# US 11,306,705 B2 · System and method for monitoring a device

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

- Number: US 11,306,705 B2
- Document ID: US11306705B2
- Status: Granted
- Kind: B2
- Application number: 16/592,815
- Filed: 2019-10-04
- Published: 2021-04-08
- Granted: 2022-04-19
- Earliest claimed: 2019-10-04
- Assignees: City University of Hong Kong
- Applicants: City University of Hong Kong
- Inventors: Xin Liu; Zijun Zhang
- Web page: https://patentgrep.com/patent/US11306705B2

## CPC

- F03D 17/00: Monitoring or testing of wind motors, e.g. diagnostics (testing during commissioning of wind motors F03D13/30)
- G05B 23/0254: {based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks}
- G06N 3/045: Combinations of networks
- G06N 3/0455: Auto-encoder networks; Encoder-decoder networks
- G06N 3/0499: Feedforward networks
- G06N 3/08: Learning methods
- G06N 3/082: modifying the architecture, e.g. adding, deleting or silencing nodes or connections
- G06N 3/09: Supervised learning
- G06N 3/096: Transfer learning
- G06N 3/0985: Hyperparameter optimisation; Meta-learning; Learning-to-learn
- F05B 2260/80: Diagnostics
- F05B 2260/84: Modelling or simulation
- F05B 2270/709: with neural networks
- G06N 3/084: Backpropagation, e.g. using gradient descent
- F03D: WIND MOTORS
- G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- F05B: INDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G

## Abstract

A system and a method for monitoring a device including the steps of obtaining operation information from a device, wherein the operation information is associated with the condition of the device in operation; and processing the operation information with a device modelling engine to determine one or more operation conditions of the device.

## Claims

### Claim 1

A method for monitoring a device comprising the steps of:
obtaining operation information from a device, wherein the operation information is associated with the condition of the device in operation; and
processing the operation information with a device-specific matching network in a device modelling engine to determine or predict one or more operation conditions of the device; wherein the device modelling engine includes a plurality of device-specific matching networks each arranged for processing operation information associated with a respective one of the plurality of devices, the plurality of device-specific matching networks including the device-specific matching network; wherein each of the plurality of device-specific matching networks is generated by:
training of a base matching network with a domain data set including operation information of the plurality of devices to generate a pre-trained base matching network, and
training of the pre-trained base matching network with a target data set including operation information of the respective device.

### Claim 2 (depends on 1)

The method for monitoring the device in accordance with claim 1, wherein the based matching network is a deep neural network.

### Claim 3 (depends on 2)

The method for monitoring the device in accordance with claim 2, wherein the structure of the deep neural network is adjusted based on a stopping criterion.

### Claim 4 (depends on 3)

The method for monitoring the device in accordance with claim 3, wherein the stopping criterion is determined based on a reconstruction error obtained from the deep neural network.

### Claim 5 (depends on 4)

The method for monitoring the device in accordance with claim 4, wherein the reconstruction error is obtained by inputting the domain data set into the deep neural network.

### Claim 6 (depends on 5)

The method for monitoring the device in accordance with claim 5, wherein the structure of the deep neural network is adjusted by inserting one or more hidden layers into the deep neural network.

### Claim 7 (depends on 6)

The method for monitoring the device in accordance with claim 6, wherein the one or more hidden layers inserted by adding one or more hidden nodes into the hidden layers.

### Claim 8 (depends on 7)

The method for monitoring the device in accordance with claim 7, wherein weights or biases of the deep neural network are manipulated when the deep neural network is adjusted.

### Claim 9 (depends on 1)

The method for monitoring the device in accordance with claim 1, wherein the device is a wind turbine.

### Claim 10

A system for monitoring a device comprising:
a device conditions gateway arranged to obtain operation information from a device, wherein the operation information is associated with the condition of the device in operation; and
a conditions monitoring engine arranged to process the operation information with a device-specific matching network in a device modelling engine to determine or predict one or more operation conditions of the device; wherein the device modelling engine includes a plurality of device-specific matching networks each arranged for processing operation information associated with a respective one of the plurality of devices, the plurality of device-specific matching networks including the device-specific matching network; wherein each of the plurality of device-specific matching networks is generated by:
training of a base matching network with a domain data set including operation information of the plurality of devices to generate a pre-trained base matching network, and
training of the pre-trained base matching network with a target data set including operation information of the respective device.

### Claim 11 (depends on 10)

The system for monitoring the device in accordance with claim 10, wherein the based matching network is a deep neural network.

### Claim 12 (depends on 11)

The system for monitoring the device in accordance with claim 11, wherein the structure of the deep neural network is adjusted based on a stopping criterion.

### Claim 13 (depends on 12)

The system for monitoring the device in accordance with claim 12, wherein the stopping criterion is determined based on a reconstruction error obtained from the deep neural network.

### Claim 14 (depends on 13)

The system for monitoring the device in accordance with claim 13, wherein the reconstruction error is obtained by inputting the domain data set into the deep neural network.

### Claim 15 (depends on 14)

The system for monitoring the device in accordance with claim 14, wherein the structure of the deep neural network is adjusted by inserting one or more hidden layers into the deep neural network.

### Claim 16 (depends on 15)

The system for monitoring the device in accordance with claim 15, wherein the one or more hidden layers inserted by adding one or more hidden nodes into the hidden layers.

### Claim 17 (depends on 16)

The system for monitoring the device in accordance with claim 16, wherein weights or biases of the deep neural network are manipulated when the deep neural network is adjusted.

### Claim 18 (depends on 17)

The system for monitoring the device in accordance with claim 17, wherein the device is a wind turbine.

### Claim 19 (depends on 10)

The system for monitoring the device in accordance with claim 10, further comprising a pre-processor arranged to process the operation information obtained from the device conditions gateway.

### Claim 20 (depends on 19)

The system for monitoring the device in accordance with claim 19, wherein the operation information is pre-processed before the information is inputted to the device modelling engine.

### Claim 21 (depends on 20)

The system for monitoring the device in accordance with claim 20, wherein the operation information is pre-processed to remove fault data from the operation information.

### Claim 22 (depends on 10)

The system for monitoring the device in accordance with claim 10, wherein the device conditions gateway is a SCADA system.