Method for predicting power generation and remaining useful life per system and system for performing the same

Patent metadata

Field	Value
Publication number	US-11967822-B2
Application number	US-202117489805-A
Country	US
Kind code	B2
Filing date	Sep 30, 2021
Priority date	Dec 11, 2020
Publication date	Apr 23, 2024
Grant date	Apr 23, 2024

Abstract

Embodiments relate to a method for predicting power generation and remaining useful life to predict operational soundness of a power plant and a system for performing the same, the method including acquiring sensing data from each of a plurality of sensors included in a plurality of systems in the power plant, outputting each of a predicted power generation and a predicted remaining useful life from a measurement value in sensing data of an input sensor through a pre-trained prediction model, assessing the operational soundness in aspect of the power generation and the remaining useful life using a prediction result and a current result in each aspect, and determining the operational soundness of the system based on prediction uncertainty and an assessment result in aspect of the power generation and the remaining useful life.

First claim

The invention claimed is:

1. A system for predicting power generation and remaining useful life for each system to predict operational soundness of a power plant, the system comprising:
a plurality of sensors included in a plurality of systems in the power plant;
a data acquisition unit to receive sensing data from each of the plurality of sensors;
a prediction unit to output a predicted power generation from a measurement value in sensing data of an input sensor through a pre-trained power generation prediction model, and output a predicted remaining useful life from a measurement value in sensing data of an input sensor through a pre-trained life prediction model; and
a system assessment unit to assess operational soundness of the system using prediction results and current results in aspect of the power generation and the remaining useful life, wherein the system assessment unit is configured to:
assess the operational soundness in aspect of the power generation for the corresponding system at a target time based on the predicted power generation and the current power generation for each system,
assess the operational soundness in aspect of the remaining useful life for the corresponding system at the target time based on the predicted remaining useful life and a current remaining useful life for each system, and
assess the operational soundness of the corresponding system based on prediction uncertainty and assessment results in aspect of the power generation and the remaining useful life.

2. The system according to claim 1 , wherein the power generation prediction model is configured to calculate the power generation of the corresponding system at a future target time as the predicted power generation when an input dataset including the measurement value of the input sensor is inputted, and
the life prediction model is configured to calculate the remaining useful life of the corresponding system at the future target time as the predicted remaining useful life when an input dataset including the measurement value of the input sensor is inputted.

3. The system according to claim 2 , wherein the prediction unit is further configured to:
calculate prediction uncertainty in the predicted power generation based on the predicted power generation calculated by the power generation prediction model, and
calculate prediction uncertainty in the predicted remaining useful life based on the predicted remaining useful life calculated by the life prediction model.

4. The system according to claim 3 , wherein the prediction unit calculates the prediction uncertainty in the corresponding prediction result based on at least one of intrinsic uncertainty in the data or epistemic uncertainty in the model.

5. The system according to claim 1 , wherein the data acquisition unit is further configured to, for each system,
form an input dataset including the measurement value of the sensor having correlation with the power generation of the system, and
form an input dataset including the measurement value of the sensor having correlation with the remaining useful life of the system.

6. The system according to claim 1 , wherein the system assessment unit is configured to:
calculate a power generation assessment index using a difference between the current power generation and the predicted power generation to assess the operational soundness in aspect of the power generation for the system, and
determine a power generation assessment level based on the power generation assessment index.

7. The system according to claim 1 , wherein the system assessment unit determines a remaining useful life assessment level by comparing the current remaining useful life with the predicted remaining useful life.

8. The system according to claim 1 , wherein in case of false rejection, the system assessment unit determines the assessment level at a highest risk rank as the operational soundness of the system.

9. The system according to claim 1 , wherein the system assessment unit determines the operational soundness of the system on preferential basis of the prediction uncertainty to determine the operational soundness of the system, and
wherein when the operational soundness of the system is determined on preferential basis of the prediction uncertainty, the assessment of the operational soundness of the corresponding system is completed.

10. The system according to claim 1 , wherein the system assessment unit determines the operational soundness of the system further using a difference between the assessment results in aspect of the power generation and the remaining useful life or the risk rank to determine the operational soundness of the system based on the prediction uncertainty and the assessment results in aspect of the power generation and the remaining useful life.

11. A method for predicting power generation and remaining useful life for each system to predict operational soundness of a power plant, performed by a processor, the method comprising:
acquiring sensing data from each of a plurality of sensors included in a plurality of systems in the power plant;
calculating a predicted power generation and a predicted remaining useful life from a measurement value in sensing data of an input sensor through a pre-trained prediction model, and calculating prediction uncertainty in the corresponding prediction result;
assessing the operational soundness in aspect of the power generation using a prediction result and a current result in aspect of the power generation and assessing the operational soundness in aspect of the remaining useful life using a prediction result and a current result in aspect of the remaining useful life; and
determining the operational soundness of the system based on the prediction uncertainty and assessment results in aspect of the power generation and the remaining useful life, wherein the operational soundness of the system is determined on preferential basis of the prediction uncertainty to determine the operational soundness of the system, and wherein when the operational soundness of the system is determined on preferential basis of the prediction uncertainty, the assessment of the operational soundness of the corresponding system is completed.

12. The method for predicting power generation and remaining useful life according to claim 11 , further comprising:
before inputting the sensing data to the prediction model, for each of the plurality of systems, selecting the sensing data of the sensor associated with the power generation of each system and the sensing data of the sensor associated with the remaining useful life of each system.

13. The method for predicting power generation and remaining useful life according to claim 11 , wherein calculating the predicted power generation and the predicted remaining useful life comprises:
calculating the predicted power generation from a measurement value in the sensing data of the input sensor through a pre-trained power generation prediction model;
calculating the prediction uncertainty in the predicted power generation based on the calculated predicted power generation;
calculating the predicted remaining useful life from the measurement value in the sensing data of the input sensor through a pre-trained life prediction model; and
calculating the prediction uncertainty in the predicted remaining useful life based on the calculated predicted remaining useful life.

14. The method for predicting power generation and remaining useful life according to claim 11 , wherein in case that both labeled data associated with the predicted power generation and the predicted remaining

Assignees

• Korea Inst Sci & Tech

Inventors

• Kim Ig Jae
• Choi Heeseung
• Choi Yeji
• Kim Jisoo

Classifications

• H02J2103/30

  Simulating, planning, modelling, reliability check or computer assisted design [CAD] of electric power networks · CPC title

• H02J3/004 Primary

  Generation forecast, e.g. methods or systems for forecasting future energy generation · CPC title

• G05B13/0265

  the criterion being a learning criterion · CPC title

• G05B13/048

  using a predictor · CPC title

• H02J2203/20

  Electricity · mapped topic