtwinbasedonmachinelearningenhancedreducedordermodelforrealtimesimulationofmagneticbearings AT yvesdupuis ahybridtwinbasedonmachinelearningenhancedreducedordermodelforrealtimesimulationofmagneticbearings AT joelmouterde ahybridtwinbasedonmachinelearningenhancedreducedordermodelforrealtimesimulationofmagneticbearings AT joaquimdasilva ahybridtwinbasedonmachinelearningenhancedreducedordermodelforrealtimesimulationofmagneticbearings AT franciscochinesta ahybridtwinbasedonmachinelearningenhancedreducedordermodelforrealtimesimulationofmagneticbearings AT chadyghnatios hybridtwinbasedonmachinelearningenhancedreducedordermodelforrealtimesimulationofmagneticbearings AT sebastianrodriguez hybridtwinbasedonmachinelearningenhancedreducedordermodelforrealtimesimulationofmagneticbearings AT jerometomezyk hybridtwinbasedonmachinelearningenhancedreducedordermodelforrealtimesimulationofmagneticbearings AT yvesdupuis hybridtwinbasedonmachinelearningenhancedreducedordermodelforrealtimesimulationofmagneticbearings AT joelmouterde hybridtwinbasedonmachinelearningenhancedreducedordermodelforrealtimesimulationofmagneticbearings AT joaquimdasilva hybridtwinbasedonmachinelearningenhancedreducedordermodelforrealtimesimulationofmagneticbearings AT franciscochinesta hybridtwinbasedonmachinelearningenhancedreducedordermodelforrealtimesimulationofmagneticbearings |