Uncertainty Quantification and Sensitivity Analysis in a Nonlinear Finite-Element Model of a Permanent Magnet Synchronous Machine

Beltrán-Pulido, Andrés; Aliprantis, Dionysios; Bilionis, Ilias; Muñoz, Andrés Pedreño; Leonardi, F.; Avery, Seth M.

doi:10.1109/tec.2020.3001914

Cited by 16 publications

(9 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our ten-dimensional parametric magnetostatics problem is challenging, albeit it is probably solvable in a more computationally efficient way by nonintrusive, regression-based approaches such as combining Gaussian process regression with principal component analysis [24]. However, we know that regression approaches do not scale well with increasing dimensions.…”

Section: Discussionmentioning

confidence: 99%

“…The design of high-performance electromechanical energy conversion devices, such as electric vehicle or aircraft motors, typically requires conducting parametric studies based on first principles within an optimization framework [1][2][3][4][5][6][7][8][9][10][11][12][13][14]. More recently, numerical methods for uncertainty quantification and sensitivity analysis of electric machines have been proposed [15][16][17][18][19][20][21][22][23][24][25]. Such studies rely on exhaustive evaluation of the underlying physical models.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Physics-informed neural networks for solving parametric magnetostatic problems

Beltrán-Pulido¹,

Bilionis²,

Aliprantis³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

The optimal design of magnetic devices becomes intractable using current computational methods when the number of design parameters is high. The emerging physics-informed deep learning framework has the potential to alleviate this curse of dimensionality. The objective of this paper is to investigate the ability of physics-informed neural networks to learn the magnetic field response as a function of design parameters in the context of a two-dimensional (2-D) magnetostatic problem. Our approach is as follows. We derive the variational principle for 2-D parametric magnetostatic problems, and prove the existence and uniqueness of the solution that satisfies the equations of the governing physics, i.e., Maxwell's equations. We use a deep neural network (DNN) to represent the magnetic field as a function of space and a total of ten parameters that describe geometric features and operating point conditions. We train the DNN by minimizing the physics-informed loss function

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Physics-informed neural networks for solving parametric magnetostatic problems

Beltrán-Pulido¹,

Bilionis²,

Aliprantis³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…In terms of machine design, several GPR applications can be found [18], [24]- [28]. However, most of the stated approaches lack a documentation, how the Kriging models were set up.…”

Section: Introductionmentioning

confidence: 99%

Single-Objective Machine Design with Gaussian Process Regression and Bayesian Optimization (Extended Version)

Rossmann,

Kamper,

Hackl

2024

Preprint

View full text Add to dashboard Cite

Due to the complex rotor design of reluctance synchronous machines, a finite element analysis is essential for the accurate calculation of machine relevant performance objectives. Reluctance synchronous machines tend to have a large torque ripple, if this objective is not considered during the machine design. This necessitates a large number of simulation steps, resulting in a high computational burden and a long simulation time per design evaluation. Therefore, an efficient optimization algorithm is required. This paper proposes a complete framework for single-objective machine design optimization using Gaussian process regression and Bayesian optimization. Focusing on reluctance synchronous machine design, different kernel functions (squared exponential, Matern, rational quadratic) and hyperparameter configurations ´ are assessed for regression accuracy of the optimization objectives mean torque, torque ripple, and power factor. The impact of noise in the input data on the regression results is also investigated. Bayesian optimization with the infill criterion Expected Improvement is finally used to perform machine design optimization for 18 design variables. Bayesian optimization outperforms the classical algorithms such as genetic or particle swarm algorithms. It results in a faster design optimization, even for such a high number of design variables.

show abstract

“…In recent years, for the modeling and analysis of motors, many studies have used finite element analysis (FEA) to solve the problem of the nonlinear relationship of motors. [15][16][17][18][19]. Compared to linear modeling and analysis using a traditional mathematical model, motor modeling and analysis using the FE method achieves higher fidelity [20].…”

Section: Introductionmentioning

confidence: 99%

Nonlinear Modeling and Analysis of Wound-Rotor Synchronous Starter/Generator (WRSSG) in Generating State for More Electric Aircraft

Du,

Liu,

et al. 2023

Actuators

View full text Add to dashboard Cite

The nonlinear modeling and analyzing of wound-rotor synchronous starter/generators (WRSSGs) plays a vital role in the analysis and monitoring of aircraft power systems. Moreover, they are of great significance with regard to the establishment of a future aircraft smart grid. However, owing to its nonlinear, high-dimensional, and strong coupling characteristics, this modeling has always remained in the frequency domain stage and the progress of more intuitive time domain modeling has been slow. This paper presents a nonlinear model of a WRSSG in a generating state. When the WRSSG is in power generation mode, most cases indicate that the aircraft is in flight mode. The establishment of the nonlinear model of the system in the power generation state is of great significance for the research of the health management and state monitoring of the aircraft power system and can improve the safety and reliability of the aircraft during flight. The model uses FE analysis and neural network to solve the nonlinear problem of the motor in the system and uses the improved variable parameter average model to solve the nonlinear problem of the rotating rectifier. According to the principle of signal transmission, a time domain model for the whole system is developed. Finally, the model is compiled by the RT-LAB real-time simulator. The nonlinear model performs well when compared with FE analysis results and tested against the MIL-STD-704F standard. The proposed nonlinear model and analysis results can be used for the condition monitoring and fault diagnosis of aircraft power systems. The hardware-in-the-loop test platform based on an accurate nonlinear model is a feasible means to study the failure of expensive equipment, and it can aid the study of irreversible failures of equipment at a low cost.

show abstract

Uncertainty Quantification and Sensitivity Analysis in a Nonlinear Finite-Element Model of a Permanent Magnet Synchronous Machine

Cited by 16 publications

References 28 publications

Physics-informed neural networks for solving parametric magnetostatic problems

Physics-informed neural networks for solving parametric magnetostatic problems

Single-Objective Machine Design with Gaussian Process Regression and Bayesian Optimization (Extended Version)

Nonlinear Modeling and Analysis of Wound-Rotor Synchronous Starter/Generator (WRSSG) in Generating State for More Electric Aircraft

Contact Info

Product

Resources

About