The aim of the work is to investigate the effectiveness of selected classification algorithms and their extensions in assessing microstructure of castings. Experiments were carried out in which the prepared algorithms and machine learning methods were tested in various conditions and configurations, as well as for various input data, which are photos of castings (photos of the microstructure) or information about the material (e.g., type, composition). As shown by the literature review, there are few scientific papers on this subject (i.e., in the use of machine learning to assess the quality of the microstructure and the obtained strength properties of cast iron). The effectiveness of machine learning algorithms in assessing the quality of castings will be tested using the most universal methods. Results obtained by classic machine learning methods and by neural networks will be compared with each other, taking into account aspects such as interpretability of results, ease of model implementation, algorithm simplicity, and learning time.

The structural stability of ausferrite in thin-walled Austempered Ductile Iron (ADI) castings with 5 mm wall thickness is compared to a reference casting with 25 mm wall thickness. The thin-walled and reference castings were first austenitized between 850 and 925 °C, and then austempered between 250 and 380 °C. X-ray diffraction (XRD) investigations with changing temperature were performed between − 260 up to + 450 °C to investigate the change of phase fraction, lattice parameters and strain in ausferrite. The role of the austenitization temperature on structural stability and homogeneity of the investigated ADI castings has been provided. In addition, the problem of the occurrence of “blocky” high-carbon austenite that was not completely involved during austempering, has been taken into account. Finally, it has been shown that the thin-walled castings provided higher structural homogeneity and stability if compared to the reference castings.

The aim of the study was to design and optimize the complex, two-step heat treatment of Cu–Ni (Mn, Mo) ductile iron. A method for the formation and investigation of austempered ductile iron (ADI) by means of complex two-step and as comparative, standard one-step heat treatments has been developed, using quenching dilatometer. Investigations of proceeding phase transformations using differential dilatometric and DSC analysis supported by microstructural observations, hardness and austenite volume measurements have been carried out. An analysis of the temperature sequence of the ausferrite decomposition in the one-step and two-step ADI was performed, which allowed for the separation and identification of the effects responsible for the carbon-enriched austenite decomposition. A quantitative relationship was established between basic dimensional effects revealed on the differential dilatometric curve of ausferrite decomposition, which enables prognosis and optimization of the parameters of complex ADI heat treatment variants. Verification tests were performed on a stand equipped with salt furnaces enabling a quick transfer of samples from one bath to another without changing their initial temperature. Optimization of the two-step ADI heat treatment with the use of quantitative dilatometric analysis of the ausferrite decomposition, allowed to obtain for the temperature step-down heat treatment 390 °C/15, 20 min ≫ 270 °C/130 min the excellent mechanical properties, unattainable by means of standard 1-step ADI heat treatment.

The work presents a tool enabling the selection of powder for 3D printing. The project focused on three types of powders, such as steel, nickel- and cobalt-based and aluminum-based. An important aspect during the research was the possibility of obtaining the mechanical parameters. During the work, the possibility of using the selected algorithm based on artificial intelligence like Random Forest, Decision Tree, K-Nearest Neighbors, Fuzzy K-Nearest Neighbors, Gradient Boosting, XGBoost, AdaBoost was also checked. During the work, tests were carried out to check which algorithm would be best for use in the decision support system being developed. Cross-validation was used, as well as hyperparameter tuning using different evaluation sets. In both cases, the best model turned out to be Random Forest, whose F1 metric score is 98.66% for cross-validation and 99.10% after tuning on the test set. This model can be considered the most promising in solving this problem. The first result is a more accurate estimate of how the model will behave for new data, while the second model talks about possible improvement after optimization or possible overtraining to the parameters.

This paper presents an assessment of the possibility of using digital image classifiers for tomographic images concerning ductile iron castings. The results of this work can help the development of an efficient system suggestion allowing for decision making regarding the qualitative assessment of the casting process parameters. Special attention should be focused on the fact that automatic classification in the case of ductile iron castings is difficult to perform. The biggest problem in this aspect is the high similarity of the void image, which may be a sign of a defect, and the nodular graphite image. Depending on the parameters, the tests on different photos may look similar. Presented in this article are test scenarios of the module analyzing two-dimensional tomographic images focused on the comprehensive assessment by convolutional neural network models, which are designed to classify the provided image. For the purposes of the tests, three such models were created, different from each other in terms of architecture and the number of hyperparameters and trainable parameters. The described study is a part of the decision-making system, supporting the process of qualitative analysis of the obtained cast iron castings.

The article addresses stress formation in the structural 3D-printed elements of a high-pressure die casting die mould used for production of aluminum castings. The 3D-printed elements with conformal cooling are manufactured of 18Ni300 powder. Initial numerical calculations were performed on a test die mould made of standard steel X40CrMoV5 to determine temperature distribution and stress state, providing a baseline for comparing 3D-printed 18Ni300 parts. A database for 18Ni300 material was developed, including optimal heat treatment parameters: aged at 560 °C for 8 h. The resulting tensile strength of approximately ~1600 MPa, yield strength 1550 MPa, and elongation 6–7%, with properties temperature-dependent from 20 °C to 600 °C. Results show that conformal cooling increases stress gradients, highlighting the demands on fatigue strength at elevated temperatures. The study revealed that the heat treatment significantly influences the final properties, with tensile strengths of 1400–2000 MPa and elongation from 1 to 8%. While the heat treatment has a greater impact on the mechanical properties than the printing parameters, optimizing the printing settings remains crucial for ensuring density and quality in the die moulds under cyclic loads.