Codes for the paper "Modeling of forced-vibration systems using continuous-time state-space neural network"
--- Coded in the Python environment utilizing the powerful deep learning library PyTorch.
- Zener_Sin_RNN_main.py is the main Python file for the quick demonstration of the RNN model.
- Zener_CSNN_main.py is the main Python file for the SINN model.
- Zener_preparation.m is the Matlab file to generate the banded limited white noises as input signals.
- Zener_data.mat is the generated banded limited white noises.
- Zener_CSNN_2layer1_2state_2neuron1_6000_10s_loss2834em1.pt is the trained library of CSNN model, users could load it in Zener_CSNN_main.py to genetate Zener_CSNN_2layer1_2state_2neuron1.mat quickly.
- Zener_CSNN_2layer1_2state_2neuron1.mat stores the simulation results.
- Zener_CSNN_plot.py plots the results.
- num_CSNN_main.py is the main Python file for the CSNN model.
- num_CSNN_2layer1_2state_2neuron1_3000_loss5316em2.pt is the trained library of CSNN model, users could load it in num_main.py to genetate num_data_pred1_2layer1_2state_2neuron1.mat quickly.
- num_preparation.py prepares the data.
- results_num_ag2u.mat is the data file coming from the PhyCNN paper "Physics-guided convolutional neural network (PhyCNN) for data-driven seismic response modeling".
- numa_ref.mat stores the reference results.
- num_PhyCNN.mat stores the prediction results using the PhyCNN model.
- num_CSNN_2layer1_2state_2neuron1.mat stores the prediction results using the CSNN model.
- num_CSNN_plot.py plots the results.
- exp_CSNN_main.py is the main Python file for the CSNN model.
- exp_CSNN_1layer_3state_3neuron1_acc_roof_10000_loss9413ep3.pt is the trained library of CSNN model, users could load it in exp_CSNN_main.py to genetate exp_CSNN_1layer_3state_3neuron1.mat quickly.
- results_exp_ag2utt.mat is the data file coming from the PhyCNN paper
- exp_ref.mat stores the reference results.
- exp_PhyCNN.mat stores the prediction results using the PhyCNN model.
- exp_CSNN_1layer_3state_3neuron1.mat stores the prediction results using the CSNN model.
- exp_CSNN_plot.py plots the results.