Distribution network maintenance planning

• US 8,799,042 B2
• Filed: 08/08/2011
• Issued: 08/05/2014
• Est. Priority Date: 08/08/2011
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Hampapur, Arun, Kalagnanam, Jayant R., Yashchin, Emmanuel, Zhu, Yada
• Primary Examiner(s): Sterrett, Jonathan G
• Assistant Examiner(s): Anderson, Folashade
• Application Number:
  US13/205,190
• Publication Number:
  US 20130041705A1
• Time in Patent Office: 1,093 Days
• Field of Search: 705/7.12, 705/7.41
• US Class Current: 705/7.12
• CPC Class Codes: G06Q 10/06 Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
• CPC Class Codes: G06Q 10/10 Office automation; Time management

Abstract

A system, method and computer program product for maintaining an infrastructure of components. The system receives structured data, unstructured data, and infrastructure data from a database. The system runs at least one statistical and optimization modeler with one or more of: the received structured data, the received unstructured data and the received infrastructure data, in order to calculate a health index of at least one component of the infrastructure. The health index represents a health attribute of the at least one component. The system establishes at least one maintenance plan of the infrastructure, based on the calculated health index. Each established maintenance plan is associated with at least one health index. The system compares health indices of the at least one established maintenance plan. The system selects a plan, among the at least one established maintenance plan, whose health index is a maximum among the compared health indices.

Claims

1. A method for maintaining an infrastructure of components, the method comprising:

receiving, by a processor device, structured data, unstructured data, and infrastructure data from a database;

running, by said processor device, at least one statistical and optimization modeler with one or more of;

the received structured data, the received unstructured data and the received infrastructure data, in order to calculate a health index of at least one component of the infrastructure, the health index representing a health attribute of the at least one component, the calculation of the health index of the at least one component of the infrastructure including;

HI (health index)=−

(F×

R+W)×

1, where F is a vector representing failure probabilities of the at least one component in the infrastructure, R represents a spatial correlation value of the at least one component and the W represents a time elapsed since a last maintenance of the at least one component, and 1 represents a vector of ones, wherein the vector F is a function of λ

exp (Σ

_i=1^kβ

_iZ_i), where λ

is any positive real number, Z_i is a covariate, i is an index, k is a maximum number of the index, and β

_i is a variable representing one or more of;

an age of the at least one component, a diameter of the at least one component, a length of the at least one component, a number of previous failures of the at least one component;

establishing at least one maintenance plan of the infrastructure, based on the calculated health index, each established maintenance plan being associated with at least one health index;

comparing health indices of the at least one established maintenance plan; and

selecting a plan, among the at least one established maintenance plan, whose health index is a maximum among the compared health indices,

wherein a computing system including at least one memory device and said processor device performs the establishing, the comparing, and the selecting.

2. The method according to claim 1, wherein the statistical and optimization modeler includes one or more of:

a statistical modeler, a nonlinear optimizer, a stochastic optimizer, and a heuristic optimizer.

3. The method according to claim 2, wherein the established maintenance plan specifies one or more of:

a subset of components, a timing, intervention actions, and a resource requirement, the subset of components representing parts that need an inspection or repair, the timing representing when the inspection or repair is performed, the intervention actions representing actions that can be taken to perform the inspection or repair, the resource requirement representing materials and equipment needed to perform the inspection or repair.

4. The method according to claim 3, further comprising:

specifying, according to the selected plan, at least one determined action and at least one undetermined action;

implementing the at least one determined action;

re-calculating the health index after the implementation of the at least one determined action;

comparing the re-calculated health index against a pre-determined threshold; and

determining whether to implement the at least one undetermined action based on the comparison.

5. The method according to claim 4, further comprising:

determining at least one constraint associated with performing maintenance of the infrastructure;

determining a long-term maintenance plan associated with the infrastructure, based on the calculated health index of components in the infrastructure and the determined constraint;

estimating a cost of the determined long-term maintenance plan;

determining a first short-term maintenance plan associated with the infrastructure, the determined first short-term maintenance plan specifying one or more of;

the subset of components, the timing, the intervention actions, and the resource requirement;

performing an action, among the intervention actions, specified in the determined first short-term maintenance plan;

updating the determined constraint after performing the specified action;

re-calculating the health index of the components in the infrastructure after performing the specified action; and

repeating the step of determining the constraint, the step of determining the long-term maintenance plan, the step of estimating the cost, the step of determining the first short-term maintenance plan, the step of performing, the step of updating, and the step of re-calculating.

6. The method according to claim 5, wherein the constraint includes one or more of:

an availability of spare parts, labor cost, cost of the spare parts, equipment cost, an availability of labor, a traffic condition, a weather condition, a safety condition, an availability of equipment, and a population density.

7. The method according to claim 5, wherein the determining the long-term maintenance plan includes:

aggregating pre-determined short-term maintenance plans of the infrastructure, based on the determined constraint and the calculated health index, wherein the determined long-term maintenance plan specifies a group of components to maintain in each short-term.

8. The method according to claim 5, wherein the estimating the cost includes:

adding a labor cost for the determined long-term maintenance plan and equipment and material cost for the determined long-term maintenance plan.

9. The method according to claim 5, wherein the determining the first short-term maintenance plan includes:

running the stochastic optimizer with cost parameters, a scenario tree, the determined constraint, a current situation associated with the infrastructure, the calculated health index, and the determined long-term plan and outputting, from the stochastic optimizer, the first short-term maintenance plan.

10. The method according to claim 9, wherein the scenario tree lists the intervention actions under the calculated health index, the scenario tree including a plurality of nodes, a root node in the scenario tree representing the calculated health index, a first level in the scenario tree including at least one node representing at least one action that can be taken under the calculated health index, a second level of the scenario tree including at least one node representing a health index calculated after an action at the first level of the scenario tree is performed, a third level of the scenario tree including at least one node representing at least one action that can be taken under the health index at the second level of the scenario tree, a leaf node in the scenario tree representing spare parts and a workforce needed to perform an action or option described in a corresponding parent node.

11. The method according to claim 10, further comprising:

determining a second short-term maintenance plan after performing the action according to the determined first short-term maintenance plan; and

repeating the step of determining the first short-term maintenance plan, the step of performing the action, and the step of the determining the second short-term maintenance plan.

12. The method according to claim 11, wherein the step of determining the second short-term maintenance plan includes:

running the nonlinear optimizer with the cost parameters, the scenario tree, the re-calculated health index, the updated constraint, the current situation associated with the infrastructure, and the determined first long-term plan and outputting, from the nonlinear optimizer, the second short-term maintenance plan.

13. The method according to claim 5, wherein the step of specifying further comprises:

generating a work schedule and a work contract governing the intervention actions of workers that provide maintenances on the infrastructure based on the determined first short-term maintenance plan, the generated work schedule specifying one or more of;

a list of components to inspect or repair, actions to be taken to inspect or repair the components in the list, an estimated time to conduct the inspection or repair, an estimated time that the inspection or repair is completed, required materials and equipments to perform the inspection or repair, the generated work contract specifying one or more of;

types of skilled workers needed to perform the inspection or repair, an estimation of a number of work hours needed per each type of skilled workers until the inspection or repair is completed, a labor cost per each skilled worker needed, an additional cost associated with the at least one undetermined action, and an estimated cost to perform an initial investigation of the components in the list.

14. The method according to claim 1, wherein the structured data includes one or more of:

a physical property of the component, an operational condition of the component, an environmental factor of the component, a maintenance history record of the component, parts and suppliers of the component, and a human activity associated with the component.

15. The method according to claim 1, wherein the infrastructure data includes one or more of:

a service reliability of the infrastructure, an availability of the infrastructure, a status of deterioration of the infrastructure, an intervention history of the infrastructure, and an impact of a failure of the infrastructure.

16. The method according to claim 1, wherein the unstructured data includes one or more of:

expert knowledge associated with the component, engineering knowledge associated with the component, equipment information of the component, manufacturer information of the component, an installation date of the component, and an operating condition of the component.

17. The method according to claim 1, wherein the health index is computed as a mathematical function of the structured data, the infrastructure data, and the unstructured data.

18. A system for maintaining an infrastructure of components, the system comprising:

at least one memory device;

at least one processor connected to the memory device,

wherein the processor is configured to;

receive structured data, unstructured data, and infrastructure data from a database;

run at least one statistical and optimization modeler with one or more of;

the received structured data, the received unstructured data and the received infrastructure data, in order to calculate a health index of at least one component of the infrastructure, the health index representing a health attribute of the at least one component, the calculation of the health index of the at least one component of the infrastructure including;

HI (health index)=−

(F×

R+W)×

1, where F is a vector representing failure probabilities of the at least one component in the infrastructure, R represents a spatial correlation value of the at least one component and the W represents a time elapsed since a last maintenance of the at least one component, and 1 represents a vector of ones, wherein the vector F is a function of λ

exp (Σ

_i=1^kβ

_iZ_i), where λ

is any positive real number, Z_i is a covariate, i is an index, k is a maximum number of the index, and β

_i is a variable representing one or more of;

an age of the at least one component, a diameter of the at least one component, a length of the at least one component, a number of previous failures of the at least one component;

establish at least one maintenance plan of the infrastructure, based on the calculated health index, each established maintenance plan being associated with at least one health index;

compare health indices of the at least one established maintenance plan; and

select a plan, among the at least one established maintenance plan, whose health index is a maximum among the compared health indices.

19. The system according to claim 18, wherein the statistical and optimization modeler includes one or more of:

a statistical modeler, a nonlinear optimizer, a stochastic optimizer, and a heuristic optimizer.

20. The system according to claim 19, wherein the established maintenance plan specifies one or more of:

a subset of components, a timing, intervention actions, and a resource requirement, the subset of components representing parts that need an inspection or repair, the timing representing when the inspection or repair is performed, the intervention actions representing actions that can be taken to perform the inspection or repair, the resource requirement representing materials and equipment needed to perform the inspection or repair.

21. The system according to claim 20, wherein the processor is further configured to:

specify, according to the selected plan, at least one determined action and at least one undetermined action;

implement the at least one determined action;

re-calculate the health index after the implementation of the at least one determined action;

compare the re-calculated health index against a pre-determined threshold; and

determine whether to implement the at least one undetermined action based on the comparison.

22. The system according to claim 21, wherein the processor is further configured to:

determine at least one constraint associated with performing maintenance of the infrastructure;

determine a long-term maintenance plan associated with the infrastructure, based on the calculated health index of components in the infrastructure and the determined constraint;

estimate a cost of the determined long-term maintenance plan;

determine a first short-term maintenance plan associated with the infrastructure, the determined first short-term maintenance plan specifying one or more of;

the subset of components, the timing, the intervention actions, and the resource requirement;

perform an action, among the intervention actions, specified in the determined first short-term maintenance plan;

update the determined constraint after performing the specified action;

re-calculate the health index of the components in the infrastructure after performing the specified action; and

repeatedly perform determining the constraint, determining the long-term maintenance plan, estimating the cost, determining the first short-term maintenance plan, performing the action, updating the determined constraint, and re-calculating the health index.

23. A computer program product for maintaining an infrastructure of components, the computer program product comprising a storage medium readable by a processing circuit and storing instructions run by the processing circuit for performing a method, the method comprising:

receiving structured data, unstructured data, and infrastructure data from a database;

running at least one statistical and optimization modeler with one or more of;

the received structured data, the received unstructured data and the received infrastructure data, in order to calculate a health index of at least one component of the infrastructure, the health index representing a health attribute of the at least one component,the calculation of the health index of the at least one component of the infrastructure including;

HI (health index)=×

(F×

R+W)×

1, where F is a vector representing failure probabilities of the at least one component in the infrastructure, R represents a spatial correlation value of the at least one component and the W represents a time elapsed since a last maintenance of the at least one component, and 1 represents a vector of ones, wherein the vector F is a function of λ

exp (Σ

_i=1^kβ

_iZ_i), where λ

is any positive real number, Z_i is a covariate, i is an index, k is a maximum number of the index, and β

_i is a variable representing one or more of;

an age of the at least one component, a diameter of the at least one component, a length of the at least one component, a number of previous failures of the at least one component;

establishing at least one maintenance plan of the infrastructure, based on the calculated health index, each established maintenance plan being associated with at least one health index;

comparing health indices of the at least one established maintenance plan; and

selecting a plan, among the at least one established maintenance plan, whose health index is a maximum among the compared health indices.

24. The computer program product according to claim 23, wherein the method further comprises:

specifying, according to the selected plan, at least one determined action and at least one undetermined action;

implementing the at least one determined action;

re-calculating the health index after the implementation of the at least one determined action;

comparing the re-calculated health index against a pre-determined threshold; and

determining whether to implement the at least one undetermined action based on the comparison.