Battery materials screening

1. A computing device, comprising:

at least one memory configured to store computer program instructions; and

at least one processing device operatively coupled to the at least one memory, the at least one processing device configured to execute the computer program instructions to:

receive microstructure generation parameters for a plurality of microstructures, where the microstructure generation parameters include microstructure characteristics;

generate microstructure statistics using a first artificial intelligence (“AI”) model, where the received microstructure generation parameters are inputs for the first AI model;

predict microstructure properties using a second AI model for the microstructures based on the generated microstructure statistics, the received microstructure generation parameters, and battery cell characteristics; and

determine whether at least one of the microstructures is within a predefined energy profile range based on the predicted microstructure properties;

validate the at least one of the microstructures by computer simulation, where results of the validated microstructure are used to retrain the second AI model.

2. The computing device of claim 1 , wherein the first AI model is a feed-forward neural network or a Gaussian process regression.

3. The computing device of claim 1 , wherein the second AI model is a feed-forward neural network or a Gaussian process regression.

4. The computing device of claim 1 , wherein the microstructure generation parameters include particle size distribution.

5. The computing device of claim 1 , wherein the microstructure characteristics include a first main particle diameter mean, a second main particle diameter mean, a first main scaling factor, a second main scaling factor, a particle diameter standard deviation for a first main particles, a particle diameter standard deviation for a second main particle, a solid volume percentage, a binder weight percentage, a binder contact angle, a binder anisotropy factor, a random seed, and combinations thereof.

6. The computing device of claim 1 , wherein the battery cell characteristics include a cathode thickness, a Lithium-ion (Li+) concentration, an applied charge rate (C-rate), a diffusion coefficient of the active material, and combinations thereof.

7. The computing device of claim 1 , wherein the generated microstructure statistics include an active surface area, an effective diffusivity, a conductivity of an electrolyte, a volume fraction of an active material, a volume fraction of a binder material, a volume fraction of the electrolyte, and combinations thereof.

8. The computing device of claim 1 , wherein the predicted microstructure properties include a specific power, a specific energy, discharge curves, and combinations thereof.

9. The computing device of claim 1 , further comprising the computer program instructions:

define an energy profile range;

predict at least one microstructure using at least one complex AI model; and

output the predicted microstructure,

wherein the at least one complex AI model is retrained using uncertainty and new data generation from the second AI model.

10. The computing device of claim 9 , wherein the at least one complex AI model is a selected one from a group of a convolutional neural network, a deep neural network, a generative adversarial network, and a variation autoencoder.

11. The computing device of claim 1 , wherein in the predict microstructure properties computer program instruction, further comprises the sub-instructions of:

predict energy profiles for the microstructures using the second AI model, wherein the second AI model is a simple neural network;

select microstructure candidates from the microstructures based on the predicted energy profiles; and

predict energy profiles for the selected microstructure candidates using a deep neural network.

12. A computer-implemented method for battery material screening, comprising:

receiving microstructure generation parameters for a plurality of microstructures, where the microstructure generation parameters include microstructure characteristics;

generating microstructure statistics using a first artificial intelligence (“AI”) model, where the received microstructure generation parameters are inputs for the first AI model;

predicting microstructure properties using a second AI model for the microstructures based on the generated microstructure statistics, the received microstructure generation parameters, and battery cell characteristics; and

determining whether at least one of the microstructures is within a predefined energy profile range based on the predicted microstructure properties;

validating the at least one of the microstructures by computer simulation, wherein results of the validated microstructure are used to retrain the second AI model.

13. The computer-implemented method of claim 12 , wherein the first AI model is a feed-forward neural network or a Gaussian process regression, and wherein the second AI model is a feed-forward neural network or a Gaussian process regression.

14. The computer-implemented method of claim 13 , wherein the microstructure generation parameters include particle size distribution, wherein the microstructure characteristics include a first main particle diameter mean, a second main particle diameter mean, a first main scaling factor, a second main scaling factor, a particle diameter standard deviation for a first main particles, a particle diameter standard deviation for a second main particle, a solid volume percentage, a binder weight percentage, a binder contact angle, a binder anisotropy factor, a random seed, and combinations thereof, wherein the battery cell characteristics include a cathode thickness, a Lithium-ion (Li+) concentration, an applied charge rate (C-rate), a diffusion coefficient of the active material, and combinations thereof, wherein the generated microstructure statistics include an active surface area, an effective diffusivity, a conductivity of an electrolyte, a volume fraction of an active material, a volume fraction of a binder material, a volume fraction of the electrolyte, and combinations thereof, and wherein the predicted microstructure properties include a specific power, a specific energy, discharge curves, and combinations thereof.

15. The computer-implemented method of claim 13 , further comprising the steps of:

defining an energy profile range;

predicting at least one microstructure using at least one complex AI model; and

outputting the predicted microstructure,

wherein the at least one complex AI model is retrained using uncertainty and new data generation from the second AI model.

16. A non-transitory computer readable medium encoded with instructions that when executed by at least one processor causes the processor to carry out the following operations:

receive microstructure generation parameters for a plurality of microstructures, where the microstructure generation parameters include microstructure characteristics;

generate microstructure statistics using a first artificial intelligence (“AI”) model, where the received microstructure generation parameters are inputs for the first AI model;

predict microstructure properties using a second AI model for the microstructures based on the generated microstructure statistics, the received microstructure generation parameters, and battery cell characteristics; and

determine whether at least one of the microstructures is within a predefined energy profile range based on the predicted microstructure properties;

validate the at least one of the microstructures by computer simulation, wherein results of the validated microstructure are used to retrain the second AI model.

17. The non-transitory computer readable medium of claim 16 , wherein the first AI model is a feed-forward neural network or a Gaussian process regression, wherein the second AI model is a feed-forward neural network or a Gaussian process regression, wherein the microstructure generation parameters include particle size distribution, wherein the microstructure characteristics include a first main particle diameter mean, a second main particle diameter mean, a first main scaling factor, a second main scaling factor, a particle diameter standard deviation for a first main particles, a particle diameter standard deviation for a second main particle, a solid volume percentage, a binder weight percentage, a binder contact angle, a binder anisotropy factor, a random seed, and combinations thereof, wherein the battery cell characteristics include a cathode thickness, a Lithium-ion (Li+) concentration, an applied charge rate (C-rate), a diffusion coefficient of the active material, and combinations thereof, wherein the generated microstructure statistics include an active surface area, an effective diffusivity, a conductivity of an electrolyte, a volume fraction of an active material, a volume fraction of a binder material, a volume fraction of the electrolyte, and combinations thereof, and wherein the predicted microstructure properties include a specific power, a specific energy, discharge curves, and combinations thereof.

18. The non-transitory computer readable medium of claim 17 , further comprising the operations of:

define an energy profile range;

predict at least one microstructure using at least one complex AI model; and

output the predicted microstructure,

wherein the at least one complex AI model is retrained using uncertainty and new data generation from the second AI model.