Data for: Multivariate interpolation and machine learning models for extreme defects-based fatigue life prediction of Ti6Al4V specimens fabricated by SLM

Data for the paper:

"Multivariate interpolation and machine learning models for extreme defects-based fatigue life prediction of Ti6Al4V specimens fabricated by SLM"
by: J. Horňas, A. Materna, J. Glinz, M. Yosifov, S. Senck

DOI: https://doi.org/10.1016/j.engfracmech.2024.110756

The paper proposed a novel methodology for extreme defects-based fatigue life prediction of Ti6Al4V specimens fabricated by selective laser melting (SLM) technique. The introduced framework is represented by a combination of a traditional ranking method based on the extreme values of stress intensity factor with related defect parameters (size, distance from the free surface, sphericity and compactness), training set augmentation using variational autoencoder (VAE) and various data-driven models (multivariate interpolation and machine learning algorithms). The defects were observed using micro-computed tomography (µ-CT) prior to the fatigue tests.

The following data are shared:

1) Experimental (raw) dataset:
Experimnetal_Raw_Dataset.csv

2) Experimental (features scaled by Min-Max normalization) dataset:
Experimnetal_Scaled_Dataset.csv

3) Experimental training set (features scaled by Min-Max normalization):
Experimental_Training_Set.csv

4) Augmented training set generated by variational autoencoder (VAE):
Augmented_Training_Set_by_VAE.csv

5) Final training set (experimental + augmented):
Final_Training_Set.csv

6) Training subsets for features (X) and target (y) during hyperparameters tuning using tree-structured Parzen estimator (TPE) with 23-fold cross-validation (CV):
X_Training_Subset_Cross_Validation_Number_01.csv,
... ,
X_Training_Subset_Cross_Validation_Number_23.csv

y_Training_Subset_Cross_Validation_Number_01.csv,
... ,
y_Training_Subset_Cross_Validation_Number_23.csv

7) Validation subsets for features (X) and target (y) during hyperparameters tuning using tree-structured Parzen estimator (TPE) with 23-fold cross-validation (CV):
X_Validation_Subset_Cross_Validation_Number_01.csv,
... ,
X_Validation_Subset_Cross_Validation_Number_23.csv

y_Validation_Subset_Cross_Validation_Number_01.csv,
... ,
y_Validation_Subset_Cross_Validation_Number_23.csv

8) Training sets of features (X) and target (y) for models with tuned hyperparamaters:
X_Training_Set.csv
y_Training_Set.csv

9) Test sets (unseen data) of features (X) and target (y) for final evaluation of prediction accuracy:
X_Test_Set.csv
y_Test_Set.csv