Equipment control based on failure determination

• US 10,466,142 B2
• Filed: 07/13/2016
• Issued: 11/05/2019
• Est. Priority Date: 07/13/2016
• Status: Active Grant

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Hiruta, Tomoaki, Gupta, Chetan
• Primary Examiner(s): Oh, Jaehwan
• Application Number:
  US15/208,662
• Publication Number:
  US 20180017467A1
• Time in Patent Office: 1,210 Days
• CPC Class Codes: G01M 99/008 by doing functionality tests
• CPC Class Codes: G05B 17/02 electric
• CPC Class Codes: G05B 23/0221 Preprocessing measurements, e.g. data collection rate adjustment; Standardization of measurements; Time series or signal analysis, e.g. frequency analysis or wavelets; Trustworthiness of measurements; Indexes therefor; Measurements using easily measured parameters to estimate parameters difficult to measure; Virtual sensor creation; De-noising; Sensor fusion; Unconventional preprocessing inherently present in specific fault detection methods like PCA-based methods
• CPC Class Codes: G05B 23/0254 based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks
• CPC Class Codes: 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

Abstract

In some examples, a computing device may generate simulated data based on a physical model of equipment. For example, the simulated data may include a plurality of probability distributions of a plurality of degrees of failure, respectively, for at least one failure mode of the equipment. In addition, the computing device may receive sensor data indicating a measured metric of the equipment. The computing device may compare the received sensor data with the simulated data to determine a failure mode and a degree of failure of the equipment. At least partially based on the determined failure mode and degree of failure of the equipment, the computing device may send at least one of a notification or a control signal.

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, program the one or more processors to perform operations comprising;

receiving a physical model of an equipment including a brake;

generating simulated data corresponding to braking times for the brake based on the physical model, the simulated data including a plurality of probability distributions of a plurality of degrees of failure, respectively, for at least one failure mode of the brake;

determining a difference between the simulated data for the brake generated based on the physical model and historical data for the brake maintained in a data structure;

receiving sensor data including at least one measured metric indicating a condition of the brake;

normalizing the received sensor data based on the difference between the simulated data and the historical data;

comparing the normalized sensor data with the simulated data to determine a failure mode and a degree of failure of the brake; and

sending a control signal for controlling the equipment based at least in part on the determined failure mode and degree of failure of the brake.

2. The system as recited in claim 1, wherein the operation of determining the difference between the simulated data and the historical data for the brake comprises:

receiving the historical data corresponding to the brake, the historical data including historical sensor data associated with the brake, the historical data further including historical maintenance data indicating respective conditions of the brake corresponding to the historical sensor data;

determining a first probability distribution for the simulated data having a degree of failure categorized as normal operation;

determining a second probability distribution for the historical data having a condition categorized as normal operation; and

determining the difference based on a deviation between the first probability distribution and the second probability distribution.

3. The system as recited in claim 1, the operations further comprising:

receiving the physical model as a candidate physical model generated at least partially based on experimental data corresponding to the brake;

executing a simulation using the candidate physical model to obtain test simulated data;

determining a difference between the test simulated data and the experimental data; and

receiving the physical model based at least partially on updating one or more parameters of the candidate physical model based on determining that the difference between the test simulated data and the experimental data exceeds a threshold.

4. The system as recited in claim 1, wherein comparing the normalized sensor data with the simulated data to determine the degree of failure of the brake comprises:

generating a statistical model based on the simulated data, the statistical model including a plurality of the probability distributions for a plurality of the degrees of failure, respectively, for a plurality of the failure modes of the brake; and

comparing the normalized sensor data with the plurality of the probability distributions to determine a failure mode and degree of failure having a maximum probability.

5. The system as recited in claim 1, the operations further comprising:

comparing the failure mode and the degree of failure of the brake with one or more thresholds to determine a condition of the brake; and

at least partially based on the condition, sending a notification to a computing device, the notification including an indication of the condition.

6. The system as recited in claim 1, wherein the brake is a component of a machine, the operations further comprising:

comparing the failure mode and the degree of failure of the brake with one or more thresholds to determine that wear of the brake exceeds a threshold of acceptable wear; and

at least partially based on determining that wear of the brake exceeds the threshold of acceptable wear, sending the control signal to the machine of which the brake is the component to cause the machine to cease operation.

7. 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, program the one or more processors to;

determine a difference between simulated data generated by a physical model of an equipment and historical data for the equipment;

generate a statistical model based on the simulated data, the statistical model including a plurality of probability distributions for a plurality of degrees of failure, respectively, for at least one failure mode of the equipment;

receive sensor data indicating a measured metric of the equipment;

normalize the received sensor data based on the difference;

based on the statistical model, compare the normalized sensor data with the simulated data to determine a degree of failure of the equipment; and

perform at least one action based on the determined degree of failure.

8. The system as recited in claim 7, wherein performing the at least one action comprises at least one of:

sending a control signal to the equipment for controlling the equipment based on the determined degree of failure;

or

sending a notification to at least one computing device associated with monitoring the equipment, the notification including an indication of the condition of the equipment.

9. The system as recited in claim 7, wherein the one or more processors are further programmed to:

compare the degree of failure of the equipment with one or more thresholds to determine a condition of the equipment; and

determine the at least one action to perform at least partially based on the determined condition.

10. The system as recited in claim 7, wherein the normalized sensor data includes a distribution of sensor measurements, and the one or more processors are further programmed to compare the normalized sensor data with the simulated data to determine a degree of failure of the equipment by:

determining a distance between a normalized distribution of the received sensor data and the plurality of probability distributions for the plurality of degrees of failure to determine relative entropies; and

selecting a degree of failure based on the relative entropies.

11. The system as recited in claim 7, wherein the received sensor data includes an individual sensor measurement, and the one or more processors are further programmed to compare the received sensor data with the simulated data to determine a degree of failure of the equipment by:

determining probability densities of the plurality of probability distributions of the plurality of degrees of failure;

determining posterior probabilities for the probability densities for the plurality of probability distributions, respectively; and

selecting a degree of failure based on a maximum posterior probability.

12. The system as recited in claim 7, wherein at least a first one of the processors is remote from the equipment and/or the sensors, such as over a network, while at least a second one of the processors is co-located with the equipment such connected to communicate directly with the equipment and/or the sensors.

13. An apparatus comprising:

one or more processors; and

a non-transitory computer-readable medium storing executable instructions, which, when executed by the one or more processors, configure the one or more processors to perform operations comprising;

using a physical model of an equipment to generate simulated data of equipment operation;

determining a difference between the simulated data generated from the physical model of the equipment and historical data for the equipment, the historical data including historical sensor data associated with the equipment and historical maintenance data indicating respective conditions of the equipment corresponding to the historical sensor data;

generating a statistical model using at least the simulated data, the statistical model including a plurality of probability distributions for a plurality of degrees of failure, respectively, for at least one failure mode of the equipment;

receiving sensor data indicating a measured metric of the equipment;

normalizing the received sensor data based on the difference between the simulated data and the historical data;

based on the statistical model, comparing the normalized sensor data with the simulated data to determine a degree of failure of the equipment; and

performing at least one action based on the determined degree of failure.

14. The apparatus as recited in claim 13, wherein the operation of performing the at least one action comprises at least one of:

sending a control signal to the equipment for controlling the equipment based on the determined degree of failure;

or

sending a notification to at least one computing device associated with monitoring the equipment, the notification including an indication of the condition of the equipment.

15. The apparatus as recited in claim 13, wherein the operation of determining the difference between the simulated data and the historical data for the equipment comprises:

receiving the historical data corresponding to the equipment from a data structure maintained in the non-transitory computer-readable medium;

determining a first probability distribution for the simulated data having a degree of failure categorized as normal operation;

determining a second probability distribution for the historical data having a condition categorized as normal operation; and

determining the difference based on a deviation between the first probability distribution and the second probability distribution.

16. The apparatus as recited in claim 13, the operations further comprising:

receiving the physical model as a candidate physical model generated at least partially based on experimental data corresponding to the equipment;

executing a simulation using the candidate physical model to obtain test simulated data;

determining a difference between the test simulated data and the experimental data; and

receiving the physical model based at least partially on updating one or more parameters of the candidate physical model based on determining that the difference between the test simulated data and the experimental data exceeds a threshold.

17. The apparatus as recited in claim 13, wherein the statistical model includes a plurality of the probability distributions for a plurality of the degrees of failure, respectively, for a plurality of the failure modes of the equipment, the operations further comprising comparing the normalized sensor data with the plurality of the probability distributions to determine a failure mode and degree of failure having a maximum probability.

18. The apparatus as recited in claim 13, wherein the normalized sensor data includes a distribution of sensor measurements, and the one or more processors are further programmed to compare the normalized sensor data with the simulated data to determine a degree of failure of the equipment by:

determining a distance between a normalized distribution of the received sensor data and the plurality of probability distributions for the plurality of degrees of failure to determine relative entropies; and

selecting a degree of failure based on the relative entropies.

19. The apparatus as recited in claim 13, wherein the equipment is a component of a machine, the operations further comprising:

comparing the failure mode and the degree of failure of the equipment with one or more thresholds to determine that wear of the equipment exceeds a threshold of acceptable wear; and

at least partially based on determining that wear of the equipment exceeds the threshold of acceptable wear, performing the at least one action by sending a control signal to the machine of which the equipment is the component to cause the machine to cease operation.