Physics-augmented deep learning for high-speed electromagnetic simulation and optimization

2021 ◽  
Author(s):  
Mingkun Chen ◽  
Robert Lupoiu ◽  
Chenkai Mao ◽  
Der-Han Huang ◽  
Jiaqi Jiang ◽  
...  
Keyword(s):  
Neural Network ◽  
Electromagnetic Field ◽  
Electric Fields ◽  
High Speed ◽  
Near Field ◽  
Physical Constraints ◽  
Simulation And Optimization ◽  
Field Distributions ◽  
The Neural Network ◽  
Nanostructure Arrays

Abstract The calculation of electromagnetic field distributions within structured media is central to the optimization and validation of photonic devices. We introduce WaveY-Net, a hybrid data- and physics-augmented convolutional neural network that can predict electromagnetic field distributions with ultra fast speeds and high accuracy for entire classes of dielectric photonic structures. This accuracy is achieved by training the neural network to learn only the magnetic near-field distributions of a system and to use a discrete formalism of Maxwell's equations in two ways: as physical constraints in the loss function and as a means to calculate the electric fields from the magnetic fields. As a model system, we construct a surrogate simulator for periodic silicon nanostructure arrays and show that the high speed simulator can be directly and effectively used in the local and global freeform optimization of metagratings. We anticipate that physics-augmented networks will serve as a viable Maxwell simulator replacement for many classes of photonic systems, transforming the way they are designed.

scholarly journals Axle Temperature Monitoring and Neural Network Prediction Analysis for High-Speed Train under Operation

Symmetry ◽  
2020 ◽  
Vol 12 (10) ◽  
pp. 1662
Author(s):  
Wei Hao ◽  
Feng Liu
Keyword(s):  
Neural Network ◽  
Network Model ◽  
Neural Network Model ◽  
High Speed ◽  
Prediction Method ◽  
High Speed Train ◽  
The Neural Network ◽  
Neural Network Prediction ◽  
Network Prediction ◽  
Trained Neural Network

Predicting the axle temperature states of the high-speed train under operation in advance and evaluating working states of axle bearings is important for improving the safety of train operation and reducing accident risks. The method of monitoring the axle temperature of a train under operation, combined with the neural network prediction method, was applied. A total of 36 sensors were arranged at key positions such as the axle bearings of the train gearbox and the driving end of the traction motor. The positions of the sensors were symmetrical. Axle temperature measurements over 11 days with more than 38,000 km were obtained. The law of the change of the axle temperature in each section was obtained in different environments. The resultant data from the previous 10 days were used to train the neural network model, and a total of 800 samples were randomly selected from eight typical locations for the prediction of axle temperature over the following 3 min. In addition, the results predicted by the neural network method and the GM (1,1) method were compared. The results show that the predicted temperature of the trained neural network model is in good agreement with the experimental temperature, with higher precision than that of the GM (1,1) method, indicating that the proposed method is sufficiently accurate and can be a reliable tool for predicting axle temperature.

Aircraft Mounted Neural Network Electrostatic Discharge (ESD) Location

2012 ◽  
Vol 721 ◽  
pp. 331-336
Author(s):  
Paul Ratnamahilan Polycarp Hoole ◽  
Nur Farah Aziz ◽  
Velappa Ganapathy ◽  
Kanesan Jeevan ◽  
Ramiah Harikrishnan ◽  
...  
Keyword(s):  
Neural Network ◽  
Electrical Activity ◽  
Electric Fields ◽  
Array Antenna ◽  
Lightning Activity ◽  
Electronic Systems ◽  
Distance Range ◽  
The Neural Network ◽  
Unique Shape ◽  
Element Array

Abstract. Cloud to ground and cloud to cloud lightning flashes pose a threat to the aircraft body and the electronic systems inside the aircraft. In this paper we present a single unit, as opposed to a three unit, lightning locator mounted on the aircraft that uses the wave-shapes of electromagnetic fields radiated by lightning and electrical activity ahead of the aircraft to locate the distance range of lightning activity. A three element array antenna scans the area ahead of the aircraft to narrow down the area ahead where the lightning or threatening electrical activity is. Moreover, the unique shape of the electric fields depending on the distance from the lightning activity is used by a neural network to train and recognize the distance range of the lightning activity from the aircraft on which the lightning detector is mounted. The combined use of the three element array antenna and the neural network provides the required knowledge of lightning activity for the pilot to take evasive action.

Study on Decreasing Vibration of High-Speed Train Semi-Active Suspension System

2011 ◽  
Vol 230-232 ◽  
pp. 1104-1109
Author(s):  
Zhen Ping Fan ◽  
Heng Zeng ◽  
Jian Wei Yang ◽  
Jie Li
Keyword(s):  
Neural Network ◽  
Control Strategy ◽  
High Speed ◽  
Driving Safety ◽  
High Speed Train ◽  
Proportional Valve ◽  
Joint Simulation ◽  
The Neural Network ◽  
Adaptive Control Strategy ◽  
Active Damper

Lateral semi-active damper is designed by author based on the electro-hydraulic proportional valve, from the perspective angle of improving vehicle comfort; its purpose is to ensure vehicle driving safety. At the same time, the neural network adaptive control strategy is used for joint simulation of semi-active damper. The results show that lateral semi-active damper with the train body has significantly improved compared to the traditional passive lateral damper acceleration.

Neural Network Emulation of a Magnetically Suspended Rotor

2004 ◽  
Vol 126 (2) ◽  
pp. 373-384 ◽  
Author(s):  
A. Escalante ◽  
V. Guzma´n ◽  
M. Parada ◽  
L. Medina ◽  
S. E. Diaz
Keyword(s):  
Neural Network ◽  
Virtual Instrument ◽  
High Speed ◽  
Magnetic Bearing ◽  
Magnetic Bearings ◽  
System Response ◽  
The Neural Network ◽  
Artificial Neurons ◽  
Position Data ◽  
System 2

The use of magnetic bearings in high speed/low friction applications is increasing in industry. Magnetic bearings are sophisticated electromechanical systems, and modeling magnetic bearings using standard techniques is complex and time consuming. In this work a neural network is designed and trained to emulate the operation of a complete system (magnetic bearing, PID controller, and power amplifiers). The neural network is simulated and integrated into a virtual instrument that will be used in the laboratory both as a teaching and a research tool. The main aims in this work are: (1) determining the minimum amount of artificial neurons required in the neural network to emulate the magnetic bearing system, (2) determining the more appropriate ANN training method for this application, and (3) determining the errors produced when a neural network trained to emulate system operation with a balanced rotor is used to predict system response when operating with an unbalanced rotor. The neural network is trained using as input the position data from the proximity sensors; neural network outputs are the control signals to the coil amplifiers.

Controller Design Based on Mechatronic Model

2014 ◽  
Vol 526 ◽  
pp. 351-356
Author(s):  
Li Xi Yue ◽  
Jian Hui Zhou ◽  
Yan Nan Lu ◽  
Chong Chong Ji ◽  
Zhi Yong Yu
Keyword(s):  
Neural Network ◽  
High Speed ◽  
Controller Design ◽  
Design Method ◽  
Nonlinear Modeling ◽  
Digital Design ◽  
The Neural Network ◽  
Key Issues ◽  
Neural Network Algorithm ◽  
Computed Torque

The dissertation deals with some key issues relevant to the controller design and digital design method for a newly patented high-speed parallel manipulator. Meanwhile, a Virtual Prototyping based co-simulation platform is also established according to the ADAMS and Matlab/Simulink software. In order to promote the ability that the manipulator traces the prescribed trajectory, a model based computed torque controller is described in detail, and a neural network algorithm is also used to optimize controller parameters real-timely under the consideration of systematic nonlinear, modeling error and outer disturbance. The neural network based computed torque controller increases the robustness of system dramatically.

The Control Technology of Air-Conditioning Based RBF Adaptive Neural Network

2012 ◽  
Vol 490-495 ◽  
pp. 693-697
Author(s):  
Jie Jia Li ◽  
Hao Chen ◽  
Dai Yan Liu
Keyword(s):  
Neural Network ◽  
Air Conditioning ◽  
High Speed ◽  
Delay System ◽  
Network Control ◽  
Comfort Level ◽  
Control Technology ◽  
Control Effect ◽  
The Neural Network ◽  
Conditioning Control

The central air-conditioning control system is a nonlinear, large inertia, delay system, controller the performance, relates directly to the control effect, energy loss and comfort level. This paper analyzes the working principle and characteristics of the air-conditioning system, and needle air-conditioning control deficiencies, with a prediction neural network control technology, established the neural network predictive controller model. Through the combination of sensor and controller predict, the simulation results show that the neural network predictive control has the characteristic of high speed and high control accuracy, and a strong ability to adapt and so on.

scholarly journals Application Research on Fault Diagnosis of the Filter Unit Based on Intelligent Algorithm of GA and WNN

2015 ◽  
Vol 9 (1) ◽  
pp. 922-926 ◽  
Author(s):  
Zhao Xuejun ◽  
Wang Mingfang ◽  
Wang Jie ◽  
Tong Chuangming ◽  
Yuan Xiujiu
Keyword(s):  
Neural Network ◽  
Fault Diagnosis ◽  
High Speed ◽  
Analog Circuit ◽  
Optimal Solution ◽  
Intelligent Algorithm ◽  
The Neural Network ◽  
Filter Unit ◽  
Band Sequence ◽  
Neural Network Input

This paper focuses on the potential of GA algorithm for adaptive random global search, and WNN resolution as well as the ability of fault tolerance to build a multi intelligent algorithm based on the GA-WNN model using the filter unit of analog circuit for fault diagnosis. Construction of GA-WNN model was divided into two stages; in the first stage GA was used to optimize the initial weights, threshold, expansion factor and translation factor of WNN structure; while in the second stage, initially, based on WNN training and learning, global optimal solution was obtained. In the process of using analog output signal by using wavelet decomposition, the absolute value of coefficient of each frequency band sequence was obtained along with the energy characteristics of the cross joint, with a combination of feature vectors as the input of the neural network. Through the pretreatment method, in order to reduce the neural network input, neural grid size of neurons was reduced in each layer and the convergence speed of neural network was increased. The experimental results show that the method can diagnose single and multiple soft faults of the circuit, with high speed and high precision.

Research on Optimal CDMA Multiuser Detection Based on Stochastic Hopfield Neural Network

2019 ◽  
Vol 12 (3) ◽  
pp. 233-240
Author(s):  
Tongke Fan
Keyword(s):  
Neural Network ◽  
Local Convergence ◽  
High Speed ◽  
Hopfield Neural Network ◽  
Communication Model ◽  
State Variables ◽  
Stochastic Disturbance ◽  
Detection Techniques ◽  
The Neural Network ◽  
Searching Ability

Background: Most of the common multi-user detection techniques have the shortcomings of large computation and slow operation. For Hopfield neural networks, there are some problems such as high-speed searching ability and parallel processing, but there are local convergence problems. Objective: The stochastic Hopfield neural network avoids local convergence by introducing noise into the state variables and then achieves the optimal detection. Methods: Based on the study of CDMA communication model, this paper presents and models the problem of multi-user detection. Then a new stochastic Hopfield neural network is obtained by introducing a stochastic disturbance into the traditional Hopfield neural network. Finally, the problem of CDMA multi-user detection is simulated. Conclusion: The results show that the introduction of stochastic disturbance into Hopfield neural network can help the neural network to jump out of the local minimum, thus achieving the minimum and improving the performance of the neural network.

Sign in / Sign up

Export Citation Format

Share Document