Method and system for determining ply-by-ply damage in a composite structure

1 . A computer-implemented method for determining ply-by-ply damage in a three-dimensional (3D) composite structure, method comprising:

receiving a 3D geometric model at a computing system, wherein the 3D geometric model specifies geometric aspects of the 3D composite structure, and wherein the 3D geometric model facilitates performance of finite element analysis (FEA);

applying each material failure model (MFM) of a plurality of MFMs to the 3D geometric model using FEA logic of the computing system, wherein the plurality of MFMs are configured to predict different types of damage to the 3D composite structure due to an applied force or stress, and wherein each MFM relates the applied force or stress to a particular type of damage to a material;

generating a respective damage prediction image for each ply of the 3D composite structure using the FEA logic and for each MFM of the plurality of MFMs;

inputting a plurality of damage prediction images from the plurality of MFMs into an input layer of a deep learning model implemented by machine learning logic of the computing system; and

generating an aggregate prediction of damage to the 3D composite structure using the deep learning model, wherein:

the aggregate prediction is based on the plurality of damage prediction images,

the aggregate prediction comprises a 3D composite image, and

the 3D composite image identifies a respective damage location and a respective extent of the damage for each ply of the 3D composite structure.

2 . The computer-implemented method according to claim 1 , wherein generating the aggregate prediction comprises:

outputting the 3D composite image.

3 . The computer-implemented method according to claim 1 , wherein an MFM of the plurality of MFMs predicts stressed-based damage to the 3D composite structure due to the applied force or stress.

4 . The computer-implemented method according to claim 1 , wherein an MFM of the plurality of MFMs predicts fracture-based damage to the 3D composite structure due to the applied force or stress.

5 . The computer-implemented method according to claim 1 , wherein the 3D geometric model specifies each ply of the 3D composite structure with one element.

6 . The computer-implemented method according to claim 1 , wherein each damage prediction image of the plurality of damage prediction images depicts the damage that is predicted according to a particular MFM to occur to the 3D composite structure.

7 . The computer-implemented method according to claim 1 , wherein each damage prediction image of the plurality of damage prediction images depicts a binary mask indicative of where the damage is predicted according to a particular MFM to occur to the 3D composite structure.

8 . The computer-implemented method according to claim 1 , wherein each damage prediction image of the plurality of damage prediction images depicts a gradient that represents different amounts of damage that are predicted according to a particular MFM to occur to the 3D composite structure.

9 . The computer-implemented method according to claim 1 , wherein each damage prediction image of the plurality of damage prediction images depicts a respective composite image of damage predicted according to a particular MFM to occur on each layer of the 3D composite structure.

10 . The computer-implemented method according to claim 1 , wherein:

the 3D geometric model specifies geometric aspects of a composite panel; and

the plurality of MFMs predict different types of damage to the composite panel due to a ballistic impact.

11 . A computing system that for determining ply-by-ply damage in a three-dimensional (3D) composite structure, the computing system comprising:

one or more instruction storage devices for storing instruction code; and

one or more processors in communication with the one or more instruction storage devices, wherein execution of the instruction code by the one or more processors causes the computing system to perform operations comprising:

receiving a 3D geometric model at a computing system, wherein the 3D geometric model specifies geometric aspects of the 3D composite structure, and wherein the 3D geometric model facilitates performance of finite element analysis (FEA);

applying each material failure model (MFM) of a plurality of MFMs to the 3D geometric model using FEA logic of the computing system, wherein the plurality of MFMs are configured to predict different types of damage to the 3D composite structure due to an applied force or stress, and wherein each MFM relates the applied force or stress to a particular type of damage to a material;

generating a respective damage prediction image for each ply of the 3D composite structure using the FEA logic and for each MFM of the plurality of MFMs;

inputting a plurality of damage prediction images from the plurality of MFMs into an input layer of a deep learning model implemented by machine learning logic of the computing system; and

generating an aggregate prediction of damage to the 3D composite structure using the deep learning model, wherein:

the aggregate prediction is based on the plurality of damage prediction images,

the aggregate prediction comprises a 3D composite image, and

the 3D composite image identifies a respective damage location and a respective extent of the damage for each ply of the 3D composite structure.

12 . The computing system according to claim 11 , wherein generating the aggregate prediction comprises:

outputting the 3D composite image.

13 . The computing system according to claim 11 , wherein an MFM of the plurality of MFMs predicts stressed-based damage to the 3D composite structure due to the applied force or stress.

14 . The computing system according to claim 11 , wherein an MFM of the plurality of MFMs predicts fracture-based damage to the 3D composite structure due to the applied force or stress.

15 . The computing system according to claim 11 , wherein the 3D geometric model specifies each ply of the 3D composite structure with one element.

16 . The computing system according to claim 11 , wherein each damage prediction image of the plurality of damage prediction images depicts the damage that is predicted according to a particular MFM to occur to the 3D composite structure.

17 . The computing system according to claim 11 , wherein each damage prediction image of the plurality of damage prediction images depicts a binary mask indicative of where the damage is predicted according to a particular MFM to occur to the 3D composite structure.

18 . The computing system according to claim 11 , wherein each damage prediction image of the plurality of damage prediction images depicts a gradient that represents different amounts of damage that are predicted according to a particular MFM to occur to the 3D composite structure.

19 . The computing system according to claim 11 , wherein each damage prediction image of the plurality of damage prediction images depicts a respective composite image of damage predicted according to a particular MFM to occur on each layer of the 3D composite structure.

20 . A non-transitory computer-readable medium that stores instruction code that facilitates determining ply-by-ply damage in a three-dimensional (3D) composite structure, wherein execution of the instruction code by one or more processors of a computing system causes the computing system to perform operations comprising:

receiving a 3D geometric model at a computing system, wherein the 3D geometric model specifies geometric aspects of the 3D composite structure, and wherein the 3D geometric model facilitates performance of finite element analysis (FEA);

applying each material failure model (MFM) of a plurality of MFMs to the 3D geometric model using FEA logic of the computing system, wherein the plurality of MFMs are configured to predict different types of damage to the 3D composite structure due to an applied force or stress, and wherein each MFM relates the applied force or stress to a particular type of damage to a material;

generating a respective damage prediction image for each ply of the 3D composite structure using the FEA logic and for each MFM of the plurality of MFMs;

inputting a plurality of damage prediction images from the plurality of MFMs into an input layer of a deep learning model implemented by machine learning logic of the computing system; and

generating an aggregate prediction of damage to the 3D composite structure using the deep learning model, wherein:

the aggregate prediction is based on the plurality of damage prediction images,

the aggregate prediction comprises a 3D composite image, and

the 3D composite image identifies a respective damage location and a respective extent of the damage for each ply of the 3D composite structure.