METHOD AND SYSTEM FOR FAILURE PREDICTION WITH AN AGENT

• US 20110029824A1
• Filed: 07/30/2010
• Published: 02/03/2011
• Est. Priority Date: 08/03/2009
• Status: Active Grant

• Current Assignee: Siemens Healthineers AG (Siemens AG)
• Original Assignee: Siemens AG
• Inventors: Zirkel, Werner, Schöler, Thorsten
• Granted Patent: US 8,291,264 B2
• Time in Patent Office: Days
• US Class Current: 714/47.3
• CPC Class Codes: G05B 23/0224 Process history based detection method, e.g. whereby history implies the availability of large amounts of data

Abstract

A system, a computer program product and a method for failure prediction are implemented on an agent. The agent is installed on a machine to be monitored. The method includes detecting service data on the machine. A reference database is accessed, and a failure pattern is provided. The detected service data are analyzed in view of the provided failure pattern by applying a correlation mechanism, such that a prediction for future failures is generated and depicted as a result.

Claims

1. A computer-implemented method for failure prediction, implemented as an agent, wherein the agent is installed on a machine to be monitored, the computer-implemented method comprising:

detecting service data on the machine;

accessing a reference database, in which reference service data relating to previously detected failures and maintenance related events are stored;

providing a failure pattern;

analyzing the detected service data in view of the failure pattern, the analyzing comprising applying a correlation mechanism; and

generating a prediction for future failures as a result of the analyzed detected service data.

2. The method of claim 1, further comprising generating a notification informing about a future failure and sending the notification to the machine, a service unit or to another device.

3. The method of claim 1, wherein the machine comprises a plurality of sub-modules, the plurality of sub-modules interacting with each other, and

wherein the failure pattern comprises events of all or selected sub-modules and the interactions between all or the selected sub-modules.

4. The method of claim 1, wherein the failure pattern is generated by taking into account machine-related parameters and events and prior failure-related, statistical data.

5. The method of claim 1, wherein the failure pattern is generated by taking into account different time phases, during which previously detected failures and events have occurred and during which machine-related service data are detected or have occurred.

6. The method of claim 1, wherein the detected service data comprise context data with respect to the machine or an actual functionality or a present application of the machine.

7. The method of claim 1, wherein the detected service data comprise:

sensor data including temperature, voltage or speed values or material defects;

computer processing data including resources, memory capacity, processing power or data transfer; and

nominal, time-based event data, generated by sub-modules of the machine, the sub-modules including embedded hardware controllers or robotic devices.

8. The method of claim 1, wherein the correlation mechanism uses a threshold for machine-related parameters or service data, and

wherein the threshold represents a limit for correct operation of the machine.

9. The method of claim 1, wherein the correlation mechanism is based on a rule-based or knowledge-based system, and

wherein the rule-based system is operable to process association models derived from data mining classification algorithms, the data mining classification algorithms including decision tree algorithms, rule-based algorithms, time-based algorithms, or time-series models.

10. The method of claim 1, wherein the correlation mechanism uses statistical or data mining procedures, and

wherein the statistical procedures comprise analysis of frequency distributions, significance tests and a generalized sequential pattern algorithm.

11. The method of claim 1, wherein the generated prediction for future failures is generated for different time frames using a sliding windows approach.

12. The method of claim 1, wherein the machine is a medical apparatus.

13. The method of claim 1, wherein the result is used to pre-plan upcoming service actions with respect to the machine.

14. The method of claim 1, further comprising providing procedures for autonomous behavior of the agent, taking into account performance characteristics and self-adjustment to client applications on the machine or to the machine.

15. The method of claim 1, wherein the method is used for different machines.

16. A system for failure prediction with an agent, wherein the agent is installed on a machine to be monitored, the system comprising:

a detection unit for detecting service data on the machine;

a reference database, in which reference service data relating to previously detected failures and maintenance related events are stored; and

a processing unit configured for;

providing a failure pattern;

analyzing the detected service data in view of the failure pattern as a function of correlation, and

generating a prediction for future failures as a result of the analyzed detected service data.

17. The system of claim 16, wherein the system or the processing unit is configured to provide decision support functions and support consecutive actions triggered by an agent alert, based on the data of the agent.

18. The system of claim 16, wherein the system further comprises a central service unit for failure prediction, the central service unit being in data exchange with a plurality of agents for a plurality of different machines, so that failure patterns are provided from the plurality of different machines.

19. A non-transitory computer program product for failure prediction, implemented on an agent, wherein the agent is installed on a machine to be monitored, wherein the computer program product comprises s computer-readable storage medium having computer-readable program code portions stored therein, the computer-readable program code portions comprising:

detecting service data on the machine;

accessing a reference database, in which reference service data relating to previously detected failures and maintenance related events are stored;

providing at least one failure pattern;

applying a correlation mechanism to the detected service data with the failure pattern; and

generating a prediction for future failures as a result of the analyzed detected service data.