Systems and methods for reducing manufacturing failure rates

• US 10,482,382 B2
• Filed: 05/09/2017
• Issued: 11/19/2019
• Est. Priority Date: 05/09/2017
• Status: Active Grant

• Current Assignee: Palantir Technologies Incorporated
• Original Assignee: Palantir Technologies Incorporated
• Inventors: Seaton, William, Wiltsche, Clemens, Novick, Myles, Patel, Rootul
• Primary Examiner(s): Pellett, Daniel T
• Application Number:
  US15/590,959
• Publication Number:
  US 20180330241A1
• Time in Patent Office: 924 Days
• Field of Search: None
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition
• CPC Class Codes: G06Q 10/06 Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling

Abstract

Systems and methods are provided for reducing failure rates of a manufactured products. Manufactured products may be clustered together according to similarities in their production data. Manufactured product clusters may be analyzed to determine mechanisms for failure rate reduction, including adjustments to test quality parameters, product formulas, and product processes. Recommended product adjustments may be provided.

Claims

1. A system for reducing failure rates of a manufactured product comprising:

one or more processors; and

memory storing instructions that, when executed by the one or more processors, cause the system to perform;

receiving, from a database, a first product data set associated with a first batch of a first product, the first product data set including a first product formula, a plurality of first product examples, and first test results of the plurality of first product examples, the first test results being associated with a post-manufacturing test, at least a portion of the first product data set being based on first sensor data recorded by one or more first sensors, the first sensor data being stored as time-series data of first sensor measurements;

receiving, from the database, a second product data set associated with a second batch of a second product, the second product being different from the first product, the second product data set including a second product formula, a plurality of second product examples, and second test results of the plurality of second product examples, the second test results being associated with the post-manufacturing test, at least a portion of the second product data set being based on second sensor data recorded by one or more second sensors, the second sensor data being stored as time-series data of second sensor measurements;

clustering, by machine learning, a first product cluster including the first product data set and the second product data set according to a comparison between the first product formula and the second product formula;

determining, based on the clustering, a failure rate reduction mechanism of at least one product of the first product cluster, wherein the determining the failure rate reduction mechanism comprises;

identifying, based on the clustering, the post-manufacturing test from a plurality of different post-manufacturing tests based on one or more failure results of the post-manufacturing test included in the first test results and the second test results; and

modifying, based on the clustering, a passing quality parameter range of the post-manufacturing test such that the failing result is within a modified passing quality parameter range of the post-manufacturing test.

2. The system of claim 1, wherein to cluster, by machine learning, the first product cluster, the system is further caused to:

receive a plurality of additional product data sets including a plurality of additional product formulas;

score the first product formula according to a weight of ingredients of the first product formula and a frequency of ingredients of the first product formula, the second product formula, and the plurality of additional product formulas;

score the second product formula according to a weight of ingredients of the second product formula and a frequency of ingredients of the first product formula, the second product formula, and the plurality of additional product formulas; and

cluster, by machine learning, the first product cluster according to a comparison of the score of the first product formula and the score of the second product formula.

3. The system of claim 1, wherein the system is further caused to:

receive a plurality of additional product data sets including a plurality of additional test results; and

identify the post-manufacturing test according to a frequency of failing results for the post-manufacturing test among the first test results, the second test results, and the plurality of additional test results.

4. The system of claim 1, wherein the system is further caused to:

select one of the first product or second product as a selected product for failure rate reduction according to a comparison between the first test results of the plurality of first product examples and the second test results of the plurality of second product examples; and

determine the failure rate reduction mechanism of the selected product.

5. The system of claim 4, wherein to determine the failure rate reduction mechanism of the selected product, the system is further caused to:

determine a formula adjustment of the selected product.

6. The system of claim 4, wherein to determine the failure rate reduction mechanism of the selected product, the system is further caused to:

determine a process adjustment of the selected product.

7. The system of claim 1, wherein the system is further caused to:

generate a correlation model according to an in-common ingredient in the first product formula and the second product formula and an in-common test result of the first test results and the second test results.

8. The system of claim 7, wherein the system is further caused to:

predict a test result of the first test results according to the correlation model;

compare the predicted test result to an actual test result of the first test results; and

determine a test failure mode according to the comparison.

9. A computer implemented method for reducing failure rates of a manufactured product, the method being performed on a computer system having one or more physical processors programmed with computer program instructions that, when executed by the one or more physical processors, cause the computer system to perform the method, the method comprising:

receiving, from a database, a first product data set associated with a first batch of a first product, the first product data set including a first product formula, a plurality of first product examples, and first test results of the plurality of first product examples, the first test results being associated with a post-manufacturing test, at least a portion of the first product data set being based on first sensor data recorded by one or more first sensors, the first sensor data being stored as time-series data of first sensor measurements;

receiving, from the database, a second product data set associated with a second batch of a second product, the second product being different from the first product, the second product data set including a second product formula, a plurality of second product examples, and second test results of the plurality of second product examples, the second test results being associated with the post-manufacturing test, at least a portion of the second product data set being based on second sensor data recorded by one or more second sensors, the second sensor data being stored as time-series data of second sensor measurements;

clustering, by machine learning, a first product cluster including the first product data set and the second product data set according to a comparison between the first product formula and the second product formula;

determining, based on the clustering, a failure rate reduction mechanism of at least one product of the first product cluster, wherein the determining the failure rate reduction mechanism comprises;

identifying, based on the clustering, the post-manufacturing test from a plurality of different post-manufacturing tests based on one or more failure results of the post-manufacturing test included in the first test results and the second test results; and

modifying, based on the clustering, a passing quality parameter range of the post-manufacturing test such that the failing result is within a modified passing quality parameter range of the post-manufacturing test.

10. The computer implemented method of claim 9, clustering, by machine learning, the first product cluster further comprises:

receiving, by the computer system, a plurality of additional product data sets including a plurality of additional product formulas;

scoring, by the computer system, the first product formula according to a weight of ingredients of the first product formula and a frequency of ingredients of the first product formula, the second product formula, and the plurality of additional product formulas;

scoring, by the computer system, the second product formula according to a weight of ingredients of the second product formula and a frequency of ingredients of the first product formula, the second product formula, and the plurality of additional product formulas; and

clustering, by machine learning by the computer system, the first product cluster according to a comparison of the score of the first product and the score of the second product.

11. The computer implemented method of claim 9, wherein the method further comprises:

receiving, by the computer system, a plurality of additional product data sets including a plurality of additional test results; and

identifying, by the computer system, the post-manufacturing test according to a frequency of failing results for the post-manufacturing test among the first test results, the second test results, and the plurality of additional test results.

12. The computer implemented method of claim 9, wherein the method further comprises:

selecting, by the computer system, one of the first product or second product as a selected product for failure rate reduction according to a comparison between the first test results of the plurality of first product examples and the second test results of the plurality of second product examples; and

determining, by the computer system, the failure rate reduction mechanism of the selected product.

13. The computer implemented method of claim 12, wherein determining the failure rate reduction mechanism of the selected product includes determining a formula adjustment of the selected product.

14. The computer implemented method of claim 12, wherein determining the failure rate reduction mechanism of the selected product includes determining a process adjustment of the selected product.

15. The computer implemented method of claim 9, wherein the method further comprises:

generating, by the computer system, a correlation model according to an in-common ingredient in the first product formula and the second product formula and an in-common test result of the first test results and the second test results.

16. The computer implemented method of claim 15, wherein the method further comprises:

predicting, by the computer system, a test result of the first test results according to the correlation model;

comparing, by the computer system, the predicted test result to an actual test result of the first test results; and

determining, by the computer system, a test failure mode according to the comparison.

17. A non-transitory computer readable medium comprising instructions that, when executed, cause one or more processors to perform:

receiving, from a database, a first product data set associated with a first batch of a first product, the first product data set including a first product formula, a plurality of first product examples, and first test results of the plurality of first product examples, the first test results being associated with a post-manufacturing test, at least a portion of the first product data set being based on first sensor data recorded by one or more first sensors, the first sensor data being stored as time-series data of first sensor measurements;

receiving, from the database, a second product data set associated with a second batch of a second product, the second product being different from the first product, the second product data set including a second product formula, a plurality of second product examples, and second test results of the plurality of second product examples, the second test results being associated with the post-manufacturing test, at least a portion of the second product data set being based on second sensor data recorded by one or more second sensors, the second sensor data being stored as time-series data of second sensor measurements;

clustering, by machine learning, a first product cluster including the first product data set and the second product data set according to a comparison between the first product formula and the second product formula;

determining, based on the clustering, a failure rate reduction mechanism of at least one product of the first product cluster, wherein the determining the failure rate reduction mechanism comprises;

identifying, based on the clustering, the post-manufacturing test from a plurality of different post-manufacturing tests based on one or more failure results of the post-manufacturing test included in the first test results and the second test results; and

modifying, based on the clustering, a passing quality parameter range of the post-manufacturing test such that the failing result is within a modified passing quality parameter range of the post-manufacturing test.