Automated electromagnetic interference filter design

Patent metadata

Field	Value
Publication number	US-10706197-B2
Application number	US-201815987945-A
Country	US
Kind code	B2
Filing date	May 24, 2018
Priority date	May 24, 2018
Publication date	Jul 7, 2020
Grant date	Jul 7, 2020

Abstract

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.

First claim

What is claimed:

1. A system comprising:
one or more processors;
one or more non-transitory computer-readable media storing instructions that, when executed by the one or more processors, program the one or more processors to perform operations comprising:
receiving a plurality of parameters for a filter design, the plurality of parameters including a noise parameter;
determining a plurality of candidate filter configurations that match at least one of the parameters;
determining, 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 amount associated with the candidate filter configuration;
ranking the candidate filter configurations based on the estimated EMI noise amount determined as outputs of the machine learning model for each respective candidate filter configuration;
selecting a subset of highest ranked candidate filter configurations of the candidate filter configurations based on the ranking determined based on the outputs of the machine learning model;
executing respective simulations of respective candidate filter configurations included in the selected subset of highest ranked candidate filter configurations by executing the respective simulations of the respective candidate filter configurations with a simulation of a system in which the respective candidate filter configurations are intended to be used to determine a simulated EMI noise for the respective candidate filter configurations; and
sending information related to at least one of the respective candidate filter configurations to a computing device.

2. The system as recited in claim 1 , wherein the plurality of parameters include at least one of a size of a filter or a cost of the filter.

3. The system as recited in claim 2 , wherein the operation of determining a plurality of candidate filter configurations that match at least one of the parameters further comprises:
traversing a data structure that includes filter configurations including filter topology information and filter component information; and
identifying as one of the candidate filter configurations a respective one of the filter configurations having a size and cost that is less than a size and cost specified in the plurality of parameters.

4. The system as recited in claim 1 , wherein the operation of ranking the candidate filter configurations comprises ranking the candidate filter configurations in order from lowest estimated EMI noise amount to highest estimated EMI noise amount.

5. The system as recited in claim 1 , wherein the operation of selecting the subset of highest ranked candidate filter configurations based on the ranking further comprises selecting K of the candidate filter configurations having the lowest estimated EMI noise amount, wherein K is a number specified by at least one of:
a user who submitted the parameters;
an administrator; or
a system setting.

6. The system as recited in claim 1 , the operations further comprising;
executing a plurality of simulations using a respective plurality of filter configurations to generate simulation training data for the machine learning model; and
training the machine learning model, at least in part with the simulation training data.

7. The system as recited in claim 1 , wherein sending the information related to at least one of the respective candidate filter configurations to the computing device causes, at least partially, the computing device to send the at least one respective candidate filter configuration to an automated manufacturing system to generate a prototype of the at least one respective candidate filter configuration.

8. A method comprising:
receiving, by one or more processors, a plurality of parameters for a filter design, the plurality of parameters including a noise parameter;
determining a plurality of candidate filter configurations based on at least one of the parameters;
determining, 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 the candidate filter configuration;
ranking the candidate filter configurations based on the estimated EMI noise amount determined as outputs of the machine learning model for each respective candidate filter configuration;
selecting a subset of highest ranked candidate filter configurations of the candidate filter configurations based on the ranking determined based on the outputs of the machine learning model;
executing respective simulations of respective candidate filter configurations included in the selected subset of highest ranked candidate filter configurations by executing the respective simulations of the respective candidate filter configurations with a simulation of a system in which the respective candidate filter configurations are intended to be used to determine a simulated EMI noise for the respective candidate filter configurations; and
sending information related to at least one of the respective candidate filter configurations to a computing device.

9. The method as recited in claim 8 , wherein sending the information related to the at least one respective candidate filter configuration includes determining that the simulated EMI noise for the respective candidate filter configuration is less than the noise parameter.

10. The method as recited in claim 8 , wherein ranking the candidate filter configurations based on the estimated EMI noise comprises ranking the candidate filter configurations in order from lowest estimated EMI noise amount to highest estimated EMI noise amount.

11. The method as recited in claim 8 , wherein the plurality of parameters include at least one of a size of a filter or a cost of the filter.

12. The method as recited in claim 11 , wherein determining a plurality of candidate filter configurations based on at least one of the parameters further comprises:
traversing a data structure that includes filter configurations including filter topology information and filter component information; and
identifying as one of the candidate filter configurations a respective one of the filter configurations having a size and cost that is less than a size and cost specified in the plurality of parameters.

13. The method as recited in claim 8 , wherein selecting the subset of highest ranked candidate filter configurations based on the ranking further comprises selecting K of the candidate filter configurations having a lower estimated EMI noise level than remaining candidate filter configurations, wherein K is an integer specified by at least one of:
a user who submitted the parameters;
an administrator; or
a system setting.

14. The method as recited in claim 8 , further comprising
executing a plurality of simulations using a respective plurality of filter configurations to generate simulation training data for the machine learning model; and
training the machine learning model, at least in part with the simulation training data.

15. One or more computer-readable media storing instructions that, when executed by one or more processors, program the one or more processors to perform operations comprising:
receiving, from a computing device, a plurality of parameters for a filter design, the plurality of parameters including a noise parameter;
determining a plurality of candidate filter configurations based on at least one of the parameters;
determining, for each candidate filter configuration of the plurality of

Assignees

• Hitachi Ltd

Inventors

• Li Jia
• Ma Tianye
• Xiao Yuan
• Chen Heming

Classifications

• H02M7/5387

  in a bridge configuration · CPC title

• H02M1/44

  Circuits or arrangements for compensating for electromagnetic interference in converters or inverters · CPC title

• G06F30/367 Primary

  Design verification, e.g. using simulation, simulation program with integrated circuit emphasis [SPICE], direct methods or relaxation methods · CPC title

• G06F30/327

  Logic synthesis; Behaviour synthesis, e.g. mapping logic, HDL to netlist, high-level language to RTL or netlist · CPC title

• G01R31/002

  where the device under test is an electronic circuit · CPC title