# US 2021/0342244 A1 · Distributed architecture for fault monitoring

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0342244 A1
- Document ID: US20210342244A1
- Status: Published application
- Kind: A1
- Application number: 17/378,230
- Filed: 2021-07-16
- Published: 2021-11-04
- Earliest claimed: 2018-03-09
- Assignees: Toyota Motor Engineering & Manufacturing North America, Inc.; University of Connecticut
- Applicants: Toyota Motor Engineering & Manufacturing North America, Inc.; University of Connecticut
- Inventors: Ercan M. Dede; Shallesh N. Joshi; Lingyl Zhang; Weiqiang Chen; Krishna Pattipatti; Ali M. Bazzi
- Web page: https://patentgrep.com/patent/US20210342244A1

## CPC

- G06F 11/0739: {in a data processing system embedded in automotive or aircraft systems}
- G06F 11/3062: {where the monitored property is the power consumption (power management in a computing system G06F1/3203)}
- G06F 11/0706: {the processing taking place on a specific hardware platform or in a specific software environment}
- G06F 11/0778: {Dumping, i.e. gathering error/state information after a fault for later diagnosis}
- G06F 11/079: {Root cause analysis, i.e. error or fault diagnosis (in a hardware test environment G06F11/22; in a software test environment G06F11/36)}
- G06F 11/3013: {where the computing system is an embedded system, i.e. a combination of hardware and software dedicated to perform a certain function in mobile devices, printers, automotive or aircraft systems (testing or monitoring of control systems or parts thereof G05B23/02)}
- G06F 11/3055: {Monitoring arrangements for monitoring the status of the computing system or of the computing system component, e.g. monitoring if the computing system is on, off, available, not available (error or fault processing without redundancy G06F11/0703; error detection or correction by redundancy in data representation G06F11/08; error detection or correction of the data by redundancy in operations G06F11/14; error detection or correction by redundancy in hardware G06F11/16)}
- G06F 17/18: for evaluating statistical data {, e.g. average values, frequency distributions, probability functions, regression analysis (forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04)}
- G06N 20/00: Machine learning
- H04W 4/38: for collecting sensor information
- G06F 11/3452: {Performance evaluation by statistical analysis}
- Y02D 10/00: Energy efficient computing, e.g. low power processors, power management or thermal management
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- H04W: WIRELESS COMMUNICATION NETWORKS (broadcast communication H04H; communication systems using wireless links for non-selective communication, e.g. wireless extensions H04M1/72)
- Y02D: CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE

## Abstract

Systems and methods for detecting an anomaly in a power semiconductor device are disclosed. A system includes a server computing device and one or more local components communicatively coupled to the server computing device. Each local component includes sensors positioned adjacent to the power semiconductor device for sensing properties thereof. Each local component receives data corresponding to one or more sensed properties of the power semiconductor device from the sensors and transmits the data to the server computing device. The server computing device utilizes the data, via a machine learning algorithm, to generate a set of eigenvalues and associated eigenvectors and select a selected set of eigenvalues and associated eigenvectors. Each local component conducts a statistical analysis of the selected set of eigenvalues and associated eigenvectors to determine that the data is indicative of the anomaly.

## Claims

### Claim at position 1

-20. (canceled)

### Claim 21

A distributed architecture system, comprising:
a power semiconductor device that, during a course of operation, exhibits properties comprising a device cycling current, a change in device temperature, a normalized junction-to-ambient thermal resistance, a power step, a device maximum junction temperature, a device minimum junction temperature, a Vcold measurement, a Vhot measurement, and a Von measurement; and
a local component comprising one or more sensors positioned adjacent to the power semiconductor device for sensing one or more of the properties of the power semiconductor device and generating sensed data corresponding thereto, wherein the local component designates a first portion of the sensed data as testing data and designates a second portion of the sensed data as training data, transmits the training data to a server and retains the testing data, receives a selected set of eigenvalues and associated eigenvectors generated by the server based on the properties, and conducts a statistical analysis of the selected set of eigenvalues and associated eigenvectors to determine that the testing data is indicative of an anomaly.

### Claim 22 (depends on 21)

The distributed architecture system of claim 21, wherein the local component further generates a reduced data matrix from the selected set of eigenvalues and associated eigenvectors and conducts the statistical analysis from the reduced data matrix.

### Claim 23 (depends on 21)

The distributed architecture system of claim 21, wherein the local component conducts the statistical analysis from a reduced data matrix generated by the server from the selected set of eigenvalues and associated eigenvectors.

### Claim 24 (depends on 21)

The distributed architecture system of claim 21, wherein the selected set of eigenvalues and associated eigenvectors are generated using one or more of a machine learning algorithm, a fault diagnostic algorithm, and a fault prognostic algorithm.

### Claim 25 (depends on 21)

The distributed architecture system of claim 21, wherein the selected set of eigenvalues and associated eigenvectors are generated using a principal component analysis algorithm.

### Claim 26 (depends on 21)

The distributed architecture system of claim 21, wherein the local component determines that the testing data is indicative of the anomaly when the statistical analysis indicates a trend towards a threshold.

### Claim 27 (depends on 21)

The distributed architecture system of claim 21, wherein the local component and the one or more sensors are located in a vehicle, a robot, a mobility system, a manufacturing system, or a utility system.

### Claim 28 (depends on 21)

The distributed architecture system of claim 21, wherein the statistical analysis is a T2 and Q statistical analysis.

### Claim 29 (depends on 21)

The distributed architecture system of claim 21, wherein the local component conducts the statistical analysis from a reduced data matrix generated by the server from the selected set of eigenvalues and associated eigenvectors.

### Claim 30

A method of detecting an anomaly in a power semiconductor device, the method comprising:
receiving sensed data from one or more sensors positioned adjacent to the power semiconductor device, the sensed data corresponding to sensed properties of the power semiconductor device comprising one or more of a device cycling current, a change in device temperature, a normalized junction-to-ambient thermal resistance, a power step, a device maximum junction temperature, a device minimum junction temperature, a Vcold measurement, a Vhot measurement, and a Von measurement;
designating a first portion of the sensed data as testing data and a second portion of the sensed data as training data;
transmitting the training data to a server that is remotely located from the processing device and retaining the testing data;
receiving processed data from the server, the processed data comprising information generated as a result of an application of an algorithm to the sensed data; and
statistically analyzing the processed data to determine whether the retained testing data is indicative of the anomaly.

### Claim 31 (depends on 30)

The method of claim 30, wherein receiving the processed data comprises receiving a selected set of eigenvalues and associated eigenvectors based on the properties from the server.

### Claim 32 (depends on 31)

The method of claim 31, further comprising:
generating a reduced data matrix from the selected set of eigenvalues and associated eigenvectors.

### Claim 33 (depends on 32)

The method of claim 32, wherein statistically analyzing the processed data comprises analyzing the reduced data matrix to determine whether the testing data is indicative of the anomaly.

### Claim 34 (depends on 31)

The method of claim 31, wherein the selected set of eigenvalues and associated eigenvectors are generated using one or more of a machine learning algorithm, a fault diagnostic algorithm, a the fault prognostic algorithm.

### Claim 35 (depends on 31)

The method of claim 31, wherein the selected set of eigenvalues and associated eigenvectors are generated using a principal component analysis algorithm

### Claim 36 (depends on 31)

The method of claim 31, wherein statistically analyzing the selected set of eigenvalues and associated eigenvectors to determine whether the testing data is indicative of the anomaly comprises determining whether statistical analysis results indicate a trend towards a threshold.

### Claim 37 (depends on 30)

The method of claim 30, further comprising:
receiving a reduced data matrix generated by the server.

### Claim 38

A distributed architecture system, comprising:
a power semiconductor device that, during a course of operation, exhibits properties comprising a device cycling current, a change in device temperature, a normalized junction-to-ambient thermal resistance, a power step, a device maximum junction temperature, a device minimum junction temperature, a Vcold measurement, a Vhot measurement, and a Von measurement;
a local component comprising one or more sensors positioned adjacent to the power semiconductor device for sensing one or more of the properties of the power semiconductor device and generating sensed data corresponding thereto; and
a server communicatively coupled to the local component,
wherein:
the local component:
designates a first portion of the sensed data as testing data and designates a second portion of the sensed data as training data, and
transmits the training data to the server and retains the testing data,
the server:
generates a set of eigenvalues and associated eigenvectors based on the properties by applying one or more of a machine learning algorithm, a fault diagnostic algorithm, and a fault prognostic algorithm to the training data, and
selects a selected set of eigenvalues and associated eigenvectors, and
the local component conducts a statistical analysis of the selected set of eigenvalues and associated eigenvectors to determine that the testing data is indicative of the anomaly.

### Claim 39 (depends on 38)

The distributed architecture system of claim 38, wherein the one or more of the machine learning algorithm, the fault diagnostic algorithm, and the fault prognostic algorithm is a principal component analysis algorithm.

### Claim 40 (depends on 38)

The distributed architecture system of claim 38, wherein the local component further generates a reduced data matrix from the selected set of eigenvalues and associated eigenvectors and conducts the statistical analysis from the reduced data matrix.
