Prognostic maintenance system and method
A system and method accepts data from direct fault and parameters measurements and historical data in formatted and unformatted form organized by device or system class, specific unit and unit subsystems, and by operation to learn how to predict or correct specified maintenance or other operations. Disclosed examples use known data and learned information historical data to correct or elaborate service requests based on specified and likely (based on learned information) needed maintenance operations. The system and method are applicable to any device requiring prognostic maintenance or other service based on service orders and direct measurement of operating parameters.
1. An improved maintenance method, comprising the steps of:
performing a learning process, including the steps of:
a) receiving existing information associated with maintenance activities previously performed on one or more devices or systems;
b) performing one or more operations on the information received in a) to generate formatted maintenance information;
c) applying one or more learning algorithms to the formatted maintenance information to predict or correct the way in which a maintenance operation should be performed on the one or more devices or systems; and
performing an execution process, including the steps of:
d) receiving a maintenance request associated with a target device or system;
e) filtering the request to correct for any errors in the request; and
f) performing a maintenance operation on the target device or system using the result of c).
2. The improved maintenance method of claim 1 , wherein the existing maintenance information received in a) includes receiving human- or machine-generated maintenance records in free-form or fixed record formats.
3. The improved maintenance method of claim 1 , wherein the existing maintenance information received in a) includes one or more of the following:
previous maintenance operations performed,
diagnostic information indicating the type of maintenance performed, and
direct measurements of faults that implicate required maintenance.
4. The improved maintenance method of claim 1 , wherein the existing maintenance information received in a) includes one or more of the following:
prior readings or measurements,
data source descriptions,
fixed-sized text fields,
variable-sized text fields, or
natural language information.
5. The improved maintenance method of claim 4 , wherein the natural language information includes text forming words, phrases, or sentences, typed or captured by optical character recognition or verbally captured and converted by speech-to-text translation.
6. The improved maintenance method of claim 1 , wherein the existing maintenance information in a) is received in one or more of the following ways:
on paper,
in electronic or computer records,
in database systems,
in a collection directories, folders or files,
data storage accessible through a network; or
a stand-alone filing system.
7. The improved maintenance method of claim 1 , wherein one of the operations performed in b) to generate formatted maintenance information includes sorting the existing maintenance information received in a) into groups that pertain to the same or similar devices or systems.
8. The improved maintenance method of claim 1 , wherein one of the operations performed in b) includes parsing the received information into similar or related datasets.
9. The improved maintenance method of claim 1 , wherein one of the operations performed in b) includes parsing the received information into one or more machine-standardized forms.
10. The improved maintenance method of claim 1 , wherein the one or more learning algorithms applied in c) includes the use of an adaptive learning technique or technology to associate and correct discrepancy narratives contained in the formatted maintenance information.
11. The improved maintenance method of claim 1 , wherein the one or more learning algorithms applied in c) includes analyzing a dataset of prior maintenance records to generate a set of rules, filters, or networks generate a corrected and more complete version of a maintenance request received in d).
12. The improved maintenance method of claim 1 , wherein the one or more learning algorithms applied in c) includes analyzing a dataset of prior maintenance records to predict the maintenance that will have to be performed so that the correct material, operational steps, or personnel will be on hand when the maintenance is performed.
13. The improved maintenance method of claim 1 , wherein the learning process further includes:
rating or scoring the a result of the learning process; and
iterating steps a), b), c), or a combination thereof, to increase the rating or scoring.
14. The improved maintenance method of claim 1 , wherein the devices or systems include aircraft, vehicles, ships, or subsystems thereof.
15. The improved maintenance method of claim 1 , wherein the learning process does not require human intervention.
16. The improved maintenance method of claim 1 , wherein the learning process includes performing one or more of the following functions:
applying knowledge learned to correct new maintenance requests,
adding historical knowledge to new maintenance requests so that they more fully describe work likely work to be performed, and
coding the final resulting work order record to be more fully correct.
17. The improved maintenance method of claim 1 , wherein the algorithm applied in c) includes one or more of the following
decision trees,
support vector machines,
genetic algorithms,
adaptive neural nets,
convolutional nets,
recurrent nets, and
adaptive or associative algorithms.