Data for Doctoral Thesis (J. HORŇAS): A Machine Learning Approach for Fatigue Life Prediction of Additively Manufactured Metal Parts

Title of Doctoral Thesis: A Machine Learning Approach for Fatigue Life Prediction of Additively Manufactured Metal Parts
Author: Jan HORŇAS

Affiliation:
Czech Technical University in Prague
Faculty of Nuclear Sciences and Physical Engineering
Trojanova 13, 12000 Prague, Czechia

The root directory contains a README.pdf file describing the datasets, along with several Comma-Separated Values (CSV) files corresponding to the experimental raw dataset, training (baseline or augmented) and test sets for each permutation (P1, ..., P5), see Tab. 1 in README.pdf. The structure is as follows:

/
├── README.pdf
├── 00_Experimental_Raw_Dataset/
├── 01_Baseline_Dataset/
├── 02_Augmented_Dataset_8_instances/
├── 03_Augmented_Dataset_23_instances/
└── 04_Augmented_Dataset_69_instances/
├── X_Scaled_Augmented_Training_Set_P1_69_instances.csv
├── X_Scaled_Experimental_Test_Set_P1.csv
├── X_Scaled_Experimental_Training_Set_P1.csv
├── y_Augmented_Training_Set_P1_69_instances.csv
├── y_Experimental_Test_Set_P1.csv
├── y_Experimental_Training_Set_P1.csv
├── ...
├── X_Scaled_Augmented_Training_Set_P5_69_instances.csv
├── X_Scaled_Experimental_Test_Set_P5.csv
├── X_Scaled_Experimental_Training_Set_P5.csv
├── y_Augmented_Training_Set_P5_69_instances.csv
├── y_Experimental_Test_Set_P5.csv
└── y_Experimental_Training_Set_P5.csv

The final training datasets used in doctoral thesis are formed by concatenating the experimental and augmented data (generated by variational autoencoder) as described by (Eq. 1) and (Eq. 2) in the README.pdf file.

The labels of individual columns—representing features (X) and target(y)—in the CSV files are denoted as listed in Tab. 2 in README.pdf file.

---------- IF ANY DATASET IS USED, PLEASE CITE ALL RELATED REFERENCES LISTED BELOW ----------

[1] Jan Horňas. “A Machine Learning Approach for Fatigue Life Prediction of Additively Manufactured
Metal Parts”. Doctoral thesis. Prague: Czech Technical University in Prague, 2025.
URL: https://hdl.handle.net/10467/177291.

[2] Jan Horňas et al. “Multivariate interpolation and machine learning models for extreme defects-based
fatigue life prediction of Ti6Al4V specimens fabricated by SLM”. In: Engineering Fracture Mechanics
314 (2025). ISSN: 0013-7944. DOI: https://doi.org/10.1016/j.engfracmech.2024.110756.

[3] Jan Horňas et al. “A machine learning based approach with an augmented dataset for fatigue life
prediction of additively manufactured Ti-6Al-4V samples”. In: Engineering Fracture Mechanics 293
(2023). ISSN: 0013-7944. DOI: https://doi.org/10.1016/j.engfracmech.2023.109709.

[4] Jan Horňas et al. “Modelling fatigue life prediction of additively manufactured Ti-6Al-4V samples
using machine learning approach”. In: International Journal of Fatigue 169 (2023). ISSN: 0142-1123.
DOI: https://doi.org/10.1016/j.ijfatigue.2022.107483.