System and method for maintenance planning and failure prediction for equipment subject to periodic failure risk

• US 9,058,569 B2
• Filed: 01/28/2013
• Issued: 06/16/2015
• Est. Priority Date: 12/11/2012
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Hosking, Jonathan R., Kalagnanam, Jayant R., Zhu, Yada
• Primary Examiner(s): JONES, HUGH M
• Application Number:
  US13/751,298
• Publication Number:
  US 20140163936A1
• Time in Patent Office: 869 Days
• Field of Search: 703/2
• US Class Current: 1/1
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202
• CPC Class Codes: G06Q 10/0635 Risk analysis of enterprise or organisation activities
• CPC Class Codes: G06Q 10/20 Administration of product repair or maintenance
• CPC Class Codes: Y02P 90/80 Management or planning

Abstract

Systems and methods for failure prediction and maintenance planning are provided. A system for failure prediction and maintenance planning, comprises a statistical modeling module comprising a periodic impact evaluation module capable of identifying periodic effects on the failure risk, a balance equation systems module capable of constructing balance equations with respect to phases of failure times, and an initial phase estimation module capable of estimating an unknown initial phase, wherein one or more of the modules are implemented on a computer system comprising a memory and at least one processor coupled to the memory.

Claims

1. A system for failure prediction for equipment subject to periodically varying failure risk, comprising:

a memory and at least one processor coupled to the memory;

a data center, which executes via the at least one processor and receives collected prior failure data of the equipment in connection with at least one of an environmental condition and an operating condition, wherein the collected prior failure data was collected during a time period excluding a portion of a failure history of the equipment; and

a statistical modeling module, which executes via the at least one processor, is operatively coupled to the data center and comprises;

a periodic impact evaluation module, which executes via the at least one processor and identifies periodic effects on the failure risk corresponding to the time period excluding the portion of the failure history of the equipment;

a balance equation systems module, which executes via the at least one processor and constructs balance equations with respect to phase values of failure times, wherein;

the phase values represent a phase at respective failure times during a time period that the equipment is in operation; and

the phase values are elements of a set corresponding to a duration of a cycle;

an initial phase estimation module, which executes via the at least one processor and estimates, based on the balance equations, an unknown initial phase value at a start of the time period excluding the portion of the failure history of the equipment; and

a failure time distribution module, which executes via the at least one processor and uses the estimated initial phase value when predicting a subsequent failure time distribution of the equipment;

wherein the balance equations are constructed according to the following equation;

2. The system of claim 1, wherein the statistical modeling module further comprises a periodic impact modeling module, which executes via the at least one processor and models periodic patterns using a Fourier series.

3. The system of claim 2, wherein the failure time distribution modeling module models a hazard rate with the periodic patterns and uses the estimated initial phase value when modeling the hazard rate.

4. The system of claim 2, wherein the statistical modeling module further comprises a discrepancy measure module, which executes via the at least one processor and constructs a measure of discrepancy between actual and modeled failure times.

5. The system of claim 4, wherein parameters of the periodic patterns are estimated by evaluating the measure of discrepancy.

6. The system of claim 1, wherein the periodic effects are identified based on a Fourier transformation of failure intensity.

7. The system of claim 1, wherein the periodic effects are due to at least one of the environmental or operational conditions collected during the time period.

8. The system of claim 7, wherein the unknown initial phase value is a phase value of a last failure time before the time period.

9. The system of claim 1, further comprising an optimization module, which executes via the at least one processor, is connected with the statistical modeling module and receives failure prediction data from the statistical modeling module.

10. The system of claim 9, wherein the optimization module optimizes the failure prediction data, and wherein the optimization module outputs the optimized failure prediction data for transmission to at least one of a capital planning module, a spare parts demand module, and a maintenance scheduling module.

11. The system of claim 1, further comprising an optimization module connected with the data center, wherein the optimization module executes via the at least one processor, optimizes data from the data center, and outputs the optimized data for transmission to at least one of a capital planning module, a spare parts demand module, and a maintenance scheduling module.

12. The system of claim 11, wherein the data from the data center includes at least one of spare parts data, supplier data and growth plan data.

13. The system of claim 1, further comprising an optimization module connected with the statistical modeling module and the data center, wherein the optimization module executes via the at least one processor, and performs at least one of dynamic programming, stochastic programming, nonlinear programming, heuristics algorithm processing and Markov Chain Monte Carlo processing on failure prediction data from the statistical modeling module and on data from the data center.

14. An article of manufacture comprising a non-transitory computer readable storage medium comprising program code tangibly embodied thereon, which when executed by a computer, performs method steps for failure prediction for equipment subject to periodically varying failure risk, the method steps comprising:

receiving collected prior failure data of the equipment in connection with at least one of an environmental condition and an operating condition, wherein the collected prior failure data was collected during a time period excluding a portion of a failure history of the equipment;

identifying periodic effects on the failure risk corresponding to the time period excluding the portion of the failure history of the equipment and modeling periodic patterns;

modeling failure times with the periodic patterns;

constructing balance equations with respect to phase values of the failure times, wherein;

the phase values represent a phase at respective failure times during a time period that the equipment is in operation; and

the phase values are elements of a set corresponding to a duration of a cycle;

estimating, based on the balance equations, an unknown initial phase value at a start of the time period excluding the portion of the failure history of the equipment; and

using the estimated initial phase value when predicting a subsequent failure time distribution of the equipment;

wherein the balance equations are constructed according to the following equation;

15. An apparatus for failure prediction for equipment subject to periodically varying failure risk, comprising:

a memory; and

a processor coupled to the memory and configured to execute code stored in the memory for;

receiving collected prior failure data of the equipment in connection with at least one of an environmental condition and an operating condition, wherein the collected prior failure data was collected during a time period excluding a portion of a failure history of the equipment;

identifying periodic effects on the failure risk corresponding to the time period excluding the portion of the failure history of the equipment and modeling periodic patterns;

modeling failure times with the periodic patterns;

constructing balance equations with respect to phase values of the failure times, wherein;

the phase values represent a phase at respective failure times during a time period that the equipment is in operation; and

the phase values are elements of a set corresponding to a duration of a cycle;

estimating, based on the balance equations, an unknown initial phase value at a start of the time period excluding the portion of the failure history of the equipment; and

using the estimated initial phase value when predicting a subsequent failure time distribution of the equipment;

wherein the balance equations are constructed according to the following equation;