Augmented reliability models for design and manufacturing
A method for generating an augmented reliability performance model for a product includes obtaining a reliability performance model for the product, developing a reliability prediction machine learning model for predicting reliability performance of the product based on data obtained from manufacturing and testing of the product, and obtaining, from development of the machine learning model, feature names for the machine learning model and their predictive power values. The feature names may correspond to features from the data obtained from manufacturing and testing of the product. The method may further include extracting a set of feature names corresponding to features having highest predictive power values from the feature names, and generating the augmented reliability performance model for the product by modifying the reliability performance model to incorporate model parameters derived from the set of feature names.
1. A method for generating an augmented reliability performance model for a product, the method comprising:
obtaining a reliability performance model for the product;
developing a reliability prediction machine learning model for predicting reliability performance of the product based on data obtained from manufacturing and testing of the product;
obtaining, from development of the machine learning model, feature names for the machine learning model and their predictive power values, wherein the feature names correspond to features from the data obtained from manufacturing and testing of the product;
extracting a set of feature names corresponding to features having highest predictive power values from the feature names; and
generating the augmented reliability performance model for the product by modifying the reliability performance model to incorporate one or more model parameters derived from the set of feature names.
2. The method of claim 1 , wherein extracting the set of feature names corresponding to the features having highest predictive power values comprises extracting one or more feature names corresponding to the features having predictive power values exceeding a threshold predictive power value.
3. The method of claim 1 , wherein the feature names for the machine learning model comprise metadata of the machine learning model.
4. The method of claim 1 , wherein extracting the set of feature names corresponding to the features having highest predictive power values comprises extracting a predetermined number of feature names.
5. The method of claim 1 , wherein modifying the reliability performance model to incorporate model parameters derived from the set of feature names comprises adding one or more new model parameters derived from the set of feature names.
6. The method of claim 1 , wherein modifying the reliability performance model to incorporate model parameters derived from the set of feature names comprises modifying one or more existing model parameters based on information derived from the set of feature names.
7. The method of claim 1 , further comprising:
translating the set of feature names into one or more design parameters; and
incorporating the one or more design parameters into the augmented reliability performance model.
8. The method of claim 7 , wherein incorporating the one or more design parameters into the augmented reliability performance model comprises adding one or more new model parameters to the augmented reliability performance model.
9. The method of claim 7 , wherein incorporating the one or more design parameters into the augmented reliability performance model comprises modifying one or more existing parameters of the augmented reliability performance model.
10. The method of claim 1 , further comprising:
extracting one or more feature names corresponding to features having lowest predictive power from the feature names; and
modifying the reliability performance model by removing one or more existing model parameters based on the one or more feature names to generate the augmented reliability performance model for the product.
11. A method for designing a product, the method comprising:
obtaining a design reliability model for the product, wherein obtaining the design reliability model comprises:
developing a reliability prediction machine learning model for predicting reliability performance of the product based on data obtained from manufacturing and testing of the product;
obtaining, from development of the machine learning model, feature names for the machine learning model and their predictive power values, wherein the feature names correspond to features from the data obtained from manufacturing and testing of the product;
extracting a set of feature names corresponding to the features having highest predictive power from the feature names; and
modifying a reliability performance model by incorporating one or more model parameters derived from the set of feature names to generate the design reliability model for the product; and
utilizing the design reliability model for at least one of design of the product and design of a process for manufacturing the product.
12. The method of claim 11 , wherein extracting the set of feature names corresponding to the features having highest predictive power values comprises extracting one or more feature names having predictive power values exceeding a threshold predictive power value.
13. The method of claim 11 , wherein the feature names for the machine learning model comprise metadata of the machine learning model.
14. The method of claim 11 , wherein extracting the set of feature names corresponding to the features having highest predictive power values comprises extracting a predetermined number of feature names.
15. The method of claim 11 , wherein modifying the reliability performance model to incorporate one or more model parameters derived from the set of feature names comprises adding one or more new model parameters derived from the set of feature names.
16. The method of claim 11 , wherein modifying the reliability performance model to incorporate one or more model parameters derived from the set of feature names comprises modifying one or more existing model parameters based on information derived from the set of feature names.
17. The method of claim 11 , further comprising:
translating the set of feature names into one or more design parameters; and
incorporating the one or more design parameters into the design reliability model.
18. The method of claim 17 , wherein incorporating the one or more design parameters into the design reliability model comprises adding one or more new model parameters to the design reliability model.
19. The method of claim 17 , wherein incorporating the one or more design parameters into the design reliability model comprises modifying one or more existing parameters of the design reliability model.
20. The method for claim 11 , further comprising:
extracting one or more feature names corresponding to the features having lowest predictive power from the feature names; and
modifying the reliability performance model by removing one or more existing model parameters based on the one or more feature names to generate the design reliability model for the product.