Equipment repair management and execution

• US 11,544,676 B2
• Filed: 12/30/2019
• Issued: 01/03/2023
• Est. Priority Date: 09/24/2018
• Status: Active Grant

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Ghosh, Dipanjan, Farahat, Ahmed Khairy, Zhang, Chi, Vieira, Marcos, Gupta, Chetan
• Primary Examiner(s): Mincarelli, Jan P
• Application Number:
  US16/729,657
• Publication Number:
  US 20200134574A1
• Time in Patent Office: 1,100 Days
• Field of Search: None
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06Q 10/0635 Risk analysis of enterprise or organisation activities
• CPC Class Codes: G06Q 10/067 Enterprise or organisation modelling
• CPC Class Codes: G06Q 10/20 Administration of product repair or maintenance

Abstract

In some examples, a computer system may receive historical repair data and may extract features from the historical repair data for use as training data. The computer system may determine, from the historical repair data, a repair hierarchy including a plurality of repair levels which includes repair actions as one of the repair levels. Furthermore, the computer system may train the machine learning model, which performs multiple tasks for predicting values of individual levels of the repair hierarchy, by tuning parameters of the machine learning model using the training data.

Claims

1. A system comprising:

one or more processors; and

one or more non-transitory computer-readable media maintaining executable instructions, which, when executed by the one or more processors, configure the one or more processors to perform operations including;

receiving historical repair data;

extracting features from the historical repair data as training data;

determining, from the historical repair data, a repair hierarchy including a plurality of repair levels which includes one or more repair actions as one of the repair levels;

training a single machine learning model for determining the one or more repair actions, the single machine learning model trained to perform multiple tasks for predicting values for individual repair levels of the repair hierarchy, the single machine learning model trained by tuning parameters of the single machine learning model using the training data, wherein the training the single machine learning model includes;

determining, for each task of the multiple tasks that correspond to each of the individual repair levels of the repair hierarchy, a respective loss within the single machine learning model for each individual repair level of the repair hierarchy, wherein the respective loss indicates an error within the single machine learning model;

combining the respective losses in the single machine learning model determined for each of the individual repair levels of the repair hierarchy to determine a joint loss function; and

optimizing the joint loss function during training of the single machine learning model to update the parameters of the single machine learning model to perform predictions at each of the individual repair levels of the repair hierarchy.

2. The system as recited in claim 1, wherein the single machine learning model is configured to:

determine a repair plan based on the one or more repair actions and a respective probability of success determined for each repair action of the one or more repair actions; and

perform at least one of;

sending an order for a part for a repair;

sending a communication to assign labor to perform the repair;

sending a communication to schedule a repair time for the repair;

or

remotely initiating a procedure on equipment to be repaired to effectuate, at least partially, the repair.

3. The system as recited in claim 1, wherein:

the historical repair data is for equipment, and the plurality of repair levels in the repair hierarchy includes;

one or more systems of the equipment as a first repair level of the repair hierarchy on which the single machine learning model performs a prediction;

one or more subsystems of the one or more systems of the equipment as second repair level of the repair hierarchy on which the single machine learning model performs a prediction; and

the one or more repair actions to perform on the equipment as a third repair level of the repair hierarchy on which the single machine learning model performs a prediction.

4. The system as recited in claim 1, the operations further comprising, during the training of the single machine learning model, employing a backpropagation algorithm to calculate gradients with respect to each learnable variable in the single machine learning model using the joint loss function.

5. The system as recited in claim 1, wherein optimizing the joint loss function during the training of the single machine learning model is conducted by minimizing the joint loss function to minimize, concurrently, the respective losses in the single machine learning model that correspond to the individual repair levels of the repair hierarchy during the training of the single machine learning model.

6. The system as recited in claim 1, wherein the joint loss function is determined based on a sum of a cross entropy loss of each repair level of the repair hierarchy.

7. A method for building a machine learning model for determining one or more repair actions corresponding to a repair request, the method comprising:

receiving, by one or more processors, historical repair data;

extracting features from the historical repair data as training data;

determining, from the historical repair data, a repair hierarchy including a plurality of repair levels which includes the one or more repair actions as one of the repair levels; and

training a single machine learning model as the machine learning model for determining the one or more repair actions, the single machine learning model trained to perform multiple tasks for predicting values for individual repair levels of the repair hierarchy, the single machine learning model trained by tuning parameters of the single machine learning model using the training data, wherein the training the single machine learning model includes;

determining, for each task of the multiple tasks that correspond to each of the individual repair levels of the repair hierarchy, a respective loss within the single machine learning model for each individual repair level of the repair hierarchy, wherein the respective loss indicates an error within the single machine learning model;

combining the respective losses in the single machine learning model determined for each of the individual repair levels of the repair hierarchy to determine a joint loss function; and

optimizing the joint loss function during training of the single machine learning model to update the parameters of the single machine learning model to perform predictions at each of the individual repair levels of the repair hierarchy.

8. The method as recited in claim 7, wherein the single machine learning model is configured to determine a repair plan based on the one or more repair actions and a respective probability of success determined for each repair action of the one or more repair actions.

9. The method as recited in claim 8, wherein the single machine learning model is configured to perform at least one of:

sending an order for a part for a repair;

sending a communication to assign labor to perform the repair;

sending a communication to schedule a repair time for the repair;

or

remotely initiating a procedure on equipment to be repaired to effectuate, at least partially, the repair.

10. The method as recited in claim 7, further comprising, during the training of the single machine learning model, employing a backpropagation algorithm to calculate gradients with respect to each learnable variable in the single machine learning model using the joint loss function.

11. The method as recited in claim 7, wherein optimizing the joint loss function during the training of the single machine learning model is conducted by minimizing the joint loss function to minimize, concurrently, the respective losses in the single machine learning model that correspond to the individual repair levels of the repair hierarchy during the training of the single machine learning model.

12. The method as recited in claim 7, wherein the joint loss function is determined based on a sum of cross entropy loss of each repair level of the repair hierarchy.

20. The method as recited in claim 7, wherein:

the historical repair data is for equipment, and the plurality of repair levels in the repair hierarchy includes;

one or more systems of the equipment as a first repair level of the repair hierarchy on which the single machine learning model performs a prediction;

one or more subsystems of the one or more systems of the equipment as second repair level of the repair hierarchy on which the single machine learning model performs a prediction; and

the one or more repair actions to perform on the equipment as a third repair level of the repair hierarchy on which the single machine learning model performs a prediction.

13. One or more non-transitory computer-readable media storing instructions that, when executed by one or more processors, program the one or more processors to perform operations comprising:

receiving historical repair data;

extracting features from the historical repair data as training data;

determining, from the historical repair data, a repair hierarchy including a plurality of repair levels which includes one or more repair actions as one of the repair levels; and

training a single machine learning model for determining the one or more repair actions, the single machine learning model trained to perform multiple tasks for predicting values for individual repair levels of the repair hierarchy, the single machine learning model trained by tuning parameters of the single machine learning model using the training data, wherein the training the single machine learning model includes;

determining, for each task of the multiple tasks that correspond to each of the individual repair levels of the repair hierarchy, a respective loss within the single machine learning model for each individual repair level of the repair hierarchy, wherein the respective loss indicates an error within the single machine learning model;

combining the respective losses in the single machine learning model determined for each of the individual repair levels of the repair hierarchy to determine a joint loss function; and

optimizing the joint loss function during training of the single machine learning model to update the parameters of the single machine learning model to perform predictions at each of the individual repair levels of the repair hierarchy.

14. The one or more non-transitory computer-readable media as recited in claim 13, wherein the single machine learning model is configured to determine a repair plan based on the one or more repair actions and a respective probability of success determined for each repair action of the one or more repair actions.

15. The one or more non-transitory computer-readable media as recited in claim 14, wherein the single machine learning model is configured to perform at least one of:

sending an order for a part for a repair;

sending a communication to assign labor to perform the repair;

sending a communication to schedule a repair time for the repair;

or

remotely initiating a procedure on equipment to be repaired to effectuate, at least partially, the repair.

16. The one or more non-transitory computer-readable media as recited in claim 13, further comprising, during the training of the single machine learning model, employing a backpropagation algorithm to calculate gradients with respect to each learnable variable in the single machine learning model using the joint loss function.

17. The one or more non-transitory computer-readable media as recited in claim 13, wherein optimizing the joint loss function during the training of the single machine learning model is conducted by minimizing the joint loss function to minimize, concurrently, the respective losses in the single machine learning model that correspond to the individual repair levels of the repair hierarchy during the training of the single machine learning model.

18. The one or more non-transitory computer-readable media as recited in claim 13, wherein the joint loss function is determined based on a sum of cross entropy loss of each repair level of the repair hierarchy.

19. The one or more non-transitory computer-readable media as recited in claim 13, wherein:

the historical repair data is for equipment, and the plurality of repair levels in the repair hierarchy includes;

one or more systems of the equipment as a first repair level of the repair hierarchy on which the single machine learning model performs a prediction;

one or more subsystems of the one or more systems of the equipment as second repair level of the repair hierarchy on which the single machine learning model performs a prediction; and

the one or more repair actions to perform on the equipment as a third repair level of the repair hierarchy on which the single machine learning model performs a prediction.