Method and apparatus for characterizing composite materials using an artificial neural network

1. A method for predicting an impact resistance of a composite material, the method comprising:

designing an artificial neural network including a plurality of neurons;

employing, by a processor, a sensitivity analysis to identify a parameter and quantitatively describe a degree of influence of the parameter on the impact resistance of the composite material;

training, by the processor, the artificial neural network to predict the impact resistance by adjusting an output of the plurality of neurons according to the parameter and the degree of influence identified in the employed sensitivity analysis;

inputting data of the composite material into the artificial neural network; and

utilizing the artificial neural network to predict the impact resistance of the composite material.

2. The method according to claim 1 , wherein the artificial neural network includes:

an input layer of neurons that receives data that is input into the artificial neural network; and

an output layer of neurons that outputs a prediction of the impact resistance of the composite material.

3. The method according to claim 1 , wherein training the artificial neural network further includes:

inputting sample data to the artificial neural network;

measuring an error between known results of the sample data and the prediction output from the artificial neural network; and

reducing the error by applying a variable weighting factor to each neuron of the plurality of neurons in the artificial neural network to adjust an output of each neuron.

4. The method according to claim 3 , wherein the error is a mean-squared error.

5. The method according to claim 1 , wherein the input data of the composite material includes:

a stacking sequence of layers in the composite material;

a layer thickness;

a number of layers in the composite material;

an orientation angle of the layers in the composite material; and

a material composition of the layers in the composite material.

6. The method according to claim 1 , wherein the artificial neural network is a feed forward network.

7. The method according to claim 2 , wherein the artificial neural network further includes a hidden layer comprising a plurality of neurons,

the hidden layer receives data output from the input layer, and

the hidden layer outputs processed data to the output layer.

8. The method according to claim 7 , wherein training the artificial neural network further includes:

inputting sample data to the input layer;

measuring an error between known results of the sample data and the prediction output from the output layer; and

reducing the error by:

managing the hidden layer such that data output from neurons in the input layer may be selected for input to individual neurons in the hidden layer, and

applying a variable weighting factor to each neuron of the plurality of neurons in the artificial neural network to adjust an output of each neuron.

9. A device for predicting an impact resistance of a composite material, the device comprising:

a processor configured to:

design an artificial neural network including a plurality of neurons;

employ a sensitivity analysis to identify a parameter and quantitatively describe a degree of influence of the parameter on the impact resistance of the composite material;

train the artificial neural network to predict the impact resistance by adjusting an output the plurality of neurons according to the parameter and the degree of influence identified in the employed sensitivity analysis;

input data of the composite material into the artificial neural network; and

utilize the artificial neural network to predict the impact resistance of the composite material.

10. A system for predicting an impact resistance of a composite material, the system comprising:

the device according to claim 9 ; and

the artificial neural network.

11. The device according to claim 9 , wherein the artificial neural network includes:

an input layer of neurons that receives data that is input into the artificial neural network; and

an output layer of neurons that outputs a prediction of the impact resistance of the composite material.

12. The device according to claim 9 , wherein training the artificial neural network further includes:

inputting sample data to the artificial neural network;

measuring an error between known results of the sample data and the prediction output from the artificial neural network; and

reducing the error by applying a variable weighting factor to each neuron of the plurality of neurons in the artificial neural network to adjust an output of each neuron.

13. A non-transitory computer readable medium storing computer readable instructions that when executed by a computer cause the computer to perform a method comprising:

designing an artificial neural network including a plurality of neurons;

employing a sensitivity analysis to identify a parameter and quantitatively describe a degree of influence of the parameter on the impact resistance of the composite material;

training the artificial neural network to predict the impact resistance by adjusting an output of the plurality of neurons according to the parameter and the degree of influence identified in the employed sensitivity analysis;

inputting data of the composite material into the artificial neural network; and

utilizing the artificial neural network to predict the impact resistance of the composite material.

14. The non-transitory computer readable medium of claim 13 , wherein the artificial neural network includes:

an input layer of neurons that receives data that is input into the artificial neural network; and

an output layer of neurons that outputs a prediction of the impact resistance of the composite material.

15. The non-transitory computer readable medium of claim 13 , wherein training the artificial neural network further includes:

inputting sample data to the artificial neural network;

measuring an error between known results of the sample data and the prediction output from the artificial neural network; and

reducing the error by applying a variable weighting factor to each neuron of the plurality of neurons in the artificial neural network to adjust an output of each neuron.

16. The non-transitory computer readable medium of claim 15 , wherein the error is a mean-squared error.

17. The non-transitory computer readable medium of claim 13 , wherein the input data of the composite material includes:

a stacking sequence of layers in the composite material;

a layer thickness;

a number of layers in the composite material;

an orientation angle of the layers in the composite material; and

a material composition of the layers in the composite material.

18. The non-transitory computer readable medium of claim 13 , wherein the artificial neural network is a feed forward network.

19. The non-transitory computer readable medium of claim 14 , wherein the artificial neural network further includes a hidden layer comprising a plurality of neurons,

the hidden layer receives data output from the input layer, and

the hidden layer outputs processed data to the output layer.

20. The non-transitory computer readable medium of claim 19 , wherein training the artificial neural network further includes:

inputting sample data to the input layer;

measuring an error between known results of the sample data and the prediction output from the output layer; and

reducing the error by:

managing the hidden layer such that data output from neurons in the input layer may be selected for input to individual neurons in the hidden layer, and

applying a variable weighting factor to each neuron of the plurality of neurons in the artificial neural network to adjust an output of each neuron.