Aging profiling engine for physical systems

• US 10,403,056 B2
• Filed: 12/07/2015
• Issued: 09/03/2019
• Est. Priority Date: 12/08/2014
• Status: Active Grant

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Yan, Tan, Jiang, Guofei, Chen, Haifeng, Zhang, Kai
• Primary Examiner(s): Suglo, Janet L
• Assistant Examiner(s): Liang, Leonard S
• Application Number:
  US14/961,519
• Publication Number:
  US 20160161374A1
• Time in Patent Office: 1,366 Days
• Field of Search: None
• CPC Class Codes: G05B 23/0232 based on qualitative trend analysis, e.g. system evolution
• CPC Class Codes: G05B 23/0283 Predictive maintenance, e.g. involving the monitoring of a system and, based on the monitoring results, taking decisions on the maintenance schedule of the monitored system; Estimating remaining useful life [RUL] preventive maintenance, i.e. planning maintenance according to the available resources without monitoring the system G06Q10/06
• CPC Class Codes: G07C 3/00 Registering or indicating the condition or the working of machines or other apparatus, other than vehicles

Abstract

Systems and methods for managing components of physical systems, including decomposing raw time series by extracting an aging trend and a fluctuation term from the time series using an objective function of an optimization problem, the objective function minimizing reconstruction error and ensuring flatness of the fluctuation term over time. The optimization problem is transformed into a Quadratic Programming (QP) formulation including a monotonicity constraint and a non-negativity constraint, the constraints being merged together to reduce computational costs. An aging score and a confidence score are generated for the extracted aging trend to determine a severeness of aging for one or more components of the physical system, and the aging score and confidence score are fused to provide a fused ranking for the extracted aging trend for predicting future failures of the components.

Claims

1. A method for managing components of physical systems,

comprising:

decomposing raw time series by extracting an aging trend and a fluctuation term from the time series using an objective function of an optimization problem, the objective function minimizing reconstruction error and ensuring flatness of the fluctuation term over time;

transforming the optimization problem into a Quadratic Programming (QP) formulation including a monotonicity constraint and a non-negativity constraint, the transforming reducing processing requirements and increasing optimization speed by merging the constraints together prior to solving the optimization problem;

generating an aging score and a confidence score for the extracted aging trend to determine a severeness of aging for one or more components of the physical system;

fusing the aging score and confidence score to provide a fused ranking for the extracted aging trend for predicting future failures of the components, and

preventing failure of the physical systems by disabling lower ranked, aged components and rerouting tasks to higher-ranked, less-aged components based on the fused ranking.

2. The method as recited in claim 1, wherein the constraints are merged using a transition matrix.

3. The method as recited in claim 1, wherein the aging score is the slope of the aging trend.

4. The method as recited in claim 1, wherein the confidence score is based on the closeness of a slope of the extracted aging trend to a slope of the raw time series.

5. The method as recited in claim 1, wherein the aging score and confidence score are flexibly tuned using sigmoid functions according to particular application requirements.

6. The method as recited in claim 1, wherein the aging trend is a monotonic sequence representing a change of a mean value of the time series over a particular time period, and the fluctuation term is a stationary component in which a mean value is constant over a particular time period.

7. The method as recited in claim 1, wherein the Quadratic Programming (QP) formulation is transformed into a non-negative QP problem and solved using iterative updates.

8. The method as recited in claim 1, wherein a precise location of a starting point of component degradation is determined based on the ranking.

9. An aging profiling engine for managing components of physical systems, comprising:

a time series transformer for decomposing raw time series by extracting an aging trend and a fluctuation term from the time series using an objective function of an optimization problem, the objective function minimizing reconstruction error and ensuring flatness of the fluctuation term over time;

an optimizer for transforming the optimization problem into a Quadratic Programming (QP) formulation including a monotonicity constraint and a non-negativity constraint, the transforming reducing processing requirements and increasing optimization speed by merging the constraints together prior to solving the optimization problem;

one or more score generators for generating an aging score and a confidence score for the extracted aging trend to determine a severeness of aging for one or more components of the physical system;

a ranker configured to fuse the aging score and confidence score to provide a fused ranking for the extracted aging trend for predicting future failures of the components, and

a controller for preventing failure of the physical systems by disabling lower ranked, aged components and rerouting tasks to higher-ranked, less-aged components based on the fused ranking.

10. The aging profiling engine as recited in claim 9, wherein the constraints are merged using a transition matrix.

11. The aging profiling engine as recited in claim 9, wherein the aging score is the slope of the aging trend.

12. The aging profiling engine as recited in claim 9, wherein the confidence score is based on the closeness of a slope of the extracted aging trend to a slope of the raw time series.

13. The aging profiling engine as recited in claim 9, wherein the aging score and confidence score are flexibly tuned using sigmoid functions according to particular application requirements.

14. The aging profiling engine as recited in claim 9, wherein the aging trend is a monotonic sequence representing a change of a mean value of the time series over a particular time period, and the fluctuation term is a stationary component in which a mean value is constant over a particular time period.

15. The aging profiling engine as recited in claim 9, wherein the Quadratic Programming (QP) formulation is transformed into a non-negative QP problem and solved using iterative updates.

16. The aging profiling engine as recited in claim 9, wherein a precise location of a starting point of component degradation is determined based on the ranking.