Automated EMI Filter Design Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional EMI filter design methods are inefficient and time-consuming, requiring trial and error and multiple iterations to achieve optimal filter configurations that meet EMC standards, often resulting in increased weight, volume, and cost due to the need for extensive simulation and manual redesign.
Innovation Solution
An automated filter design tool utilizing a machine learning model to predict EMI noise levels and streamline the design process by selecting candidate filter configurations based on cost and size parameters, allowing for high-fidelity simulation of a reduced number of configurations and direct prototyping, thereby reducing the time and resources required for determining an optimal filter design.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional trial-and-error design methods are used to achieve optimal EMI filter configurations, then EMI noise attenuation can be improved to meet EMC standards, but design time and iteration cycles increase significantly
Solution Approach 1:
The patent applies preliminary action by pre-generating multiple candidate filter configurations with varying component values before the actual design selection process. The system pre-calculates EMI noise levels for each candidate configuration and stores them in a database, allowing designers to quickly select optimal configurations without time-consuming trial-and-error iterations during the actual design phase.
2Reliability
If multiple iterations of filter design are performed to optimize EMI filtering, then EMI noise suppression can be improved, but device complexity and development resources increase
Solution Approach 1:
The patent applies segmentation by dividing the filter design into multiple candidate configurations with systematically varied component values. Each candidate represents a segmented version of the overall design space, allowing the system to evaluate different segments (configurations) and select the optimal one without analyzing every possible combination, thus reducing design complexity while maintaining optimization capability.
3Reliability
If extensive simulation and manual redesign are performed to achieve optimal filter design, then EMI filtering performance can be improved, but weight, volume, and cost increase
Solution Approach 1:
The patent applies parameter changes by systematically varying component parameters (resistance, capacitance, inductance values) across multiple candidate configurations. The machine learning model predicts EMI noise levels based on these parameter variations, allowing the system to identify optimal parameter combinations that achieve required filtering performance with minimal component values, thereby reducing filter weight, volume, and cost.
4Reliability
If conventional design techniques with multiple iterations are used, then EMI noise attenuation can be improved, but productivity and design efficiency decrease
Solution Approach 1:
The patent applies mechanics substitution by replacing the manual trial-and-error design process with an automated machine learning-based system. The machine learning model predicts EMI noise levels for multiple candidate configurations instantly, eliminating the need for repeated manual simulations and iterations, thus dramatically improving design efficiency and productivity while maintaining or improving EMI noise attenuation performance.
Data Source
AI summary
In some examples, a system may receive a plurality of parameters for a filter design, including a noise parameter. The system may determine a plurality of candidate filter configurations based on at least one of the received parameters. The system may further determine, for each candidate filter configuration of the plurality of candidate filter configurations, based on a trained machine learning model, an estimated electromagnetic interference (EMI) noise associated with each candidate filter configuration. The system may select at least one of the candidate filter configurations based on the estimated EMI noise. In some cases, the system may perform a simulation using the selected candidate filter configuration. Based on the results of the selecting and/or the simulation, the system may send information related to the at least one selected candidate filter configuration to a computing device.