ESTIMATION OF CURRENT AND FUTURE MACHINE STATES

• US 20190311273A1
• Filed: 06/18/2019
• Published: 10/10/2019
• Est. Priority Date: 12/21/2016
• Status: Active Grant

• Current Assignee: ABB Schweiz AG (ABB Limited)
• Original Assignee: ABB Schweiz AG (ABB Limited)
• Inventors: Cohen, Andrew, Dix, Marcel
• Granted Patent: US 11,836,636 B2
• Time in Patent Office: Days
• CPC Class Codes: G05B 19/4155 characterised by programme execution, i.e. part programme or machine function execution, e.g. selection of a programme
• CPC Class Codes: G05B 2219/33139 Design of industrial communication system with expert system
• CPC Class Codes: G05B 23/0205 by means of a monitoring system capable of detecting and responding to faults
• 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: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks

Abstract

Disclosed is a computer-implemented method for generating a prediction model. The model can be for use in processing machine event data to predict behavior of a plurality of industrial machines under supervision. The prediction model can be configured to determine current and future states of the industrial machines. The method can include: extracting event features from event codes and structuring the event features into feature vectors; and generating the prediction model by clustering the feature vectors into a plurality of vector clusters, the vector clusters being assigned to respective machine states. The prediction model can be constructed based on event data from a first industrial machine and be applied to control an operating state of a second industrial machine.

Claims

1. A computer-implemented method for generating a prediction model, the model being for use in processing machine event data generated by one or more of a plurality of industrial machines sharing common properties, the method comprising:

receiving an event log comprising a plurality of codes representing events that occurred during operation of at least one of the industrial machines during an observation time interval, the event representations comprising respective time stamps and event codes;

extracting event features from the event codes and structuring the event features into feature vectors;

generating the prediction model by clustering the feature vectors into a plurality of vector clusters, the vector clusters being assigned to respective machine states;

wherein the prediction model is configured to receive, as an input, one or more of the event codes and to provide, as an output, one or more probabilities of the at least one industrial machine transitioning between a first of the machine states to a second of the machine states.

2. The method of claim 1, wherein the event codes comprise character strings.

3. The method of claim 2, wherein the extracted event features describe at least one of:

a frequency of a first event associated with a first of the event codes;

a frequency of the first event and a second event occurring within a predetermine time interval, the second event associated with a second of the event codes;

a frequency of the first event and the second event occurring within a predetermined number of intervening event codes.

4. The method of claim 1, wherein combining the event features into feature vectors comprises one or more of the following processing techniques:

a skip-gram technique, continuous bag of words processing technique, a topic modelling technique, and a pairwise co-occurrence technique.

5. The method of claim 1, wherein clustering the feature vectors comprises one or more of the following processing techniques:

k-means, fuzzy c-means, expectation-maximization clustering, affinity-propagation, density based DBScan, and density-based maximum-margin clustering.

6. The method of claim 1, wherein the prediction model comprises a state sequence model.

7. The method of claim 1, wherein the at least one industrial machine comprises a first industrial machine and the method comprises:

receiving event codes from a second industrial machine of the plurality of industrial machines;

inputting the event codes received from the second industrial machine into the prediction model and outputting, from the prediction model, one or more probabilities of the second industrial machine transitioning between the first and the second machine states.

8. The method of claim 7, comprising controlling a motor of the second industrial machine based on the one or more probabilities of the second industrial machine transitioning between the first and second machine states.

9. The method of claim 7, comprising controlling the second industrial machine based on the one or more probabilities of the second industrial machine transitioning between the first and second machine states.

11. A non-transitory computer-readable medium comprising code configured to cause a computing system comprising one or more computer devices to perform the method of claim 1.

13. A computer system comprising one or more processors configured to perform the method of claim 1.

10. A computer-implemented method for predicting behavior of a first industrial machine of a plurality of industrial machines, the plurality of industrial machines sharing common properties, the method comprising:

generating a prediction model by receiving historical event data comprising event codes from a second industrial machine of the plurality of industrial machines;

extracting event features from the event codes and structuring the event features into feature vectors;

generating a prediction model by clustering the feature vectors into a plurality of vector clusters, the vector clusters being assigned to respective machine states;

receiving, with the prediction model, one or more of the event input codes and outputting, with the prediction model, one or more probabilities of the first industrial machine transitioning between a first of the machine states to a second of the machine states;

adjusting a state of the first industrial machine based on the one or more probabilities.

12. A non-transitory computer-readable medium comprising code configured to cause a computing system comprising one or more computer devices to perform the method of claim 10.

14. A computer system comprising one or more processors configured to perform the method of claim 10.