Dazhi Wang1, Sihan Wang1, Deshan Kong1, Jiaxing Wang1, Wenhui Li1, and Michael G. Pecht2
1 School of Information Science and Engineering, Northeastern University, Shenyang, 110819, China,
2 Center for Advanced Life Cycle Engineering, University of Maryland, College Park, MD, 20740, USA
For more information about this article and related research, please contact Prof. Michael G. Pecht.
The objective of this letter is to study the prediction of the electromagnetic field and the output performance of permanent magnet eddy current devices based on a physics-informed sparse neural network (PISNN). In order to achieve this goal, a unified physical model is firstly defined according to different types of permanent magnet eddy current devices, which is equivalent to solving a parameterized magnetic quasi-static problem. A soft constraint module and a hard constraint module, composed of physical equations, are constructed. The soft constraints are then integrated into the neural network's objective function, while the hard constraint module is utilized to predict device performance and physical field. Stochastic gradient descent is used to minimize the residual of the physical equations during PISNN training. Subsequently, the structural parameters and operating parameters of PMECD are modified to verify the generalization ability of the model. Our results indicate that PISNN accurately and efficiently predicts the electromagnetic field distribution and the output torque. Furthermore, our prediction results for permanent magnet eddy current devices with different parameters demonstrate the potential of the proposed method for transfer learning.
This article is available online here and to CALCE Consortium Members for personal review.