SYSTEM FOR DETERMINING MOST PROBABLE CAUSE OF A PROBLEM IN A PLANT

• US 20100274433A1
• Filed: 04/27/2009
• Published: 10/28/2010
• Est. Priority Date: 04/27/2009
• Status: Active Grant

• Current Assignee: Toyota Motor Corporation
• Original Assignee: Toyota Motor Engineering & Manufacturing North America Incorporated (Toyota Motor Corporation)
• Inventors: Ghimire, Sandesh J., Namburu, Setu Madhavi, Prokhorov, Danil V., Qiao, Liu
• Granted Patent: US 8,209,080 B2
• Time in Patent Office: Days
• US Class Current: 701/29
• CPC Class Codes: G05B 23/024 Quantitative history assessment, e.g. mathematical relationships between available data; Functions therefor; Principal component analysis [PCA]; Partial least square [PLS]; Statistical classifiers, e.g. Bayesian networks, linear regression or correlation analysis; Neural networks
• CPC Class Codes: G05B 23/0281 Quantitative, e.g. mathematical distance; Clustering; Neural networks; Statistical analysis
• CPC Class Codes: G06N 5/042 Backward inferencing

Abstract

A system for determining a most probable cause or causes of a problem in a plant is disclosed. The system includes a plant, the plant having a plurality of subsystems that contribute to the operation of the plant, the plurality of subsystems having operating functions that produce operational signals. A plurality of sensors that are operable to detect the operational signals from the plurality of subsystems and transmit data related to the signals is also provided. An advisory system is disclosed that receives an input, the input being in the form of data from the plurality of sensors, possible input root causes of the problem, possible input symptoms of the problem and/or combinations thereof. The advisory system has an autoencoder in the form of a recurrent neural network. The recurrent neural network has sparse connectivity in a plurality of nodes, and the autoencoder is also operable to receive the input and perform multiple iterations of computations at each of the plurality of nodes as a function of the input and provide an output. The output can be in the form of possible output causes of the problem, possible output symptoms of the problem and/or combinations thereof.

Claims

1. A system for determining a most probable cause or causes of a problem for a plant, said system comprising:

a plant, said plant having a plurality of subsystems that contribute to the operation of the plant, said plurality of subsystems having operating functions that produce operational signals;

a plurality of sensors operable to detect said operational signals from said plurality of subsystems and transmit data related to said operational signals;

an advisory system operable to receive an input, said input selected from the group consisting of data from said plurality of sensors, possible input causes of the problem, possible input symptoms of the problem and combinations thereof;

said advisory system also having an autoencoder in the form of a recurrent neural network (RNN), said RNN having sparse connectivity and a plurality of nodes, said autoencoder also operable to receive said input and perform multiple iterations of computations at each of said plurality of nodes as a function of said input and provide an output;

said output selected from the group consisting of possible output causes of the problem, possible output symptoms of the problem and combinations thereof.

2. The system of claim 1, wherein said plant is a motor vehicle.

3. The system of claim 2, wherein said plurality of subsystems are selected from at least two of the group consisting of a power plant system for the motor vehicle, a drive-train and control system for the motor vehicle, an occupant safety system for the motor vehicle, an occupant comfort system for the motor vehicle and an emissions system for the motor vehicle.

4. The system of claim 1, further comprising said electronic control unit operable to transmit said data from said plurality of sensors to said advisory system.

5. The system of claim 1, further comprising a computer, said computer having said advisory system thereon.

6. The system of claim 1, wherein said input to said advisory system is selected from the group consisting of binary data, continuous data and combinations thereof.

7. The system of claim 6, wherein said input to said advisory system is received from an input preprocessor.

8. The system of claim 7, wherein said input preprocessor converts symbolic and qualitative data into quantitative data using defuzzification.

9. The system of claim 7, wherein said input preprocessor converts at least part of said data from said plurality of sensors to a fault class data or severity level data.

10. The system of claim 7, wherein said input preprocessor converts said data from said plurality of sensors, said possible input causes of the problem and said possible input symptoms of the problem into an input vector having n dimensions.

11. The system of claim 10, wherein said input vector is connected to all of said plurality of nodes within said RNN.

12. The system of claim 10, wherein said autoencoder receives said input vector having n dimensions and produces an output vector having n dimensions.

13. The system of claim 12, wherein said output vector is connected to all of said plurality of nodes within said RNN.

14. The system of claim 1, wherein said input is only said possible input symptoms.

15. The system of claim 14, wherein said output includes said possible output causes.

16. The system of claim 1, wherein said input is only a partial list of said possible input causes for the problem and only a partial list of said possible input symptoms for the problem.

17. The system of claim 16, wherein said output is a complete list of said possible output causes for the problem and a complete list of said possible output symptoms for the problem.

18. A process for determining a most probable cause or causes of a problem for a motor vehicle, the process system comprising:

providing the motor vehicle, the motor vehicle having a plurality of subsystems that contribute to the operation of the motor vehicle, the plurality of subsystems having operating functions that produce operational signals;

providing a plurality of sensors operable to detect the operational signals from the plurality of subsystems and transmit data that is a function of the signals;

providing an input preprocessor, the input preprocessor operable to receive the data from the plurality of sensors, possibly input causes to the problem of the motor vehicle and possibly input symptoms of the problem of the motor vehicle and provide an input vector having n dimensions;

providing a sparsely connected recurrent autoencoder having a plurality of computation nodes and operable to receive an input selected from the group consisting of data from said plurality of sensors, possible input causes of the problem, possible input symptoms of the problem and combinations thereof;

the sparsely connected recurrent autoencoder also operable to receive the input vector and perform multiple iterations of computations at each of the plurality of computation nodes as a function of the input vector and provide an output;

the output selected from the group consisting of possible output causes of the problem for the motor vehicle, possible output symptoms of the problem for the motor vehicle and combinations thereof.

19. The process of claim 18, wherein the output is an output vector of n dimensions.