An artificial intelligence-based electromagnetic simulation method and its electromagnetic brain
A technology of electromagnetic simulation and artificial intelligence, applied in design optimization/simulation, biological neural network model, neural architecture, etc., to achieve high precision
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Publication Date
- 2021-12-07
Smart Images

Figure 1 
Figure 2 
Figure 3
Abstract
Description
technical field
[0001] The invention belongs to the technical field of electromagnetic simulation, and relates to an artificial intelligence-based electromagnetic simulation method and its electromagnetic brain. Background technique
[0002] The third wave of information driven by wireless communication, mobile portable and Internet of Things, etc., wireless and mobile portable communication are its fundamental features, and radio frequency integrated circuits are its most critical core technology. The core basic hardware of modern high-speed wireless communication is radio frequency integrated circuit chips. As the information society enters the era of 5G communication and cloud computing, the market demand for radio frequency integrated circuit chips will continue to grow rapidly in the future, which increases the demand for radio frequency circuit design and simulation software. demand.
[0003] Document 1 (from C.C.Weng, J.J.Li, Overview of Large-Scale Computing: The Pa...
Examples
Embodiment Construction
[0028] The technology will be further described in detail below in conjunction with the accompanying drawings and examples of implementation.
[0029] Such as figure 1 As shown, the artificial intelligence-based electromagnetic simulation system of the present invention includes: an offline training module and an ultra-efficient electromagnetic analysis module.
[0030] The offline training module imports the training data set obtained from the data server into the convolutional neural network for offline training. After training, the optimal set of weights and bias parameters of the neural network is saved to provide the scattering S parameters for predicting the new structure. . The training data set includes geometry, physics, excitation information data and scattering S-parameter data, where the scattering S-parameters are obtained by calculating the full-wave electromagnetic calculation solver through three types of data: geometry, physics, and excitation.
[0031] The ...